Docstoc

Winemakers' Federation of Australia

Document Sample
Winemakers' Federation of Australia Powered By Docstoc
					                               Canberra Office
                               Level 1/24 Bougainville Street
                               Manuka ACT 2603
                               (PO Box 3891, Manuka ACT 2603)
                               Telephone: 02 6239 8300
                               Facsimile: 02 6260 7900
                               Email: wfa@wfa.org.au
                               Web: www.wfa.org.au
                               ABN 38 359 406 467




Winemakers’ Federation of Australia

               Submission to


              AUSTRALIA:
        THE HEALTHIEST COUNTRY
                BY 2020

     A discussion paper prepared by the
   National Preventative Health Taskforce
Table of Contents

1.      Executive Summary................................................................................3

2.      Introduction.............................................................................................8

3.     The National Preventative Health Taskforce and Strategy..................9
     3.1   Australia: The Healthiest Country by 2020....................................9
     3.2   Targets............................................................................................10

4.     Overview of the Australian wine sector ..............................................11
     4.1  Background and Importance to Australia ...................................11
     4.2  Rapid Growth .................................................................................11
     4.3  Regional Impacts...........................................................................12
     4.4  Winery Viability..............................................................................13
     4.5  Today’s Challenges .......................................................................14

5.     Wine and Alcohol consumption – Trends ..........................................16
     5.1  Consumption trends......................................................................16
     5.2  Short- and long-term harm from alcohol misuse........................19
     5.3  Young and under-age consumers................................................19
     5.4  Consumption by product ..............................................................21
     5.5  Location of consumption..............................................................24
     5.6  Alcohol consumption in the indigenous community .................27

6.     The Australian wine sector and alcohol policy ..................................28
     6.1   Taxation..........................................................................................28
     6.2   Labelling.........................................................................................29
     6.3   Advertising.....................................................................................29
     6.4   DrinkWise Australia.......................................................................30
     6.5   National Wine Foundation ............................................................31

7.     Collins and Lapsley Papers .................................................................32
     7.1   Social costs of alcohol abuse ......................................................32
     7.2   Avoidable costs of alcohol abuse ................................................34

8.     Discussion Paper – Specific Questions .............................................36
     8.1   National Prevention Agency .........................................................38
     8.2   Measurements ...............................................................................38
     8.3   Taxation..........................................................................................40

9.      General Comments...............................................................................43

References ...................................................................................................44

Appendix A ...................................................................................................46




                                                                                                                2
1. Executive Summary
This submission has been prepared by the Winemakers’ Federation of Australia
(WFA) on behalf of the Australian wine sector and in consultation with Creina
Stockley of The Australian Wine Research Institute (AWRI), who acted as an
independent scientific expert.

The Australian wine sector welcomes the opportunity to provide input into the
development of the National Preventative Health Strategy and looks forward to
working with the Taskforce through this submission process, and the ongoing
development of the strategy, as outlined in the Terms of Reference:

       “The taskforce will use a multidisciplinary approach, operate in a
       collaborative, open and consultative manner, and work in partnership with
       existing agencies and bodies working in associated areas.”

The issue of alcohol misuse and its impact on society is complex. Compared with
some other stakeholders, the Australian wine sector as part of the supply and sale of
an alcoholic beverage has different challenges and considerations, but brings unique
drivers and opportunities to positively influence and reinforce the responsible
consumption of wine.

Preventative Health Taskforce and Strategy

In particular, the Australian wine sector welcomes the emphasis by the Taskforce on
“reshaping attitudes and behaviours, rather than prohibiting them.”

The Australian wine sector’s submission is mindful of the Taskforce’s focus on, and
strategies for, reducing harmful alcohol consumption rather than all alcohol
consumption. These strategies will impact on wine consumers, the majority of whom
generally consume responsibly, as well as on the minority who consume
irresponsibly. Consequently, there will be an impact on wine producers. Indeed,
there will be unanticipated and unintended consequences for the Australian wine
sector which will need to be considered in the final strategy, and impacts will need to
be addressed and managed. This submission is only concerned with the policy and
strategy implications for the Australian wine sector and, therefore, on the general
approaches and alcohol specific aspects; it does not relate to the tobacco- or obesity-
specific elements of the discussion paper.

It needs to be stated succinctly in any resultant recommendation and strategies, that
while alcohol and tobacco are similarly categorised as licit drugs, this is where the
similarity ends. The fundamental difference between alcohol and tobacco is that all
tobacco consumption is harmful to human health and society. Alcohol, and
particularly wine, when consumed responsibly in moderation, is not harmful to human
health and society, and in fact can be considered as part of a healthy diet and
lifestyle for Australians.

The discussion paper has identified several broad targets. One target identified is to
reduce the prevalence of harmful drinking for all Australians by 30%. This is an
ambitious, although unclear, target and more clarity needs to be given as to whether
the target’s aim is a 30% reduction of harmful drinking in Australia or to reduce the
number of Australians who drink in a harmful manner by 30%.




                                                                                     3
Another target identified in the Discussion Paper is to contribute to the ‘closing the
gap’ target for indigenous people, reducing the 17-year life expectancy gap between
indigenous and non-indigenous people. This target is highly appropriate for a
preventative health strategy.

The Australian Wine Sector

The Australian wine sector is one of Australia’s major success stories. Over recent
decades, the wine sector has enjoyed strong production and export growth,
substantial investment, high levels of innovation and increased employment,
particularly in regional communities. The wine sector has contributed positively to
the economy, through income and employment, as well as through multiplier effects
to other industries (including tourism and retail) and the balance of payments in major
export earning.

The number of wineries has more than doubled in 10 years, to 2,300 in 2008. The
vast majority of wineries are small and located in regional areas, and around 70%
have an annual crush of less than 100 tonnes. Around 1,000 wineries have been
added since 2001, and 700 of these have a crush of less than 100 tonnes. There are
over 350 ‘medium’ wineries that crush between 100 tonnes and 20,000 tonnes, with
20 wineries crushing over 20,000 tonnes. Australia’s 5 largest wine businesses each
crush more than 100,000 tonnes annually and accounted for 49% of the estimated
2008 crush.

This demonstrates the size, variation and complexity of the Australian wine sector,
the importance to regional Australian, and the necessity to consider potential policy
and regulatory impacts across the sector.

There are significant ongoing profitability and viability pressures for Australia’s
wineries. Detrimental policy or regulatory impacts on the Australian wine sector
would have disastrous regional, employment and tourism impacts.

Consumption Trends

In Technical Report No. 3 “Preventing Alcohol-related harm in Australia: a window of
opportunity”, the Taskforce incorrectly refers to the prevention paradox to suggest
that interventions that target the majority of consumers rather than high-risk users will
result in preventing more (net) harm.

Subsequent re-analyses of the data supporting the preventative paradox theory have
shown that when patterns of alcohol consumption, especially episodes of heavy
consumption and intoxication are analysed rather than just mean consumption levels, the
preventative paradox disappears. Indeed, it has been shown that specific occasions of
heavy consumption, usually by people who generally consume moderate amounts of
alcohol in low risk ways, result in most alcohol-related harm (Single and Rohl 1997).
Accordingly, drinking pattern is often a better predictor of alcohol-related harm than just the
amount consumed (Rehm et al. 2001a). Correspondingly, a greater reduction in harm may
be achieved through the prevention of heavy high-risk consumption occasions rather than
by a reduction in the mean level of consumption.

The Australian wine sector believes that decisions on policy and regulation to
address the issue of risky consumption of alcohol must be based on sound scientific
evidence of the effectiveness of the policy. Additionally, strategies should be
implemented to target those groups that are engaging in risky consumption with a
view to changing that behaviour. The sector also supports population-based


                                                                                         4
programs that educate consumers about the health implications of drinking at risky
levels and the benefits of responsible consumption to bring about long-term,
sustained cultural change. Bringing about cultural change is a primary objective of
DrinkWise Australia, an organisation supported by the Australian wine sector.

The discussion paper identifies a mix of population and individual based strategies.
The individual-based measures include screening and effective brief health care
interventions and targeted programs for those groups at risk. The population-based
measures aim to (1) reshape consumer demand towards low-risk drinking using
social marketing and public education campaigns and (2) reshape supply towards
lower-risk products using taxation and advertising restrictions. These measures incur
individual costs while providing little individual benefit to those who consume alcohol
responsibly.

There is little evidence to demonstrate a net population benefit via increased
taxation. Responsible consumers are likely to reduce or cease consumption
depending on the level of price increase involved – leading to a decrease in
associated benefits from moderate intake. Irresponsible consumers are considered
less price sensitive and more likely to either switch products or switch substances.
This is an oversimplification of a complex set of relationships between consumers
and price and ‘alcohol’. There are in fact many different products (beer, wine, pre-
mixed spirits, spirits and others), different sub-sets of each of these, a multitude of
brands and price points and different locations for consumption. Another influencing
factor on behaviour is found in the age of the consumer and different socio-economic
status. There should be sound evidence of the effectiveness of a population-based
measure prior to inclusion in the strategy.

While it can be demonstrated that the mean consumption of alcohol correlates with the
prevalence of excessive alcohol consumption in a population (Rose et al. 1990), a
reduction in total alcohol consumption results from some and not necessarily all
individuals in a population consuming less. Unless the effect of the policy on the
consumption by alcohol abusers is known specifically, such a policy can be regarded as
a social experiment rather than a scientific prescription (Duffy 1977). Indeed when
studies have evaluated whether prices have a differential effect on light, moderate and
heavy alcohol consumption, the results suggest that both light and heavy alcohol
consumption are much less price elastic than moderate consumption (Manning et al.
1995).

In 2004 neither cask nor bottled wine was a significant product of choice for young
and underage consumers 12-17 years. Bottled wine became more popular with
females as they matured, although other beverages were still the preferred product
for the 21-24 year age group.

Wine decreased as a beverage of choice for under age consumers in the 15 to 17
year old age group between 2000 and 2004.

It is important to note that overall levels of underage consumption, and risky
consumption by younger people, have remained largely unchanged, although any
risky consumption is a cause for concern and strategies to address this consumption
should be pursued.

Wine as a product in general is consumed differently to other alcohol beverages: it is
consumed with food, in moderation, by older consumers at home or in restaurants.
Wine is consumed irresponsibly by some and the Australian wine sector does have a
role to play in addressing and promoting social responsibility.


                                                                                      5
The Australian Wine Sector and Alcohol Abuse

•   The Australian wine sector accepts that it has a responsibility to produce,
    promote and sell its product in a responsible manner that does not encourage
    excessive or irresponsible consumption. The wine sector thus understands and
    supports a mix of individual and population-based measures to ensure that
    alcohol is consumed responsibly, but as stated previously believe that the
    measures should be based on scientific evidence of their effectiveness. In
    addition, population-based measures should be confined to those that do not
    incur an unnecessary cost and/or impost on responsible consumers.

•   It is important to note that the taxation regime that generates the socially optimal
    level of consumption will not eliminate all negative alcohol impacts on society.
    Those negatives cannot be resolved through general taxation measures without
    causing a substantial welfare loss to all alcohol consumers, the overwhelming
    majority of whom are responsible consumers. Consequently more targeted, issue
    specific, policy instruments are the efficient and equitable solution.

•   While the WFA acknowledges that there is some evidence to suggest that
    warning labels may have an impact on raising awareness among consumers,
    there is clear evidence that warning labels are not effective in changing
    behaviour, particularly for at-risk groups.

•   It is also important to note that there are a number of inquiries currently underway
    in regards to warning labels on alcohol products, including the Council of
    Australian Governments (COAG) initiated Food Standards Australia New Zealand
    (FSANZ) review of mandatory health warning labels and FSANZ consideration of
    application A576 for pregnancy health advisory labels.

•   The WFA supports the current ABAC Scheme and its continuous improvement
    which provides a robust pre-vetting and complaints process for alcohol
    advertising and will continue to work as part of the Management Committee to
    strengthen the scheme at every available opportunity.

•   The Australian wine sector is a founding partner of DrinkWise Australia, an
    evidence-based organisation which is focused on promoting change towards a
    more responsible drinking culture in Australia. A key component of this aim is to
    minimise potential harms and maximise any benefits from alcohol consumption.

•   The Australian wine sector established The National Wine Foundation (NWF) in
    2001 with funds raised from the sale of the special Centenary of Federation Red
    Wine. All aspects of the Centenary wine – from the grapes, to the bottles, corks
    and other packaging and the expertise of Australia’s pre-eminent winemakers
    who created the wine – were donated by the Australian wine sector.

•   The funds are used to provide grants to targeted projects which promote positive
    social or economic outcomes, particularly in remote, regional or indigenous
    communities.

Collins and Lapsley Papers

Much of the recent discussion around health cost from alcohol abuse, and the
commentary provided in the Preventative Health Taskforce Discussion Paper, are
based on 2 papers by David J. Collins and Helen M. Lapsley (Collins and Lapsley


                                                                                        6
2008a and Collins and Lapsley 2008b). These two papers identify Social Costs of
Alcohol Abuse of $15.3 billion in Australia, and the potential to reduce these costs
over time through policy interventions as being approximately half of the total costs.

Independent analysis by Access Economics, highly regarded world experts in cost of
illness modelling and health evaluation, reveals fundamental and comprehensive
flaws in the methodological approach of both reports, and major questions over the
validity of conclusions and usefulness of the reports in any policy context. Both
reports by Access Economics have been submitted as Attachments to this
submission.




                                                                                         7
2. Introduction

This submission has been prepared by the Winemakers’ Federation of Australia
(WFA) on behalf of the Australian wine sector and in consultation with Creina
Stockley of The Australian Wine Research Institute (AWRI), who acted as an
independent scientific expert.

Established in 1990, WFA is the peak national body for the wine sector, representing
Australia's wineries on all national and international issues. WFA operates by
voluntary membership with specific representation for small, medium and large
producers. Current WFA membership collectively accounts for over 90% of wine
production in Australia and through affiliate state associations, more than 80% of
winemakers.

WFA provides policy leadership on the issues of business climate, social
responsibility, environmental sustainability, innovation, trade, and benchmarking that
are critical to the enduring viability of Australia's wineries. WFA also works closely
with Wine Grape Growers’ Australia (WGGA) on issues of common concern and
interest.

Throughout this paper, reference to the “Australia wine sector” is broadly defined to
be inclusive of all aspects of the supply chain and production of wine.




                                                                                        8
3. The National Preventative Health Taskforce and Strategy
The Australian wine sector welcomes the opportunity to provide input into the
development of the National Preventative Health Strategy and looks forward to
working with the Taskforce through this submission process, and the ongoing
development of the strategy, as outlined in the Terms of Reference:

       “The taskforce will use a multidisciplinary approach, operate in a
       collaborative, open and consultative manner, and work in partnership with
       existing agencies and bodies working in associated areas.”

The Australian wine sector, and in particular WFA, considers that the issue of alcohol
misuse and its impacts on society are complex. Compared with some other
stakeholders, the Australian wine sector as part of the supply and sale of an alcoholic
beverage has different challenges and considerations, but brings unique drivers and
opportunities to positively influence and reinforce the responsible consumption of
wine.

In particular, the Australian wine sector welcomes the emphasis by the Taskforce on
“reshaping attitudes and behaviours, rather than prohibiting them.”

3.1 Australia: The Healthiest Country by 2020

The discussion paper prepared by the National Preventative Health Taskforce,
Australia: the Healthiest Country by 2020 outlines the “case for reform in our
approach to the prevention of illness and promotion of health”. It acknowledges that
major changes are required in the way we behave as individuals, families,
communities, industries, states and as a nation. It further states that the answer does
not lie in short-term projects, but through urgent, comprehensive and sustained
action. The purpose of the discussion paper is to “test overall targets and the initial
recommendations…to inform and provoke discussion and debate between
Australians about how these targets can be achieved.”

The Australian wine sector’s submission is mindful of the Taskforce’s focus on, and
strategies for, reducing harmful alcohol consumption rather than all alcohol
consumption. These strategies will impact on wine consumers, the majority of whom
generally consume responsibly, as well as on the minority who consume
irresponsibly. Consequently, there will be an impact on wine producers. Indeed,
there will be unanticipated and unintended consequences for the Australian wine
sector which will need to be considered in the final strategy, and impacts will need to
be addressed and managed. This submission is only concerned with the policy and
strategy implications for the Australian wine sector and, therefore, on the general
approaches and alcohol specific aspects; it does not relate to the tobacco- or obesity-
specific elements of the discussion paper.

It needs to be stated succinctly in any resultant recommendation and strategies, that
while alcohol and tobacco are similarly categorised as licit drugs, this is where the
similarity ends. The fundamental difference between alcohol and tobacco is that all
tobacco consumption is harmful to human health and society. Alcohol, and
particularly wine, when consumed responsibly in moderation, is not harmful to human
health and society, and in fact can be considered as part of a healthy diet and
lifestyle for Australians.




                                                                                      9
Even the recent research manuscript published by McIntosh (2008) concedes that
despite consumer classification errors, the beneficial health effects of light to
moderate alcohol consumption are still significant. These beneficial health effects
include a reduction in the risk and incidence of cardiovascular disease, where
cardiovascular disease is the leading cause of death in the developed world,
accounting for 25−50% of all deaths. Its incidence also is increasing in developing
countries, associated with economic growth and development (Singh et al. 1996,
WHO 2002).

Accordingly, recommendations and strategies to reduce tobacco use will not
necessarily be relevant for alcohol consumption.

3.2 Targets

The discussion paper has identified several broad targets. One target identified is to
reduce the prevalence of harmful drinking for all Australians by 30%. This is an
ambitious, although unclear, target and more clarity needs to be given as to whether
the target’s aim is a 30% reduction of harmful drinking in Australia or to reduce the
number of Australians who drink in a harmful manner by 30%.

In Australia in 2004, approximately 50% of Australians were consuming at least one
alcoholic beverage per week and 9% drank daily (AIHW 2005). The majority of
consumers, however, drank on only one or two days per week (Stanford 1999).

Another target identified in the Discussion Paper is to contribute to the ‘closing the
gap’ target for indigenous people, reducing the 17-year life expectancy gap between
indigenous and non-indigenous people. This target is highly appropriate for a
preventative health strategy. It is widely accepted that indigenous people are more
likely to abstain from alcohol than non-indigenous people but, of those indigenous
people who do consume alcohol, the likelihood of abusive consumption is
significantly higher. The population of indigenous people who abuse alcohol,
therefore, require sustained and specialised strategies to address this target.
Population-based measures will almost certainly be ineffective given the commonly
multiple and compounding factors of location, socio-economic status, health,
education, employment and cultural impacts faced by indigenous people.




                                                                                      10
4. Overview of the Australian wine sector
4.1 Background and Importance to Australia

The Australian wine sector is one of Australia’s major success stories. Over recent
decades, the wine sector has enjoyed strong production and export growth,
substantial investment, high levels of innovation and increased employment,
particularly in regional communities. The wine sector has contributed positively to
the economy, through income and employment, as well as through multiplier effects
to other industries (including tourism and retail) and the balance of payments in major
export earning.

ABS Census data indicate that direct employment in the grape growing and wine
making industries was close to 30,000 in 2006. In addition, once allied supply
industries are taken into account, the number of jobs supported by the wine sector is
estimated at around 57,000 people for 2006.

Table 1: Snapshot of the Australian Wine Sector

                  1
Wineries (2008)                                                          number          2,300
  With Cellar Doors                                                      number          1,625
                          2
Direct Employment (2006)
  Grape Growing                                                          number         11,000
  Wine Making                                                            number         16,955
                          3
Wine Grape Crush (2008)                                                '000 tonnes       1,832
                      3
Wine Production (2008)                                                 million litres    1,227
                                  4
Domestic Sales - Volume (2007-08)                                      million litres     428
                              4
Domestic Sales - Value (2007-08)                                        $A million       2,680
                      5
Exports - Volume (2007-08)                                             million litres     709
                  5
Exports - Value (2007-08)                                               $A million       2,682
                      4
Imports - Volume (2007-08)                                             million litres      53
                 4
Imports - Value (2007-08)                                               $A million        432
Taxation (2008)
Net Wine Equalisation Tax                                               $A million        661
  Goods and Services Tax                                                $A million        369
Sources:
1 Australian and New Zealand Wine Industry Directory 2008
2 2006 Census of Population and Housing

3. ABS Catalogue No: 1329.0 Australian Grape Crush & Wine Production
4 ABS Cat No. 8504.0 Sales of Australian Wine & Brandy by Winemakers
5 AWBC WINEFACTS Statistics




                                                                                                 11
4.2 Rapid Growth
Export volumes have increased by over 225% in the past decade, while the export
value has increased by 170%. Meanwhile, domestic sales of Australian wine have
grown 22% by volume over the same period.
Chart 1: Domestic Sales of Australian Wine and Wine Exports


                       900                                                                                                                                                 3300

                       800                                                                                                                                                 3000
                                                                                                                                                                           2700
                       700
                                                                                                                                                                           2400
                       600
                                                                                                                                                                           2100
      Million Litres




                                                                                                                                                                                  $ Million
                       500                                                                                                                                                 1800
                       400                                                                                                                                                 1500
                                                                                                                                                                           1200
                       300
                                                                                                                                                                           900
                       200
                                                                                                                                                                           600
                       100                                                                                                                                                 300
                         0                                                                                                                                                 0
                             1994-95


                                       1995-96


                                                 1996-97


                                                           1997-98


                                                                     1998-99


                                                                               1999-00


                                                                                         2000-01


                                                                                                   2001-02


                                                                                                               2002-03


                                                                                                                         2003-04


                                                                                                                                   2004-05


                                                                                                                                             2005-06


                                                                                                                                                       2006-07


                                                                                                                                                                 2007-08
                                                 Domestic Sales Volume                                       Export Volume

                                                 Domestic Sales Value (Real)                                 Export Value (Real)




Note: Domestic Sales and Export Sales values are 2007-08 real prices. Domestic Sales value WFA estimate
Sources: ABS Catalogue No: 8504.0 Sales of Australian Wine & Brandy by Winemakers & Catalogue No: 1329.0
Australian Wine & Grape Industry, AWBC Wine Export Approval Report via WINEFACTS Statistics


4.3 Regional Impacts
The strong growth seen in the wine sector has been particularly important for
regional communities in Australia. This is indicated by the following trends:

The number of wineries has more than doubled in 10 years, to 2,300 in 2008. The
vast majority of wineries are small and located in regional areas, and around 70%
have an annual crush of less than 100 tonnes. Around 1,000 wineries have been
added since 2001, and 700 of these have a crush of less than 100 tonnes. There are
over 350 ‘medium’ wineries that crush between 100 tonnes and 20,000 tonnes, with
20 wineries crushing over 20,000 tonnes. Australia’s 5 largest wine businesses each
crush more than 100,000 tonnes annually and accounted for 49% of the estimated
2008 crush.

Employment in grape growing and wine manufacturing rose by 60% between 1991
and 1996, yet in the subsequent five year period to 2001, it almost doubled again,
from around 15,750 to close to 30,000, maintaining this in 2006.

There has also been a strong rise in regional investment in infrastructure. Around
$1.8B was invested in new vineyard development in the 1990s and $1.1B in the first
7 years of this decade. There was around $1.6B of capital investment in winery
infrastructure between 2001 and 2005.



                                                                                                                                                                                              12
Wine tourism has also seen healthy growth. Between 2000 and 2006, the number of
domestic overnight wine visitors increased at an annual average rate of 6%, while
domestic and international day visitors increased 5% and 8% respectively. In 2006
there were approximately 5.6 million international and local visitors to Australian
wineries. During the June and September quarters of 2006, $335 million of wine was
purchased, resulting in an estimated spend of $666 million by winery visitors in 2006.
(Tourism Australia).

This demonstrates the size, variation and complexity of the Australian wine sector,
the importance to regional Australian, and the necessity to consider potential policy
and regulatory impacts across the sector. A detrimental policy or regulatory impact
on the Australian wine sector would have disastrous regional, employment and
tourism impacts.

4.4 Winery Viability

An indication of the profitability of Australian wineries is provided by the ABS
Manufacturing Survey. Since 2001 the profit margin of wineries ranged between
minus 11% in 2002-03 and positive 19% in 2003-04. Over the same period the
average profit margin for total manufacturing enterprises varied between 6.5% and
8.1%. Between 2003-04 and 2004-05 the profit margin for wineries declined 12.1
percentage points.

Chart 3: Profit Margin % - 2001-02 to 2006-07




Source: ABS Catalogue No: 8221.0 Manufacturing Industry 2001-02 to 2006-07
Note: Profit margin percentage = operating profit before tax as a percentage of sales and service income.


The Australian wine sector derives a further indication of the financial performance of
wineries from the annual Deloitte Financial Benchmarking Survey.

The following table shows the proportion of loss making wineries, expressed as a
proportion of respondents in the Deloitte survey, in each size category.




                                                                                                            13
Table 2: Proportion of Loss Making Wineries by Size Category, 2003 to 2007
         (as a proportion of respondents in the Deloitte annual survey)
Winery Size                     2003                2004               2005              2006   2007
$0m-$1m                         40%                 33%                12%               63%    47%
$1m-$5m                         53%                 40%                32%               38%    55%
$5m-$10m                        17%                 44%                23%               17%    25%
$10m-$20m                       38%                 40%                38%                0%    25%
$20m+                           22%                 43%                25%               25%    20%
Source: Deloitte Annual Financial Benchmarking Survey for the Australian Wine Industry


In 2007, of those wineries in the $1-5M category that participated in the Deloitte
survey, 55% reported that they had recorded a loss, up by 17 percentage points.

Factors that have been driving this fall in margins and profitability include increasing
domestic and international competition and the highly fluctuating Australian dollar.
High and generally rising domestic taxation was also a major impost until the
introduction of the Wine Equalisation Tax Federal Producer Rebate in 2004 which
was extended in 2006, although further reform of wine tax is still a major issue. The
capital intensive nature of winery businesses, combined with strong production
growth in recent years, has resulted in a high dependence on debt to fund growth,
particularly for wineries whose revenue falls between $0-$1M and $1M-$5M. For
these companies the Deloitte survey shows bank debt has increased in 2007.

High gearing and lower margins, particularly when combined with poor cash flow, can
have a crippling effect on business, with employment and investment plans shelved
to meet debt commitments. Unfortunately, these factors are likely to constrain future
growth in the wine sector.

4.5 Today’s Challenges
The Australian wine sector faces a number of interlinked challenges and issues that it
will need to overcome, if it is to maintain strong growth and continue to increase the
substantial contribution it makes to the Australian economy and society in general.
These challenges and issues, as identified in Wine Australia: Directions to 2025,
include:

    •    a structural imbalance between the cost of production and the price
         opportunity;
    •    grape and wine supply and demand fluctuations;
    •    retail consolidation driving downward pressure on pricing and margin;
    •    slow domestic growth and a tougher market for export growth;
    •    a resurgent Old World and better resourced New World competitors;
    •    greater environmental and sustainability challenges and responsibilities,
         particularly exacerbated by climate change;
    •    changing social concerns; and
    •    changing demographics and consumer expectations.

In addition to the above, water allocations and the cost and timing of water
purchasing decisions are major challenges for irrigators.




                                                                                                  14
For the future of the Australian wine sector, it is vital that the industry sustains the
confidence and support of the broader Australian community. In order to achieve
this, it is critical to ensure that the wine sector continues to embrace a socially
responsible approach towards the production, promotion and sale of wine. It is
equally important that this approach is recognised by politicians and regulators and to
ensure that calls for increases on taxation, further restriction on advertising, or
warning messages that are not justified by any scientific evidence of delivering a
positive outcome for the community, are rejected.

All of these ongoing profitability and viability pressures for Australia’s wineries
demonstrate the continuing challenges for the Australian wine sector. Additional
detrimental policy or regulatory impacts will have a disastrous impact for hundreds of
regional businesses and the local communities they support.




                                                                                     15
5. Wine and Alcohol consumption – Trends
The 2007 National Drug Strategy Household Survey reports on alcohol consumption
patterns of Australians aged 14 years or older. The following broad conclusions can
be drawn from the data collected.

•   The proportion of the population consuming alcohol fell between 2004 and 2007
    but consumption patterns remain stable;
•   Males are twice as likely as females to consume alcohol daily with the proportion
    doing so trending down;
•   Indigenous Australians are more likely to abstain from alcohol but those that do
    consume are more likely to do so at risky levels;
•   Underage groups consume at very low levels on a daily or weekly basis;
•   Between the ages of 14-29 is the stage of life when consumers are most likely to
    drink at risk of short and long term harm;
•   Wine is not the main beverage of choice for consumers until around the age of 30
    years;
•   Those aged between 14-19 drinking at levels considered high risk prefer bottled
    spirits and liqueurs;
•   Bottled wine is preferred over cask wine for all age groups, but there is a
    significant number of consumers over 55 who drink cask wine as well as bottled
    wine;
•   Around 91% of recent drinkers had taken steps to reduce alcohol consumption
    with those drinking at low risk levels being less likely to take steps to reduce their
    consumption than those drinking at risky levels;
•   53% of wine is consumed in the consumer’s own home. In total almost 90% of
    wine is consumed either in a private home or in a restaurant; and
•   The majority of wine consumption is combined with food - 90% of bottled wine
    and 93% of red cask wine.
The following is a more detailed analysis of alcohol consumption trends in Australia
and of the information contained in the Preventative Health Taskforce discussion
paper.

5.1 Consumption trends

In the discussion paper “Australia: the healthiest country by 2020” The Preventative
Health Taskforce has identified that while alcohol consumption patterns have not
changed noticeably over the past decade, there is an increased awareness of the
problems resulting from consuming alcohol at harmful levels.

In Technical Report No. 3 “Preventing Alcohol-related harm in Australia: a window of
opportunity”, the Taskforce incorrectly refers to the prevention paradox to suggest
that interventions that target the majority of consumers rather than high-risk users will
result in preventing more (net) harm.

Indeed, Rose (1985) who identified the prevention paradox states there are two
preventative strategies, individual (targets high-risk consumers) and population-
based. The individual strategy seeks to identify and treat those at risk of harm
relative to those identified as not being at risk, while the population-based strategy


                                                                                         16
seeks to control the causes of risk. Rose also pointed out that there is a prevention
paradox when using population-based strategies. He states that a preventative
measure that may bring benefit to the population is likely to incur costs, both social
and financial for the individual and may offer so little individual benefit to each
participant that they lack the motivation to support the change and may actively resist
it. This is the prevention paradox referred to by the Task Force.

The following table shows the advantages and disadvantages of both strategies
(Rose, 1985).

Prevention by the individual (high risk)-based strategy
Advantages                                  Disadvantages
Intervention appropriate to the individual  Difficulties & costs of screening/idenfication
Subject motivation                          Palliative & temporary - not radical
Physician/health professional motivation    Limited potential for individual and population
Cost-effective use of resources             Behaviourally inappropriate
Benefit:risk ratio favourable


Prevention by the population strategy
Advantages                                    Disadvantages
Radical                                       Small benefit to individual (prevention paradox)
Large potential for population                Poor motivation of subject
Behaviourally appropriate                     Poor motivation physician/health professional


Another disadvantage of a population-based strategy is the possibility of substantial
welfare loss to the majority of consumers.

Subsequent re-analyses of the data supporting the preventative paradox theory have
shown that when patterns of alcohol consumption, especially episodes of heavy
consumption and intoxication are analysed rather than just mean consumption levels, the
preventative paradox disappears. Indeed, it has been shown that specific occasions of
heavy consumption, usually by people who generally consume moderate amounts of
alcohol in low risk ways, result in most alcohol-related harm (Single and Rohl 1997). For
example, in 1997, acute conditions such as alcohol-related road injuries related to
occasional high-risk consumption accounted for 28% of all Australian alcohol-related
deaths while chronic conditions such as cirrhosis of the liver related to sustained high-risk
consumption accounted for 42% (Chikritzhs et al. 2001). Accordingly, drinking pattern is
often a better predictor of alcohol-related harm than just the amount consumed (Rehm et
al. 2001a). Correspondingly, a greater reduction in harm may be achieved through the
prevention of heavy high-risk consumption occasions rather than by a reduction in the
mean level of consumption.

The Australian wine sector believes that decisions on policy and regulation to
address the issue of risky consumption of alcohol must be based on sound scientific
evidence of the effectiveness of the policy. Additionally, strategies should be
implemented to target those groups that are engaging in risky consumption with a
view to changing that behaviour. The sector also supports population-based
programs that educate consumers about the health implications of drinking at risky
levels and the benefits of responsible consumption to bring about long-term,
sustained cultural change. Bringing about cultural change is a primary objective of
DrinkWise Australia (refer to section 6.4 of this submission).

The discussion paper identifies a mix of population and individual based strategies.
The individual-based measures include screening and effective brief health care


                                                                                              17
interventions and targeted programs for those groups at risk. The population-based
measures aim to (1) reshape consumer demand towards low-risk drinking using
social marketing and public education campaigns and (2) reshape supply towards
lower-risk products using taxation and advertising restrictions. These measures incur
individual costs while providing little individual benefit to those who consume alcohol
responsibly.

There is little evidence to demonstrate net population benefit particularly through
increased taxation: responsible consumers are likely to reduce or cease consumption
depending on the level of price increase involved – leading to a decrease in
associated benefits from moderate intake, and irresponsible consumers are
considered less price sensitive and more likely to either switch products or switch
substances. This is an oversimplification of a complex set of relationships between
consumers and price and ‘alcohol’. There are in fact many different products (beer,
wine, pre-mixed spirits, spirits and others), different sub-sets of each of these, a
multitude of brands and price points and different locations for consumption. Another
influencing factor on behaviour is found in the age of the consumer and different
socio-economic status. There should be sound evidence of the effectiveness of a
population-based measure prior to inclusion in the strategy.

Over the past two decades, numerous studies have been published which have
analysed whether price is a determinant of alcohol consumption (Cook 1982; Maynard
1988; Progue et al. 1989; Collins et al. 1991; Richardson et al. 1994). For example,
many of the studies considered the individual alcoholic beverages, that is, beer, wine
and spirits, separately and did not examine the possibilities for beverage substitution in
response to selective tax increases (Maynard 1988). Alcohol is a complex good
composed of different types and qualities. While certain consumers respond to price
increases by altering their total consumption, others vary their choice of type or quality.
Indeed significant reductions in sales have been observed in response to price increases
but these reductions were mitigated by significant substitutions between beverages types
of qualities (Gruenewald et al. 2006).

While it can be demonstrated that the mean consumption of alcohol correlates with the
prevalence of excessive alcohol consumption in a population (Rose et al. 1990), a
reduction in total alcohol consumption results from some and not necessarily all
individuals in a population consuming less. Unless the effect of the policy on the
consumption by alcohol abusers is known specifically, such a policy can be regarded as
a social experiment rather than a scientific prescription (Duffy 1977). Indeed when
studies have evaluated whether prices have a differential effect on light, moderate and
heavy alcohol consumption, the results suggest that both light and heavy alcohol
consumption are much less price elastic than moderate consumption (Manning et al.
1995).

The preliminary findings of the 2007 National Drug Strategy Household Survey
(NDSHS) provides an indication of the consumption patterns of different age groups
by gender and identifies at-risk consumers. The survey results show that the
proportion of the population consuming alcohol daily fell quite significantly between
2004 and 2007 from 8.9% to 8.1%. The report also found that between 1991 and
2007, for Australians aged 14 years or older, that while alcohol consumption patterns
remained largely unchanged the proportion of less-frequent consumers – drinking
less than weekly – increased from 1991 to 2007 and further, the proportion of the
population that has never had a full serve of alcohol also generally increased over
this period. (Table 1, Appendix A)




                                                                                    18
In 2007, males were almost twice as likely as females to drink daily although the
proportion doing so is trending down, 12.0% in 2004 to 10.8% in 2007. Overall the
consumption patterns for each gender follow the combined trends outlined above for
the population aged 14 years and above. (Table 2, Appendix A).

5.2 Short- and long-term harm from alcohol misuse

In Australia, the National Health & Medical Research Council’s Australian Alcohol
Guidelines: Health Risks and Benefits 2001 defines consumption in terms of risk, that
is, drinking at levels of low risk, risky (moderate) and high risk in the short or long-
term. Short- term refers to occasional drinking episodes that are confined into a
single day, where Short-term harms include injuries from violence, accidents, falls,
having unprotected sex and alcohol poisoning. Long-term refers to regular and
repeated daily drinking defined by the total number of standard drinks per week,
where long-term harms include brain damage and hypertension, haemorrhagic
stroke, certain cancers, cirrhosis of the liver and pancreatitis.

For males, the consumption of up to 28 standard drinks per week is considered ‘Low
risk’, 29 to 42 per week ‘Risky’, and 43 or more per week ‘High Risk’.

For females, the consumption of up to 14 standard drinks per week is considered
‘Low risk’, 15 to 28 per week ‘Risky’, and 29 or more per week ‘High Risk’.

Given the different policy responses and program settings in dealing with consumers
under the legal age of consumption compared with legal consumption, it would be
useful to have the data broken down into age groups of 12 – 15, 16 – 17, 18 – 19,
and so on.

The proportion of both males and females who put themselves at short-term risk at
least monthly or weekly in both the 14 – 19 and 20 – 29 age groups is relatively much
higher than other age groups. (Table 4, Appendix A).

The overall pattern of consumption for all age groups at risk of long-term harm
remained relatively stable between 2001 and 2007. This does not mean that a policy
or program response is not required, as these levels of consumption should be
targeted as part of a long-term cultural change program.

5.3 Young and under-age consumers

At the time of drafting this submission the detailed findings of the 2007 National Drug
Strategy Household Survey data were not available by product category, however,
the 2004 survey data indicates the following in terms of alcohol preferences by young
and underage consumers:




                                                                                     19
Chart 4: Types of alcohol consumed by young males (data from 2004 NDSHS)




Chart 5: Types of alcohol consumed by young females (data from 2004
NDSHS)




From the above two charts it can be seen that in 2004 neither cask nor bottled wine
was a significant product of choice for young and underage consumers 12-17 years.
Bottled wine became more popular with females as they matured, although other
beverages were still the preferred product for the 21-24 year age group.




                                                                                 20
Chart 6: Changes in types of alcohol consumed by risky drinkers aged 15 to
17 from 2001 to 2004 (data from 2000-2004 NAC)




Source: Young people and alcohol: the role of culture influences. National Centre for Education and Training on
Addiction. A Roche, P Bywood, J Borlagdan et al


Chart 6 shows that wine decreased as a beverage of choice for under age
consumers in the 15 to 17 year old age group between 2000 and 2004.

It is important to note that overall levels of underage consumption, and risky
consumption by younger people, have remained largely unchanged, although any
risky consumption is a cause for concern and strategies to address this consumption
should be pursued.

5.4 Consumption by product

The following chart again demonstrates that wine is not the main beverage of choice
for consumers between the ages of 14 and 29. This age group prefers spirits with
wine becoming the preferred beverage from the age of 30 and remaining so for all
further age groups.




                                                                                                                  21
Chart 7: Participation in Beer, Wine and Spirits Consumption by Age
                 80
                                                                Beer           Wine       Spirits
                 70

                 60

                 50
    Percentage




                 40

                 30

                 20

                 10

                  0
                      14-19   20-24   25-29   30-34   35-39   40-44    45-49    50-54   55-59   60-64   65-69   70-74   75-79     80+
                                                                        Age Group
a
 Proportions of consumers within each age group. Note that the proportions do not add up to a 100 for
a given age group because drinkers may consume multiple alcohol types. Source: NDSHS (2004), Provided
by X. Zhao, Monash University


Chart 8: Participation in Cask Wine, Bottled Wine and Fortified Wine
consumption by Age
                 70
                                                               CW              BW         FW

                 60


                 50
    Percentage




                 40


                 30


                 20


                 10


                 0
                      14-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79                             80+
                                                                       Age Group
a
 Proportions of consumers within each age group. Note that the proportions do not add up to a 100 for
a given age group because drinkers may consume multiple alcohol types. Source: NDSHS (2004). Provided
by X. Zhao, Monash University

Consumers of all ages prefer bottled wine however, cask wine gains in popularity
with the older age groups - 55 onwards, possibly as it is more price accessible and



                                                                                                                         22
convenient for retirees to be able to dispense single glass serves and keep the
remaining wine for future consumption after it has been opened.

Table 4: Type of alcohol usually consumed by recent drinkers over 14 years
old and long-term risk status

                                                       Long-term risk
Age group                                         Low Risk                       Risky or high risk
                                                        Males
14-19                    Pre-mixed spirits in a can (52.8%)          Regular strength beer (74.3%)
20-29                       Regular strength beer (65.8%)            Regular strength beer (78.6%)
30-39                       Regular strength beer (59.0%)            Regular strength beer (77.0%)
40+                                 Bottled Wine (54.3%)             Regular strength beer (61.5%)

                                                         Females
14-19                    Pre-mixed spirits in a can (64.2%)        Bottled spirits & liqueurs (84.9%)
20-29                                Bottled wine (58.8%)          Bottled spirits & liqueurs (67.6%)
30-39                                Bottled wine (68.9%)                       Bottled wine (69.7%)
40+                                  Bottled wine (69.9%)                       Bottled wine (72.2%)

Notes
1. Base is recent drinkers
2. Respondents could select more than one response
Source: 2007 National Drug Strategy Household Survey


The 2007 National Drug Strategy Household Survey shows the most recent figures
on preferred alcohol type by age group. Bottled wine is preferred by males over 40
who drink at low-risk of long-term harm, while for those in this age group who
consume at high-risk levels regular strength beer is preferred. Females drinking at
low and high-risk levels prefer wine at a younger age than males with those in the 20-
29 age bracket drinking at high-risk levels preferring bottled spirits and liqueurs.




                                                                                                  23
5.5 Location of consumption
Chart 9: Wine consumers by age group and place of consumption




Source: Observations on alcohol beverage consumption: Australian demographics and consumption characteristics,
1999

The preferred location for the consumption of wine is overwhelmingly the consumer’s
own home or the home of a friend or relative, followed by restaurants. In the 18-24
age group, although wine is not their preferred product, of those who do consume
wine it is predominantly consumed in the home of a friend or relative. Of those age
groups (30+) for whom wine is the preferred beverage, over 60% of wine is
consumed in a private home, rising to over 70% by the age of 65 years.

Between the ages of 40-64, the years when the highest proportion of consumers
nominate wine as their beverage of choice, around 55% of consumers preferred
bottled wine, and within those age groups the wine is predominantly consumed at
home (53% +) or in the home of a friend or relative (approx. 10%)




                                                                                                            24
Chart 10: Wine, beer and spirit consumers by place of wine consumption




Source: Observations on alcohol beverage consumption: Australian demographics and consumption characteristics,
1999

The following table summarises the locations where wine, beer and spirits are
consumed.

Table 5: Summary of major places of consumption for alcohol beverages
                             Share of                                Share of                                Share of
                             drinking                                drinking                                drinking
Wine                        locations   Beer                        locations   Spirits and Liqueurs        locations
Own-home                       53%      Own-home                       45%      Own-home                       46%
At friends' or relatives'               Licensed premises                       Licensed premises
home                          19%       (exc. Restaurants)            27%       (exc. Restaurants)            29%
                                        At friends' or relatives'               At friends' or relatives'
Restaurants                   15%       home                          17%       home                          15%
Total*                        87%                                     89%                                     90%

*May not add due to rounding. Other alternatives not tabled.
Source: Observations on alcohol beverage consumption: Australian demographics and consumption characteristics,
1999


53% of wine is consumed in the consumer’s own home. In total almost 90% of wine
is consumed either in a private home or in a restaurant.




                                                                                                                    25
Chart 11: Wine consumers by wine style/container type consumed with/without
a meal




Source: Observations on alcohol beverage consumption: Australian demographics and consumption characteristics,
1999

Most wine consumers prefer to drink wine with food. 93% of red cask wine drinkers
and 90% of bottled wine drinkers combined wine with a meal.

Stockley (1998) states that concomitant food consumption significantly diminishes
the blood alcohol concentration achieved and the potential for harm to body cells,
organs and tissues, much of which is directly related to blood alcohol concentration.

This overall analysis of wine consumption clearly demonstrates alcohol should not be
considered as a single product: there are clearly significant differences in the ways,
manner, location and attitudes of consumers to different alcohol products across
different age-groups and different socio-economic groups. In particular, population-
based measures that attempt to address ‘alcohol misuse’ must consider significant
and varying impacts on various responsible consumers and overall potentially
detrimental societal outcomes.

Wine as a product in general is consumed differently to other alcohol beverages: it is
consumed with food, in moderation, by older consumers at home or in restaurants.
Wine is consumed irresponsibly by some and the Australian wine sector does have a
role to play in addressing and promoting social responsibility.




                                                                                                            26
5.6 Alcohol consumption in the indigenous community

Indigenous Australians are more likely to abstain completely from alcohol than non-
indigenous Australians but those who do consume alcohol are more likely to drink at
risky levels. 56% of indigenous Australians drink at a level of low risk compared to
74% of other Australians. 23% of indigenous Australians consume at high risk levels
compared to 10% of other Australians. The issue of alcohol consumption and
consumption of other substances in the rural indigenous communities and in urban
areas is a complex one requiring identification of cultural, socio economic and other
factors in order to tailor targeted programs as well as ensuring that the wider
population-based educational and social marketing measures are relevant to the
indigenous population.

Table 6: Characteristics by short- and long-term risk and indigenous status,
persons aged 14 years and older, Australia, 2004
                                                           Short-term risk        Long-term risk
                                                                    Risky or             Risky or
Characteristic             Abstainer/ex-drinker           Low risk high risk    Low risk high risk
All persons (aged 14+)             16.4                     62.9      20.7        73.7       9.9

Indigenous status
   Indigenous                       21.3                    40.0         40.0     56.0      22.7
   Other Australians                16.1                    63.3         63.3     74.1       9.7

Source: National Drug Strategy Household Survey Detailed Findings 2004

The Australian wine sector supports interventions and programs that assist
indigenous Australians to live in safe communities and improve their overall well
being, health and quality of life.

The Australian wine sector accepts that it has a responsibility to produce, promote
and sell its product in a responsible manner that does not encourage excessive or
irresponsible consumption. The wine sector thus understands and supports a mix of
individual and population-based measures to ensure that alcohol is consumed
responsibly, but as stated previously believe that the measures should be based on
scientific evidence of their effectiveness. In addition, population-based measures
should be confined to those that do not incur an unnecessary cost and/or impost on
responsible consumers.




                                                                                                   27
6. The Australian wine sector and alcohol policy
6.1 Taxation
According to the OECD1:

       “While the main purpose and the original reason for the introduction of excise
       duties (i.e taxing specific products)2 were to raise revenue, an alternative
       philosophy has emerged, recognising the application of excise duties as a
       means of influencing consumer behaviour”. Page 56

       “In general terms, beer wine and spirits are considered separate products
       within the category of alcoholic beverages.” Page 62

       “While the main characteristics and objectives ascribed to excise duties are
       approximately the same across OECD countries, their implementation,
       especially in respect to tax rates, sometimes gives rise to significant
       differences between countries.” Page 56.

       The data “in respect of excise duties on beer, wine and spirits illustrate the
       complicated computations for excise duties and show the current comparative
       rates for OECD countries.” Page 63.

       “For example excise duty on wine may vary from zero (Austria, the Czech
       Republic, Germany, Greece, Italy, Luxemburg, Portugal, the Slovak Republic,
       Spain and Switzerland) to USD 5.86 a litre (Iceland).” Page 56

These different levels of taxation reflect, in part, differences in the use of taxation to
pursue public health objectives and views about the differential contribution of harm
that can be attributed to the different forms of alcohol. All countries acknowledge a
degree of harmful alcohol consumption that imposes external costs on others. Taxes
on alcohol are an efficient and effective policy instrument to address those external
costs where they increase prices sufficiently to reflect social costs. The corrected
price will then ensure that overall alcohol consumption is at the socially optimal level.

It is important to note that the taxation regime that generates the socially optimal
level of consumption will not eliminate all negative alcohol impacts on society. Those
negatives cannot be resolved through general taxation measures without causing a
substantial welfare loss to all alcohol consumers, the overwhelming majority of whom
are responsible consumers. Consequently more targeted, issue specific, policy
instruments are the efficient and equitable solution.

Moreover taxation has not proven to be an effective policy to reduce alcohol abuse.

       “Perhaps the most compelling evidence against taxation as an effective policy
       measure against abuse comes from countries where taxation rates have
       traditionally been high. In many of these, such as the Nordic countries or
       those in Eastern Europe, alcohol consumption and harmful drinking patterns
       remain high.”3


1
  Organisation for Economic Co-operation and Development (OECD), Consumption Tax
Trends, 2006 edition
2
  author’s explanation
3
  International Center for Alcohol Policies (ICAP), ICAP Reports 18, May 2006


                                                                                        28
Focussing specifically on wine, ACIL Consulting concluded:

          “the case for a general health-related (public social cost) tax on wine is weak
          due to the favourable differences in wine consumption patterns and the
          recognised health benefits of moderate wine consumption.” 4

Finally, Access Economics in their review of the 2008 Collins and Lapsley report, The
costs of tobacco, alcohol and illicit drug abuse to Australian society in 2004/2005,
pointed out that “Alcohol taxes thus pay more than the social costs of alcohol abuse,
by a considerable margin, each year.”5

These arguments all add up to a compelling case against a general increase in
taxation on wine and strong support for specific policies that target the specific
behaviours, attitudes and groups of wine consumers who impair public health
outcomes.

6.2 Labelling

Under the current labelling requirements, all wine producers are required to state the
number of standard drinks contained in their bottle or packaging. Alcohol beverages
are also required to have an alcohol content statement included in the label.

In addition, wine producers are voluntarily adopting a standard drinks logo and
responsible consumption messages on package labels such as “Enjoy wine in
moderation” or “Drink Responsibly”.

While the WFA acknowledges that there is some evidence to suggest that warning
labels may have an impact on raising awareness among consumers, there is clear
evidence that warning labels are not effective in changing behaviour, particularly for
at-risk groups.

Given this, the WFA would only support some form of compulsory label information
with respect to health impacts of irresponsible consumption of alcohol as long as it
was introduced in conjunction with a comprehensive education and multi-media
communications campaign with significant Government funding.

It is also important to note that there are a number of inquiries currently underway in
regards to warning labels on alcohol products, including the Council of Australian
Governments (COAG) initiated Food Standards Australia New Zealand (FSANZ)
review of mandatory health warning labels and FSANZ consideration of application
A576 for pregnancy health advisory labels.

6.3 Advertising

Australia has a quasi-regulatory system for alcohol advertising. The centrepiece of
the system is the Alcohol Beverages Advertising Code (ABAC).

The ABAC Scheme aims to ensure that alcohol advertising will be conducted in a
manner which neither conflicts with, nor detracts from, the need for responsibility and
moderation in liquor merchandising and consumption and which does not encourage
consumption by underage persons.


4
    ACIL Consulting, Pathways to Profitability for Small and Medium Wineries, October 2002
5
    Access Economics, Collins and Lapsley report review, November 2008


                                                                                             29
The ABAC Scheme is administered by a management committee which is made up
of industry, advertising and government representatives. The operation of the ABAC
Scheme consists of two separate and independent elements: the pre-vetting process
and the complaints process. These two processes are, quite properly, carried out by
different people to ensure no conflict of interest issues arise between the pre-vetting
and the complaints processes, and both are separate from industry.

The issue of alcohol advertising has been the subject of a number of reviews over
the past few years, including a Victorian Parliamentary Inquiry, the New South Wales
Summit on Alcohol Abuse, the inquiry by the National Committee for the Review of
Alcohol Advertising (NCRAA), commissioned by the Ministerial Council on Drug
Strategy and the establishment of the Monitoring of Alcohol Advertising Committee.

These comprehensive reviews found that whilst improvements could be made, there
was general support for the ABAC Scheme.

In handing down its report in 2003, NCRAA stated that:

         “NCRAA is of the view that alcohol advertising should continue to be self-
         regulated in Australia and recommends that the MCDS work with the alcohol
         industry to enhance the existing self-regulatory system.” 6

NCRAA made a number of recommendations aimed at strengthening the then ABAC
Scheme. The ABAC Management Committee, in recognising changing community
expectations and the need to maintain the integrity and effectiveness of the scheme,
fully implemented all of the recommendations handed down by NCRAA.

These recommendations have resulted in a strengthened ABAC Scheme which has
seen a significant reduction in complaints against alcohol advertising over the last
few years. In 2007, the Advertising Standards Bureau received 2602 complaints
about advertisements across all product categories. Of these, only 2.44% related to
advertisements for alcohol products. This is in comparison to previous years:

Table 7: Complaints regarding alcohol advertisements on television as a
proportion of all television advertisement complains 2007

       Year       2004         2005        2006         2007
    Proportion   21.38%       7.07%       3.14%        2.44%
    of alcohol
    complaints

The WFA supports the current ABAC Scheme and its continuous improvement which
provides a robust pre-vetting and complaints process for alcohol advertising and will
continue to work as part of the Management Committee to strengthen the scheme at
every available opportunity.

6.4 DrinkWise Australia

The Australian wine sector is a founding partner of DrinkWise Australia, an evidence-
based organisation which is focused on promoting change towards a more


6
 Report to the Ministerial Council on Drug Strategy, Review of the Self-Regulatory System for
Alcohol Advertising August 2003.


                                                                                          30
responsible drinking culture in Australia. A key component of this aim is to minimise
potential harms and maximise any benefits from alcohol consumption.

Since its establishment in 2005, DrinkWise has been involved in a number of
community based projects including:

   •   Good Sports program. DrinkWise has sponsored the ADF’s Good Sports
       program for the past three years. The Good Sports program has over 200
       registered clubs in Australia and promotes a responsible attitude towards
       alcohol consumption;
   •   Be part of It – Not Out of It. DrinkWise has partnered with the NSW
       Department of Health to promote a sensible attitude towards alcohol
       consumption for young people.
   •   Sports Challenge and DrinkWise have formed a partnership to sponsor the
       community school and students in remote Timber Creek, Northern Territory.
   •   Macquarie University, with sponsorship from DrinkWise, is developing a mode
       that can be transported to other universities for the safe supply and
       management of alcohol on campus.

In addition, DrinkWise is involved in a number of research programs which are aimed
at gaining a better understanding of alcohol and its effects on society.

In June 2008, DrinkWise launched a major social change campaigned, Kids Absorb
Your Drinking, which is aimed at creating a long-term generational change towards a
more responsible drinking culture in Australia. The campaign focuses on
empowering parents to become positive role models for their children; so that the
next generation may believe ‘drinking to get drunk’ is socially unacceptable.

6.5 National Wine Foundation

The Australian wine sector established The National Wine Foundation (NWF) in 2001
with funds raised from the sale of the special Centenary of Federation Red Wine. All
aspects of the Centenary wine – from the grapes, to the bottles, corks and other
packaging and the expertise of Australia’s pre-eminent winemakers who created the
wine – were donated by the Australian wine sector.

The funds are used to provide grants to targeted projects which promote positive
social or economic outcomes, particularly in remote, regional or indigenous
communities.




                                                                                   31
7. Collins and Lapsley Papers
The Australian wine sector is supportive of efforts to develop strategies around
prevention of alcohol abuse provided they are based on sound scientific information
and evidence of efficacy.

Much of the recent discussion around health cost from alcohol abuse, and the
commentary provided in the Preventative Health Taskforce Discussion Paper, are
based on 2 papers by David J. Collins and Helen M. Lapsley (Collins and Lapsley
2008a and Collins and Lapsley 2008b). These two papers identify Social Costs of
Alcohol Abuse of $15.3 billion in Australia, and the potential to reduce these costs
over time through policy interventions as being approximately half of the total costs.

Independent analysis by Access Economics, highly regarded world experts in cost of
illness modelling and health evaluation, reveals fundamental and comprehensive
flaws in the methodological approach of both reports, and major questions over the
validity of conclusions and usefulness of the reports in any policy context. Both
reports by Access Economics have been submitted as Attachments to this
submission, with key elements highlighted in the sections below.

7.1 Social costs of alcohol abuse

The independent analysis by Access Economics on the Collins DJ, Lapsley HM
(2008a) The costs of tobacco, alcohol and illicit drug abuse to Australian Society in
2004/05 report reviews the methodology, assumptions and conclusions in the report.

The following conclusions were made by Access Economics:

Major conceptual and methodological problems

       “It is not clear that the report meets its fundamental aims, and indeed a
       second report (Collins and Lapsley, 2008b) has now been produced in order
       to address the first two objectives properly. One of the objectives not met –
       the cost of alcohol abuse that might be avoided – involves the concept of
       ‘avoidable costs’, which is only applicable for incidence-based costings or in
       forward-looking prevalence-based costings for those costs that are reversible.
       As such, any potential applications in relation to cost benefit analysis (CBA)
       or cost effectiveness analysis (CEA) of preventative health care initiatives to
       reduce abusive alcohol consumption need to take account of the limited
       reversibility of impacts, the cost of the interventions themselves and the
       relative efficacy of such measures. Emphasis should be placed on well
       targeted interventions (ie, those that actually address problem drinking) rather
       than blanket interventions that bluntly aim to reduce everyone’s consumption,
       whether or not the extent of consumption is indeed a problem. The value of
       Collins and Lapsley (2008a) is thus not clear.

       A fundamental flaw is that the report uses what it terms a ‘demographic
       approach’ to cost estimation. This uncommon terminology and approach
       means that the authors have attempted to estimate the impact of alcohol
       abuse retrospectively ie, the difference in 2004-05 if people in the 40 years
       prior had not abused. The counterfactual is thus impossible, since the past
       cannot be changed. The reason for undertaking a retrospective rather
       than forward looking approach is unclear, as is the policy usefulness of
       such an approach. The demographic approach used by Collins and



                                                                                     32
          Lapsley is likely to yield a much higher cost estimate than the human
          capital approach.” 7

          “Another conceptual problem is the definition of alcohol abuse, which is never
          clearly defined and definitions implicit in different cost elements vary
          throughout the report. The conceptual definition used by Collins and Lapsley
          is circular: ‘Abuse occurs if the community incurs net costs as a result of drug
          abuse’. The data on crime do not permit an estimate of alcohol abuse or
          intoxication, only alcohol consumption (which may not be abusive) close to
          the time of the crime. In other cost estimates it is unclear what the definition
          of abuse is; amounts and patterns of alcohol consumption are not mentioned
          in the report – eg, it is unclear where there might be concordance with the
          National Health and Medical Research Council (NHMRC) guideline
          definitions. Another report by the same authors with an inaccessible website
          reference was used to justify the assumption that 20% of alcohol consumption
          is addictive, and then an estimate of 30% was used to estimate abusive
          consumption, with no evidence or references provided for that parameter and
          sensitivity on it of an extremely high 50% only. Lack of attempts to validate
          this important parameter, potentially through primary data collection, is
          a serious flaw, as is the lack of lower sensitivity analysis.

          “In economics, private costs/benefits are borne by the individual
          producer/consumer and social costs/benefits are these plus ‘externalities’ ie,
          those costs/benefits that spill over to other entities. This concept is very well
          understood and used; however, Collins and Lapsley (2008a:8) define social
          costs as negative externalities (based on ‘Markandya and Pearce’ not
          provided in the references) and ignore private costs and private benefits.
          ‘Private’ is not defined – it is unclear if it is supposed to refer to individuals’
          cost or to firms’. There is also confused discussion of rational behaviour and
          access to information, leading to an extreme conclusion; Access Economics
          takes a ‘middle ground’ view between the rational addiction theorists and
          Collins and Lapsley, that there remains a non-zero consumption value to the
          abuser that is not limited to avoiding withdrawal effects. There is no
          mention in the report of the very large private benefits of responsible
          alcohol consumption in terms of utility value (choice, enjoyment). It is a
          major flaw to exclude private costs from total social costs and private
          benefits from total social benefits and it is incorrect to claim that costs
          borne by individuals are irrelevant to public decision-making. Access
          Economics recommends a matrix approach to the net cost of alcohol
          abuse – by cost type and bearer – with standard definitions and
          terminology so that cost elements are clearly and accurately presented,
          as per Table 2-1 on page 9.” 7

In conclusion, Access Economics states:

          “Overall the report is poorly structured with inadequate cross-referencing,
          there is a paucity of supporting evidence and references for claims, the
          referencing is incomplete and there is a very serious lack of transparency in
          calculation processes. Methodology and data used are singularly poor.

          There is apparent upward bias in most of the cost estimates and uncertainty
          is not dealt with due to grossly inadequate sensitivity analysis. Using a best
          possible approach would exclude one cost item, making the overall cost of

7
    Access Economics, Collins and Lapsley (2008a) report review, November 2008


                                                                                          33
          alcohol abuse at least 18% lower. The other findings of the report should be
          viewed with substantial scepticism.

          We agree with the final suggestion (in Collins and Lapsley’s Section 9.14) that
          the data needs are reviewed with an expert advisory panel convened, also to
          review methods, given the serious weaknesses identified in this critique. We
          would recommend that a time series of comparable estimates of the cost of
          alcohol abuse is calculated based on a sound methodology and the best
          available data.” 7


7.2 Avoidable costs of alcohol abuse

The independent analysis by Access Economics on the Collins DJ, Lapsley HM
(2008b) The avoidable costs of alcohol abuse in Australia and the potential benefits
of effective policies to reduce the social costs of alcohol report focused on three main
areas:
        1.      whether the estimate of social costs of alcohol abuse that can be
                reduced is in fact reasonable;
        2.      whether methods and data represent a correct approach to the
                primary research question; and
        3.      whether the intervention choices, conclusions and recommendations
                are appropriate in relation to the potential cost reductions (benefits)
                from a policy perspective.

The following conclusions were made by Access Economics:

Is the estimate of the social costs of alcohol abuse reasonable?

          “We concluded that the costing analysis displayed a paucity of supporting
          evidence, incomplete references for claims, lack of transparency in
          calculation processes, and singularly poor methodology and data use. There
          was apparent upward bias in most of the cost estimates and uncertainty was
          not dealt with due to inadequate sensitivity analysis. Using a best possible
          approach would exclude one cost item, making the overall cost of alcohol
          abuse at least 18% lower. The other findings of the report should be viewed
          with substantial scepticism.

          We agree with the final suggestion (Section 9.14 of Collins and Lapsley,
          2008a) that data needs are reviewed and an expert advisory panel convened,
          also to review methods, given the serious weaknesses identified.” 8

Are methods and data appropriate to the purpose?

          “The methods (‘feasible minimum’, ‘Arcadian normal’, ‘exposure-based
          comparators’) selected are not appropriate to the purpose. Conventional
          CBA and/or CEA would have been preferable, as recommended by the
          OBPR for example, for evaluating economic evidence in relation to proposed
          interventions. The methods and numbers selected to derive the conclusion
          that it would be possible to reduce the social costs of alcohol over time by
          about a half, have no valid evidence basis, in our view. Conceptually, by
          preventing all incident cases of alcohol abuse, over time once could eliminate
          all its social costs.” 8

8
    Access Economics, Collins and Lapsley (2008b) report review, December 2008


                                                                                      34
How were the interventions chosen?

       “Methods used to select so-called effective interventions are problematic and
       out of line with standard practice. Unreferenced claims are made that certain
       interventions are ‘world best practice’, ‘proven-effective’, or ‘very cost-
       effective’, without evidence. Important issues are raised such as the
       treatment of combination policies and the risk that some policies will reduce
       benefits as well as costs of consumption – yet these are ignored, thus
       overstating the potential benefits of the selected interventions.” 8

Are the intervention conclusions appropriate?

       “No confidence can be placed in any of the estimates regarding the supposed
       cost savings from interventions, given the numerous methodological flaws.” 8

In conclusion, Access Economics states that:

       “In our view, there are fundamental flaws in the methods used in this Collins
       and Lapsley (2008b) report and hence in its recommendations. There is also
       no disclaimer on the report that the findings do not represent the views of the
       Australian Government. We recommend the matter is reviewed by the
       Department of Health and Ageing and that genuine evidence only informs
       policy consideration of this important issue.” 8




                                                                                    35
        8. Discussion Paper – Specific Questions
        The Taskforce Discussion Paper Australia the Healthiest Country by 2030 raises a
        number of specific questions. These questions do not in many instances relate to the
        text in the preceding chapter, or cover all of the points raised in the preceding text.
        As such, the response to individual questions should not be seen to be endorsement
        or otherwise for the rest of the chapter. The WFA does attempt to provide specific
        answers to the questions raised in the Discussion paper.

        Questions – p. 41

             •   Do you support a focus on the suggested priorities?
             •   If you do not support these actions, or have other suggestions, what would
                 you propose we do as a nation to halt the toll of early deaths and disease
                 caused through alcohol-related harm?
             •   What are the most important issues that can engage support from individuals,
                 communities, industry and governments and drive cultural change?
             •   What prevention strategies work best for high-risk groups, particularly among
                 young people and in indigenous communities?


Priorities             Actions                               Support /    Comments
                                                             Reject
Reshape                Develop adequately funded and         Support      The wine sector supports long
consumer demand        effective social marketing and                     term social marketing campaigns
towards low-risk       public education campaigns to                      to effect cultural change in the
drinking               prevent misuse and reduce alcohol                  attitude of Australians to
                       related harm, including targeted                   drinking. This is evident in the
                       approaches and local                               sector’s contribution to
                       complementary initiatives for                      DrinkWise Australia. This item
                       different population groups                        should also include
                                                                          Commonwealth, States and
                                                                          Local Government involvement
                                                                          as well as the other identified
                                                                          groups.
Reshape supply         Review the taxation system to         Reject as    On its own, incentives to
towards lower-risk     stimulate the production and          presented    produce low-risk products are
products               consumption of low-alcohol                         not of concern. However, the
                       products                                           definition of ‘low-risk’ products is
                                                                          not provided nor is there good
                                                                          data as to products that are the
                                                                          main source of abuse. Low
                                                                          alcohol products are not
                                                                          necessarily consumed in a less
                                                                          risky fashion. Interestingly, the
                                                                          wine sector is producing more
                                                                          ‘light’ wines in response to
                                                                          consumer demand this trend is
                                                                          expected to continue.




                                                                                              36
Priorities          Actions                                 Support /    Comments
                                                            Reject
                    Develop a staged approach to            Reject       Alcohol advertising is already
                    restrict alcohol advertising                         controlled through a quasi-
                                                                         regulatory system in Australia
                                                                         that has been through a number
                                                                         of reviews in recent years.
                                                                         Alcohol advertising is already
                                                                         restricted.
                    Remove tax deductability for            Reject       Advertising is a legitimate part of
                    advertising and develop a staged                     business activities. The wine
                    approach to restrict alcohol                         sector already contributes some
                    advertising                                          $670 million per year in wine tax,
                                                                         as well as all other business
                                                                         taxes. To remove advertising as
                                                                         a legitimate aspect of selling
                                                                         wine – a legal product that, when
                                                                         consumed in moderation, is part
                                                                         of a healthy lifestyle – introduces
                                                                         a further layer of business
                                                                         expense for thousands of wine
                                                                         businesses across the country
                                                                         with no demonstration of
                                                                         associated public health
                                                                         benefits.
Improve public      Enforce legislation on responsible      Support in   The wine sector supports greater
safety              serving of alcohol, through             principle    enforcement of existing
                    intelligence led policing programs                   regulation and greater
                    that focus on the drinking                           coordination across jurisdictions.
                    establishment                                        Improvement of public safety
                                                                         should be a priority of
                                                                         coordinated action by all
                                                                         governments, and drinking
                                                                         establishments have a positive
                                                                         role to play in this outcome.
Close the gap for   Commission research on effective        Support      Targeted and local solutions for
disadvantaged       strategies to address social                         local communities should be a
communities         determinants of alcohol                              strong principle of strategies
                    consumption in Indigenous and                        endorsed by the Taskforce
                    low-income communities
Strengthen, skill   Develop and disseminate                 Support      The wine sector strongly
and support         information and training packages                    supports targeted and brief
primary health      (with a focus on screening, effective                intervention programs
care to help        brief interventions and appropriate
people in making    referral pathways for health and
healthy choices     welfare workers
Improve maternal                                            Support      The wine sector strongly
and child health                                                         supports targeted and brief
                                                                         interventions by medical
                                                                         professionals




                                                                                            37
Priorities             Actions                                 Support /    Comments
                                                               Reject
Build the evidence     Model safer patterns of alcohol         Priority     This action calls for experiments
base                   consumption in different                supported    in taxation policy without any
                       communities from changes to             but action   consideration of the potential
                       alcohol taxation arrangements and       Rejected     significant impacts on
                       understanding of the impact of                       businesses across the country,
                       different types of alcohol outlets                   including 2,300 winemakers and
                       and their density on hospitalisation,                7,500 grapegrowers. Increases
                       violence and crime rates                             in tax on wine will have major
                                                                            social and resource allocation
                                                                            impacts on communities reliant
                                                                            on the wine sector in more than
                                                                            60 regions across the country –
                                                                            without evidence that alcohol
                                                                            abuse will be positively
                                                                            impacted.
                       Collect and analyse nationally          Support in   Greater information should
                       consistent data about alcohol outlet    principle    contribute to more informed
                       density and alcohol-related health                   policy and regulatory decision
                       and safety outcomes with a view to                   making. Decisions regarding
                       regulate alcohol outlet density                      outlet density must be made on
                                                                            sound evidence.


        8.1 National Prevention Agency
        Questions p. 46

             •   Do you support the development of a National Prevention Agency to lead and
                 guide coordinated action for prevention?
             •   Is the suggested approach adequate? If not, or if you have other suggestions,
                 what else should be considered?
        The Discussion Paper recommends the establishment of a National Prevention
        Agency. This is an additional and unnecessary expense when the basic requirement
        has already been identified: clear, long-term strategies which are appropriately and
        substantially funded to enable successful implementation. There already exists an
        effective alcohol policy group and framework within the Federal Department of Health
        and Ageing, as well as the state agencies.

        The recommended Agency would add yet another layer of bureaucracy to the
        previously established Commonwealth of Australian Governments, Ministerial
        Council on Drug Strategy, Department of Health and Ageing and State Departments,
        National Health and Medical Research Council and ad-hoc groups such as the
        National Preventative Health Taskforce and Monitoring of Alcohol Advisory Council.

        8.2 Measurements
        Question – p. 52
             •   Are these measurements appropriate?
        Performance Indicator          Support /        Comments
                                       Reject



                                                                                              38
Performance Indicator         Support /       Comments
                              Reject
Tier 1                                        A general comment with respect to Tier
                                              1 performance indicators is their
                                              inability to identify specifics associated
                                              with different beverages, the manner
                                              and nature in which it is being
                                              consumed. As a general statement,
                                              however, these high level targets are
                                              relevant.
Deaths attributable to        Support in      In order to ensure comparability of
alcohol                       principle       data, consistent, long term reporting
                                              must be employed to ensure the same
                                              parameters are being analysed.
Hospital separations          Support in      Consistency in reporting frameworks
attributable to alcohol       principle       and datasets being compared is
                                              essential to ensure accurate
                                              assessment of progress
Tier 2                                        As a general statement on Tier 2
                                              performance indicators, it is essential to
                                              be able to conduct long-term
                                              comparison of base-line information,
                                              meaning the measurement and
                                              definition of terms such as “short-term
                                              harm” and “long-term harm” is
                                              maintained. For this reason, the
                                              definitions as provided in the NHMRC
                                              2001 Australian Alcohol Guidelines
                                              should be retained. Further, analysis of
                                              data should be emphasised around
                                              appropriate age groupings (i.e. 14 – 17
                                              year olds, 18 – 19, 20 – 21, etc).
Tier 3
Recall of public education    Support
and social marketing
campaigns
Taxation incentives for the   Support in      While the wine sector supports in
production and                principle but   principle the concept of taxation
consumption of low-           reject          incentives for low alcohol products,
alcohol products              increases in    such incentives should manifest as
                              taxation for    lower taxation rates for lower alcohol
                              alcohol         wine (such as a reduction in WET) or
                              beverages       research and development incentives,
                                              rather than penalties for all other
                                              products. It must also be recognised
                                              that ‘low alcohol products’ are not
                                              necessarily identifiable with lower levels
                                              of abusive consumption across different
                                              alcohol types.




                                                                                     39
Performance Indicator         Support /     Comments
                              Reject
Alcohol outlet density by     Reject        The wine sector is not aware of
city/town/region                            evidence that directly links outlet
                                            density with alcohol abuse. In some
                                            circumstances, regional viability might
                                            be heavily influenced by critical mass
                                            such as the number of cellar doors in a
                                            wine tourism region – and an increase
                                            in cellar doors does not have any
                                            correlation to alcohol abuse rates.
Legislation to restrict the   Reject –      Comprehensive restrictions on the
promotion of alcohol          duplication   promotion of alcohol already exist.
                                            Further legislative backing of existing
                                            restrictions could enhance the
                                            operation of the ABAC scheme but
                                            would also run the risk of devaluing the
                                            existing quasi-regulatory approach and
                                            detract from the existing scheme.
Systems and practices to      Support       To be developed and implemented in
proactively police licensed                 conjunction with venues and event
venues, events and harms                    organisers
Proportion of people at       Support       Definitions of short- and long-term harm
risk of short- or long-term                 must remain consistent
harm from alcohol
receiving brief
interventions
Expenditure on research       Support
and evaluation relating to
alcohol control for
indigenous communities
and other disadvantaged
populations
Enabling Infrastructure                     The wine sector supports the enabling
                                            infrastructure outlined in the Discussion
                                            Paper, with a strong emphasis on
                                            involvement of industry and the private
                                            sector in the design and evaluation of
                                            relevant programs and assessment of
                                            industry impacts from proposed policy
                                            or regulations.


8.3 Taxation

Taxation of alcohol is encompassed within the terms of reference of the Australian
Future Tax System (AFTS) Review announced on 13th May 2008 and scheduled for
completion by the end of 2009.

This tax review has been described by the Australian Government as “the most
comprehensive examination of the taxation system in over 50 years.”



                                                                                  40
The stated objective in the Terms of Reference is “to create a tax structure that will
position Australia to deal with the demographic, social, economic and environmental
challenges of the 21st century and enhance Australia’s economic and social
outcomes.” The desired outcome is for a tax system that will help boost Australia’s
productive capacity and build prosperity in an increasingly competitive global
environment.

The review encompasses all Australian Government and State taxes (except the rate
and base of the GST), Government transfer payments, tax expenditures, interactions
between federal, state and local government taxes and the interrelationship with the
proposed emissions trading system (ETS).

The guiding principles that will be used for assessing the merits of changing an
individual tax or the tax system are economic efficiency, equity, simplicity and
compliance costs. As stated in the Review terms of reference “Raising revenue
should be done so as to do least harm to economic efficiency, provide equity
(horizontal, vertical and inter- generational) and minimise complexity for taxpayers
and the community.”

The AFTS Review initial discussion paper titled Architecture of Australia’s tax and
transfer system (www.taxreview.treasury.gov.au) released on 6th August 2008
provides an overview of the principles applicable to the goal of achieving non-tax
policy objectives with indirect taxes.

The paper specifically acknowledges that “The tax system can also be used to
reinforce non-tax policies, including industry assistance, income redistribution, or
promoting public health.” page 279

The paper explains the justification for using tax measures to address externalities,
“A tax on a specific good may be beneficial if it reflects the external costs the use of
the good imposes on others. Such a tax ensures that users or producers of the good
take into account the negative effects they have on others when making production
or consumption decisions.” Page 279

However the paper goes on to state that “These taxes are not designed to eliminate
all negative consequences to society. All that is required is that the costs to society of
undertaking the activity are reflected in the economic decisions of consumers or
producers.” Page 280

Specifically addressing alcohol the paper notes, “In some cases, specific taxes apply
to consumption of products which may have negative health effects on the consumer
and others. For example, alcohol taxation is applied to intoxicating beverages…...”
“This means alcohol taxation is more closely targeting potential public health costs
and reinforcing other social policy objectives around alcohol consumption.” Page 281

The further point also is made “Taxation is not the only way to achieve corrective
action. For example, licensing restrictions on alcohol and product controls on tobacco
aim to affect consumer behaviour, as do public health campaigns. To the extent that
the externality is addressed through non-tax means, there is reduced need to apply a
corrective tax.” Page 280

In addition the paper acknowledges the impacts of such taxes, “Addressing non-tax
policy objectives with taxes on specific goods and services has complex effects. To
the degree that there is some substitutability between different forms of consumption,
non-uniform tax rates encourage the production and consumption of less taxed


                                                                                       41
goods. Moreover, there may be different impacts on different market segments.”
Page 281. “However, the taxation of specific products does have distributional
consequences.” page 281

The above provides clear evidence that alcohol taxation is within the ambit of
the AFTS Review and that it is being evaluated in the context of the entire tax
system. This is the most appropriate forum for review given that it offers the
best prospect for an alcohol tax outcome that delivers on the tax system
principles of efficiency, equity, and simplicity in addressing the Government’s
social, economic, and environmental objectives.




                                                                                 42
9. General Comments
The Australian wine sector is supportive of efforts to implement strategies around the
prevention of alcohol misuse in Australia. Indeed, the wine sector supports a
sustained, long-term effort towards changing the way Australians drink alcohol to a
more responsible drinking culture. This is the basis of wine sector investment in
DrinkWise Australia. The Discussion Paper encouragingly identifies an
“emphasis…on reshaping attitudes and behaviours, rather than prohibiting them” (p.
ix).

The Discussion Paper generally acknowledges that a collaborative rather than a cost
and control approach to addressing alcohol misuse is required: that is, cooperation
and collaboration between Governments, health groups and industry will deliver far
greater results than the prohibitionist, heavy regulation approach advocated by some
in the community. Indeed, there is a need to ensure existing regulation and agencies
are adequately enforced and resourced as identified “Equally important is the need
for strong enforcement of current laws and practices that prevent the harmful use of
alcohol” (p. 29).

There are some inaccuracies in the Discussion Paper and associated Technical
Report No 3 that relates to Alcohol that have been pointed out in this submission.
Page 5 of the Technical Report identifies that “For example, alcohol sales data, while
it is known to be collected and analysed by the alcohol beverage industry, is not
available for the purposes of this paper, nor is it easily accessed for public health
research purposes in general”.

The Winemakers’ Federation of Australia was not approached by anyone associated
with the development of the Discussion paper or Technical Report. Furthermore, any
data that is collected or distributed by WFA is available to any in the research field,
with appropriate references and subject to copyright laws. Most of the limited data
used by industry associations is commercially available through ABS, ACNielsen or
other collection agencies and is much more readily available to the significant
resources of Governments and Government Agencies that voluntarily funded industry
associations. Indeed, the Winemakers’ Federation of Australia supports calls for
Government funding to be maintained and enhanced for the collection and availability
of data.

Similarly, wine companies purchase commercial data from ABS, ACNielsen and
other data collection agencies which they are contractually prohibited from making
available to third parties. They also have proprietary brand information which of itself
is not relevant for public health purposes and is appropriately deemed commercially
in confidence. The WFA is not aware of any approach made to any alcohol company
or industry association for data as part of the compilation of the Discussion Paper or
Technical Report.




                                                                                     43
References


1. Access Economics, (November 2008), Review of the 2008 Collins and Lapsley
   report, The costs of tobacco, alcohol and illicit drug abuse to Australian society in
   2004/2005

2. Access Economics, (December 2008), Review of the 2008 Collins and Lapsley
   report The avoidable costs of alcohol abuse in Australia and the potential benefits
   of effective policies to reduce the social costs of alcohol

3. Aust Institute of Health and Welfare, National Drug Strategy Household Survey
   Detailed Findings 2004

4. Aust Institute of Health and Welfare, National Drug Strategy Household Survey
   First Findings 2007

5. Collins DJ, Lapsley HM, (1991), Estimating the economic costs of drug abuse in
   Australia. Monograph Series No. 15, Department of Community Services and
   Health

6. Cook PJ, (1982), Alcohol taxes as a public health measure. British Journal of
   Addiction, 7:245-250

7. Department of Health & Aging (2006), Aboriginal and Torres Strait Islander health
   performance framework 2006 report detailed analyses

8. Manyard A (1988), Price as a determinant of alcohol consumption. Australian
   Drug and Alcohol Review, 7:287-296

9. McIntosh J, (2008) Is alcohol consumption good for you? Results from the 2005
   Canadian Community Health Survey, Vol. 16, No 6, 2008

10. Progue TF, Sgontz LG, (1989) Taxing to control social costs: the case of alcohol.
    The American Economic Review

11. Richardson J, Crowley S, (1994) Optimum alcohol taxation: balancing
    consumption and external costs. Health Economics 3 (2) : 73-87

12. Rose G, (1985), Sick Individuals and Sick Populations, International Journal of
    Epidemiology, International Epidemiological Association 1985, Vol. 14, No. 1

13. Roche A, Bywood, P, Borlagdan , J et al, (2007), Young people and alcohol: the
    role of culture influences. National Centre for Education and Training on
    Addiction

14. Stanford L, (1999),Observations on alcohol beverage consumption: Australian
    demographics and consumption characteristics., Australian Wine & Brandy
    Corporation.

15. Stockley C, (1998) Submission to the Tax Consultative Task Force on Taxation
    Reform, WFA.




                                                                                      44
16. Winemakers’ Federation of Australia and Australian Wine and Brandy
    Corporation, (2007) Wine Australia: Direction to 2025, Australian Wine and
    Brandy Corporation




                                                                                 45
                                                      Appendix A


Table 1: Alcohol drinking status: proportion of the population aged 14 years or
older, Australia, 1991 to 2007
Drinking status                          1991         1993         1995           1998          2001       2004        2007
                                                                   (per cent)
Daily                                     10.2          8.5          8.8           8.5             8.3          8.9        8.1 #
Weekly                                    41.0         39.9         35.2          40.1          39.5        41.2        41.3
Less than weekly                          30.4         29.5         34.3          31.9          34.6        33.5        33.5
             (a)
Ex-drinker                                12.0          9.0          9.5          10.0             8.0          7.1        7.0
Never a full serve of alchol               6.5         13.0         12.2           9.4             9.6          9.3     10.1 #


(a) Has consumed at leaast a full serve of alcohol, but not in the previous 12 months.

# Difference between 2004 result and 2007 result is statistically significantly (2-tailed α = 0.05).




Table 2: Alcohol drinking status: proportion of the population aged 14 years or
older, by sex, Australia 2004, 2007
                                                   Males                          Females                        Persons
Drinking status                           2004          2007            2004             2007            2004         2007
                                                                                  (per cent)
Daily                                       12.0          10.8 #            5.8             5.5           8.9           8.1 #
Weekly                                      47.6           46.8            35.0             35.9          41.2          41.3
Less than weekly                            27.5           28.3            39.4             38.5          33.5          33.5
Ex-drinker(a)                                6.0            5.8             8.2             8.1           7.1              7.0
Never a full serve of alcohol           6.9           8.2 #           11.6           12.1           9.3         10.1 #
         (a)         has consumed at least a full service of alcohol, but not in the previous 12 months
         #          Difference between 2004 result and 2007 result is statistically significant (2 tailed – a=0.05)
         Sources: 2004 and 2007 National Drug Strategy Household Survey: First results




                                                                                                                                 46
Table 3: Alcohol drinking status: proportion of the youth population aged 12
years or older, by age and sex, Australia, 2004 and 2007
                                                                      Age group
    Drinking status                                12-15                     16-17                   18-19
                                                                      (per cent)
                                                                        Males
                                            2004           2007      2004            2007    2004        2007
    Daily                                    0.1            --        1.0            1.7      1.3            2.6
    Weekly                                   3.3           1.0        21.9           20.0     51.5       46.7
    Less than weekly                        28.3           28.8       53.2           50.9     32.8       40.9
                 (a)
    Ex-drinker                               3.3           2.7        2.9            5.2      1.4            1.5
    Never a full serve of alcohol           64.9           67.5       21.0           22.1     13.0           8.3
                                                                      Females
                                            2004           2007      2004     2007           2004        2007
    Daily                                     --           0.5         0.3            --       0.9        0.7
    Weekly                                   3.2           3.2        21.3           15.4     39.1       35.3
    Less than weekly                        29.8           26.8       57.2           63.2     48.8       51.9
                 (a)
    Ex-drinker                               2.2           2.1        3.5            3.0      1.2            1.9
    Never a full serve of alcohol           64.8           67.4       17.7           18.4     10.0       10.2
                                                                      Persons
                                            2004           2007      2004     2007           2004        2007
    Daily                                    0.1           0.2        0.6            0.8      1.1            1.6
    Weekly                                   3.3           2.1        21.6           17.3     45.4       41.1
    Less than weekly                        29.1           27.8       55.2           57.0     40.6       46.3
                 (a)
    Ex-drinker                               2.8           2.4        3.2            4.2      1.3            1.7
    Never a full serve of alcohol           64.8          67.5         19.4          20.3      11.6          9.2
       (a)        Has consumed at least a full serve of alcohol, but not in the previous 12 months.
                  Sources: 2004 and 2007 National Drug Strategy Household Survey: First results




                                                                                                                   47
Table 4: Short Term Risk - Alcohol consumption, risk of harm in the short
term: proportion of the population aged 14 years or older, by age and sex,
Australia, 2001, 2004 and 2007

                                                                                                         (b)
                                                                                        Level of Risk
                             (a)
Age             Abstainers                 Low Risk           At least yearly                 At least                 At least
Group                                                                                         Monthly                  weekly
                                                                    (per cent)
                                                                      Males
             2001   2004     2007   2001    2004      2007   2001     2004    2007     2001    2004      2007   2001    2004      2007
14-19        27.2   30.4     29.2   30.0     32.7     33.4   13.3     10.9    12.9     19.8    15.1      15.7   9.6     10.9      8.8
20-29         8.4    7.8     11.1   27.4     27.4     26.1   21.7     20.8    19.0     27.8    26.6      26.6   14.6    17.4      17.2
30-39         9.8    8.9     10.9   39.7     37.5     40.6   22.2     24.4    21.1     20.4    18.9      17.5   7.8     10.3      10.0
40-49        11.4    8.6     10.8   50.7     49.5     47.0   17.2     19.4    18.9     12.9    12.8      14.5   7.7      9.7      8.7
50-59        12.5   10.5     9.9    59.1     58.3     59.4   11.7     13.4    12.9     8.5     10.6      9.5    8.2      7.1      8.2
60+          20.1   17.3     17.3   66.5     68.1     67.8   5.8       6.5       6.4   3.6      3.8      4.5    4.0      4.3      4.0
Total        14.1   12.9     14.0   46.5     46.8     47.2   15.5     16.1    15.1     15.3    14.4      14.3   8.5      9.8      9.3
                                                                    Females
             2001   2004     2007   2001    2004      2007   2001     2004    2007     2001    2004      2007   2001    2004      2007
14-19        25.3   27.6     28.7   28.3     30.2     30.1   13.4     12.9    12.9     21.2    18.8      18.8   11.8    10.5      9.5
20-29        11.3   13.4     14.8   32.1     29.4     29.2   20.4     20.3    20.7     26.8    26.0      23.1    9.3    10.9      12.2
30-39        16.0   13.0     13.5   47.5     47.8     46.9   18.9     19.5    19.5     12.8    13.8      13.3   4.8      5.8      6.8
40-49        16.4   13.8     13.9   55.1     53.4     53.4   14.7     16.8    16.2     9.3     10.5      10.2   4.6      5.5      6.3
50-59        21.9   19.5     18.1   62.0     63.4     62.5   8.6       9.4    10.2     4.3      4.6      4.8    3.3      3.2      4.4
60+          33.0   31.6     31.1   62.3     61.3     61.3   2.0       4.0       3.9   1.3      1.5      2.1    1.3      1.6      1.5
Total        20.8   19.8     20.1   49.6     49.4     49.3   12.7     13.5    13.4     11.6    11.5      10.9   5.3      5.7      6.2
                                                                    Persons
             2001   2004     2007   2001    2004      2007   2001     2004    2007     2001    2004      2007   2001    2004      2007
14-19        26.2   29.0     29.0   29.2     31.5     31.8   13.4     11.9    12.9     20.5    16.9      17.2   10.7    10.7      9.1
20-29         9.9   10.6     12.9   29.8     28.4     27.6   21.1     20.5    19.8     27.3    26.3      24.9   12.0    14.2      14.7
30-39        13.0   11.0     12.2   43.7     42.7     43.8   20.5     21.9    20.3     16.5    16.4      15.3   6.3      8.0      8.4
40-49        13.9   11.2     12.4   52.8     51.4     50.2   16.0     18.1    17.6     11.1    11.6      12.3   6.2      7.6      7.5
50-59        17.1   15.0     14.0   60.5     60.8     61.0   10.2     11.4    11.6     6.4      7.6      7.1    5.8      5.1      6.3
60+          27.1   25.0     24.7   64.2     64.5     64.3   3.7       5.2       5.1   2.4      2.5      3.2    2.6      2.9      2.7
Total         17.5 16.4 17.1 48.1 48.2 48.3 14.1 14.8 14.2 13.4 12.9 12.6                              6.9   7.7                  7.8
     (a)   Not consumed alcohol in the previous 12 months.
     (b)   For males, the consumption of 7 or more standard drinks on any one day. For females, the consumption of
           5 or more standard drinks on any one day.
           Sources: 2001, 2004 and 2007 National Drug Strategy Household Survey: First results




                                                                                                                          48
Table 5: Long Term Harm - Alcohol consumption, risk of harm in the long
term: proportion of the population aged 14 years or older, by age and sex,
Australia 2001, 2004 and 2007
                                                                                                (b)
                                                                                Level of Risk
                                   (a)
Age Group             Abstainers                       Low Risk                     Risky                     High Risk
                                                                   (per cent)
                                                                     Males
               2001      2004        2007       2001    2004      2007   2001       2004        2007   2001     2004        2007
14-19          27.2       30.4           29.2   64.0     62.0     63.7    6.1        4.9         4.4   2.7      2.8         2.6
20-29           8.4        7.8           11.1   77.1     77.7     73.4    9.5        8.7         9.3   5.0      5.7         6.2
30-39           9.8       8.9            10.9   81.4     80.8     79.2    5.8        7.5         6.2   3.1      2.8         3.7
40-49          11.4       8.6            10.8   79.0     82.1     79.6    6.4        5.7         6.0   3.2      3.6         3.5
50-59          12.5       10.5           9.9    75.8     78.8     78.9    7.3        6.9         6.1   4.3      3.8         5.1
60+            20.1       17.3           17.3   71.9     74.8     75.3    5.4        5.2         4.9   2.6      2.7         2.5
14+            14.1       12.9           14     75.6     77.0     75.8    6.7        6.5         6.2   3.5      3.6         3.9
                                                                    Females
               2001      2004        2007       2001    2004      2007   2001       2004        2007   2001     2004        2007
14-19          25.3       27.6           28.7   60.1     60.1     60.7     9.9        8.3        6.7   4.7      4.0         3.9
20-29          11.3       13.4           14.8   73.7     71.6     68.8    10.9       12.2       11.0   4.0      2.9         5.4
30-39          16.0       13.0           13.5   75.4     77.1     75.8    6.8        7.7         7.7   1.9      2.1         3.0
40-49          16.4       13.8           13.9   73.9     75.9     74.1    7.8        8.0         9.3   1.9      2.3         2.6
50-59          21.9       19.5           18.1   70.7     73.1     72.3    5.9        6.0         6.9   1.5      1.4         2.7
60+            33.0       31.6           31.1   62.6     63.3     63.4    3.7        4.3         4.7   0.7      0.9         0.8
14+            20.8       19.8           20.1   69.8     70.6     69.4    7.2        7.5         7.6   2.2      2.1         2.8
                                                                    Persons
               2001      2004        2007       2001    2004      2007   2001       2004        2007   2001     2004        2007
14-19          26.2       29.0           29.0   62.1     61.1     62.2    8.0        6.6         5.6   3.7      3.4         3.2
20-29           9.9       10.6           12.9   75.4     74.7     71.1    10.2       10.4       10.2   4.5      4.3         5.8
30-39          13.0       11.0           12.2   78.3     78.9     77.5    6.3        7.6         7.0   2.5      2.4         3.3
40-49          13.9       11.2           12.4   76.5     79.0     76.8    7.1        6.9         7.7   2.6      3.0         3.1
50-59          17.1       15.0           14.0   73.3     76.0     75.6    6.6        6.4         6.5   2.9      2.6         3.9
60+            27.1       25.0           24.7   66.8     68.6     68.9    4.4        4.7         4.8   1.6      1.7         1.6
Total          17.5       16.4           17.1   72.7     73.7     72.6    7.0        7.0         6.9   2.9      2.8         3.4


(a) Not consumed alcohol in the previous 12 months
(b) For males, the consumption of up to 28 standard drinks per week is considered ‘Low risk’, 29 to 42 per week
    ‘Risky’, and 43 or more per week ‘High Risk’. For females, the consumption of up to 14 standard drinks per
    week is considered ‘Low risk’, 15 to 28 per week ‘Risky’, and 29 or more per week ‘High Risk’.
    Sources: 2001, 2004, and 2007 National Drug Strategy Household Survey: First results




                                                                                                                       49
                                             28 November 2008




Collins and Lapsley report review




Report by Access Economics Pty Limited for

National Alcohol Beverage
Industry Council
                                                                                          Collins and Lapsley report review




CONTENTS

Executive summary...............................................................................................................i
1.      Background .................................................................................................................1
2.      General methodological issues..................................................................................2
        2.1     Cost definitions.....................................................................................................................2
        2.2     The concept of drug abuse ..................................................................................................3
        2.3     Abusive and addictive drug use ...........................................................................................4
        2.4     Demographic and human capital approaches .....................................................................5
        2.5     General equilibrium (GE) impacts of drug abuse.................................................................6
        2.6     Private and social costs of drug abuse and their policy significance...................................7
        2.7     Cost of consumption of abused drugs .................................................................................9
        2.8     Avoidable costs of drug abuse’ ............................................................................................9
        2.9     The incidence of abuse costs ..............................................................................................9
        2.10    Active and involuntary smoking .........................................................................................10
        2.11    Valuation of life...................................................................................................................10
        2.12    Pain and suffering ..............................................................................................................11
        2.13    Welfare...............................................................................................................................11
        2.14    Revenue impacts of drug abuse ........................................................................................12
        2.15    Research, education and drug program costs...................................................................12
        2.16    New or revised methodologies and data ...........................................................................13
3.      Analysis of cost items ..............................................................................................16
        3.1     Epidemiological estimates .................................................................................................16
        3.2     Crime costs ........................................................................................................................17
        3.3     Health costs .......................................................................................................................18
        3.4     Productivity costs ...............................................................................................................18
        3.5     Road accident costs...........................................................................................................19
        3.6     Fire costs............................................................................................................................19
4.      Concluding sections of Collins and Lapsley (2008a)..............................................20
References..........................................................................................................................22


FIGURES
Figure 2-1: Incidence and prevalence approaches to measurement of annual costs                                                                      3
Figure 2-2: DWL of taxation                                                                                                                        12



While every effort has been made to ensure the accuracy of this document, the uncertain nature of economic data, forecasting
and analysis means that Access Economics Pty Limited is unable to make any warranties in relation to the information
contained herein. Access Economics Pty Limited, its employees and agents disclaim liability for any loss or damage which
may arise as a consequence of any person relying on the information contained in this document.
                                                           Collins and Lapsley report review


TABLES
Table 2-1: Matrix of cost items                                                           9




GLOSSARY OF ACRONYMS
 ABS                     Australian Bureau of Statistics
 AIHW                    Australian Institute of Health and Welfare
 BCA                     benefit cost analysis (Collins and Lapsley use this acronym)
 CBA                     cost benefit analysis
 CEA                     cost effectiveness analysis
 DALY                    disability adjusted life year
 DUCO                    Drug Use Careers of Offenders
 DUMA                    Drug Use Monitoring in Australia
 DWL                     deadweight loss
 GE                      general equilibrium
 NHMRC                   National Health and Medical Research Council
 NHS                     National Health Survey
 QALY                    quality adjusted life year
 VSL(Y)                  value of a statistical life (year)
 WTP                     willingness to pay
                                                          Collins and Lapsley report review




EXECUTIVE SUMMARY
This report reviews the methodology, assumptions and conclusions in the monograph:
Collins DJ, Lapsley HM (2008a) The costs of tobacco, alcohol and illicit drug abuse to
Australian Society in 2004/05, Report for the Department of Health and Ageing. The
emphasis of our critique is in relation to the alcohol component only.

Major conceptual and methodological problems

It is not clear that the report meets its fundamental aims, and indeed a second report (Collins
and Lapsley, 2008b) has now been produced in order to address the first two objectives
properly. One of the objectives not met – the cost of alcohol abuse that might be avoided –
involves the concept of ‘avoidable costs’, which is only applicable for incidence-based
costings or in forward-looking prevalence-based costings for those costs that are reversible.
As such, any potential applications in relation to cost benefit analysis (CBA) or cost
effectiveness analysis (CEA) of preventive health care initiatives to reduce abusive alcohol
consumption need to take account of the limited reversibility of impacts, the cost of the
interventions themselves and the relative efficacy of such measures. Emphasis should be
placed on well targeted interventions (ie, those that actually address problem drinking) rather
than blanket interventions that bluntly aim to reduce everyone’s consumption, whether or not
the extent of consumption is indeed a problem. The value of Collins and Lapsley (2008a)
is thus not clear.

A fundamental flaw is that the report uses what it terms a ‘demographic approach’ to cost
estimation. This uncommon terminology and approach means that the authors have
attempted to estimate the impact of alcohol abuse retrospectively ie, the difference in 2004-
05 if people in the 40 years prior had not abused. The counterfactual is thus impossible,
since the past cannot be changed. The reason for undertaking a retrospective rather
than forward looking approach is unclear, as is the policy usefulness of such an
approach. The demographic approach used by Collins and Lapsley is likely to yield a
much higher cost estimate than the human capital approach.

A superior and more useful forward-looking approach, where the costs of prevalent abuse in
the year 2004-05 are estimated, is how Access Economics conducts cost of illness analysis
(on either a prevalence or incidence basis depending on applications), in line with
international best practice. Collins and Lapsley mistakenly refer to this as the ‘human capital’
approach (and, moreover, refer to premature deaths only rather than all losses), although
that terminology actually refers to a specific way of measuring productivity losses where the
cost is equated with lost wages in the context of having to replace workers in more than
frictional situations, according to standard health economics texts (eg, Pizzi and Lofland,
2005:68). Access Economics uses Australian data on employment and wage rates by age
and gender to estimate productivity losses using a human capital approach and recommends
this approach in cost of illness analysis in developed countries in line with the literature.

Another conceptual problem is the definition of alcohol abuse, which is never clearly defined
and definitions implicit in different cost elements vary throughout the report. The conceptual
definition used by Collins and Lapsley is circular: ‘Abuse occurs if the community incurs net
costs as a result of drug abuse’. The data on crime do not permit an estimate of alcohol
abuse or intoxication, only alcohol consumption (which may not be abusive) close to the time
of the crime. In other cost estimates it is unclear what the definition of abuse is; amounts and
patterns of alcohol consumption are not mentioned in the report – eg, it is unclear where
there might be concordance with the National Health and Medical Research Council


                                                                                               i
                                                                     Collins and Lapsley report review


(NHMRC) guideline definitions1. Another report by the same authors with an inaccessible
website reference was used to justify the assumption that 20% of alcohol consumption is
addictive, and then an estimate of 30% was used to estimate abusive consumption, with no
evidence or references provided for that parameter and sensitivity on it of an extremely high
50% only. Lack of attempts to validate this important parameter, potentially through
primary data collection, is a serious flaw, as is the lack of lower sensitivity analysis.

In economics, private costs/benefits are borne by the individual producer/consumer and
social costs/benefits are these plus ‘externalities’ ie, those costs/benefits that spill over to
other entities. This concept is very well understood and used; however, Collins and Lapsley
(2008a:8) define social costs as negative externalities (based on ‘Markandya and Pearce’,
not provided in the references) and ignore private costs and private benefits. ‘Private’ is not
defined – it is unclear if it is supposed to refer to individuals’ cost or to firms’. There is also
confused discussion of rational behaviour and access to information, leading to an extreme
conclusion; Access Economics takes a ‘middle ground’ view between the rational addiction
theorists and Collins and Lapsley, that there remains a non-zero consumption value to the
abuser that is not limited to avoiding withdrawal effects. There is no mention in the report
of the very large private benefits of responsible alcohol consumption in terms of utility
value (choice, enjoyment). It is a major flaw to exclude private costs from total social
costs and private benefits from total social benefits and it is incorrect to claim that
costs borne by individuals are irrelevant to public decision-making. Access
Economics recommends a matrix approach to the net cost of alcohol abuse – by cost
type and bearer – with standard definitions and terminology so that cost elements are
clearly and accurately presented, as per Table 2-1 on page 9.

Data and methodological problems for specific parameters
         There are numerous problems with the methods and parameter values adopted in
         Collins and Lapsley (2008a) in relation to pain, suffering and the value of life
         (Sections 2.11 and 2.12) where Access Economics (2008c) provides an alternative,
         most recent and superior approach.
         Collins and Lapsley ignore deadweight losses (DWLs) associated with transfers,
         saying it is not possible to measure them. Access Economics suggests a method and
         values based on Lattimore (1997) and Productivity Commission (2003) (Section 2.13).
         This represents a major flaw as the distribution of costs will be substantially affected by
         transfers and the total real costs are also likely to be underestimated.
         Although it is not spelled out, the revenue estimation (summarised in the final table of
         the report’s Executive Summary) shows that taxation raised from alcohol abuse
         exceeds its cost to Government, so it is budget positive and abusers pay their way by
         some $1.4 billion (2004-05 prices) per annum. Alcohol taxes thus more than pay for
         the social costs of alcohol abuse, by a considerable margin, each year. Moreover,
         taxation of alcohol increased from $3.37 billion in 1998-99 to $5.11 billion in 2004-05, a
         52% increase (8.6% per annum), greatly exceeding population or consumption growth.

Weaknesses of individual cost estimates
         Epidemiological estimates (the number or demographic characteristics of the
         estimated abusers are not provided, which is a weakness). A major concern is that
         attributable fractions are assumed unchanged historically, which the Appendix A
         author (Pollard) describes as a ‘serious limitation’. It is also unnecessary, since it



1
    In excess of four standard drinks per day for men or two for women on average.


                                                                                                     ii
                                                          Collins and Lapsley report review


     would have been possible to use a forward-looking rather than retrospective approach
     and use the Begg et al (2007) forward projections instead. Section 3.1 details other
     concerns.
     Crime costs (estimated as $1.424 billion). The Appendix B author (Makai) cautions
     against use of the data and the substantial risk of over-estimation of this cost item. The
     many reasons for potential overestimation and caution regarding uncertainty are
     included in Section 3.2, relating mainly to the source data from two self-report studies,
     which are also highlighted by Collins and Lapsley in their Section 9 on areas that would
     benefit from future research improvement. Given the upside bias, there should have
     been sensitivity analysis for this cost element as well as greater transparency in
     the calculation process.
     Productivity losses (estimated as $3.579 billion in the workforce and $1.571 billion in
     the household). These estimates have four sources of upward bias: (1) estimates of
     absenteeism are very high and rely on the oldest (1990) Australian Bureau of Statistics
     (ABS) National Health Survey (NHS) data when alternative 2004-05 data are available;
     (2) unpaid work is estimated based on the (high) work absenteeism estimates,
     although someone may simply defer household work till they are feeling better; (3) the
     ‘demographic’ approach is used rather than the human capital approach; and
     (4) replacement cost valuation is used when opportunity cost valuation is lower and
     preferable. There is inadequate detail regarding mortality, morbidity and national
     accounts data to replicate or validate the method, a serious flaw.
     Health expenditure (estimated as $1.977 billion) is estimated from a number of
     different data sources, and suffers from omissions and incompleteness that is identified
     by Collins and Lapsley as an area for improvement in Sections 9.8, 9.9 and 9.11.
     These weaknesses would not exist if the more appropriate Australian Institute of Health
     and Welfare (AIHW, 2005) dataset had been adopted, which would have permitted
     detailed breakdown of actual alcohol-attributable costs by alcohol-related condition and
     by type of health system expenditure, using a more comprehensive and
     methodologically superior top-down attribution and enabling estimates of omitted items
     such as over-the-counter pharmaceuticals, pathology and imaging, allied health and
     other health expenditures. An example of how this can be done is provided in Access
     Economics (2008b) in relation to the cost of obesity.
     Road accidents (estimated as $2.202 billion). This cost item is based on an out-dated
     Bureau of Transport Economics (2000) report (see Sections 2.11, 2.12. 2.16.5 and
     3.5). It would be preferable to use unit cost data for road accidents from Austroads
     (2007, 2006, 2003).
     Resources used in abusive consumption (estimated as $1.689 billion). Collins and
     Lapsley include the cost of purchasing abused substances in their cost estimates.
     While this is appropriate for smoking and illicit drugs, where any consumption is
     harmful, it is inappropriate for alcohol since the problem is pattern of consumption, not
     average consumption and hence misallocation of resources in production. Hence,
     Collins and Lapsley over-estimate the total costs of alcohol consumption by
     18.4% and this $1.7 billion item should be completely excluded.

Presentational weaknesses and conclusion

Overall the report is poorly structured with inadequate cross-referencing, there is a paucity of
supporting evidence and references for claims, the referencing is incomplete and there is a
very serious lack of transparency in calculation processes. Methodology and data use are
singularly poor.



                                                                                               iii
                                                        Collins and Lapsley report review


There is apparent upward bias in most of the cost estimates and uncertainty is not dealt with
due to grossly inadequate sensitivity analysis. Using a best possible approach would exclude
one cost item, making the overall cost of alcohol abuse at least 18% lower. The other
findings of the report should be viewed with substantial scepticism.

We agree with the final suggestion (in Collins and Lapsley’s Section 9.14) that the data
needs are reviewed and an expert advisory panel convened, also to review methods, given
the serious weaknesses identified in this critique. We would recommend that a time series of
comparable estimates of the cost of alcohol abuse is calculated based on a sound
methodology and the best available data.


Access Economics
November 2008




                                                                                            iv
                                                           Collins and Lapsley report review




1. BACKGROUND
The purpose of this report is to undertake a review of the methodology, assumptions and
conclusions in the monograph: Collins DJ, Lapsley HM (2008a) The costs of tobacco, alcohol
and illicit drug abuse to Australian Society in 2004/05, Report for the Department of Health
and Ageing.

The monograph is the fourth study by Collins and Lapsley of the social costs of drug abuse in
Australia, presenting estimates on the costs of alcohol (and tobacco and illicit drugs) for
2004-05. The emphasis of our critique is in relation to the alcohol component only, although
necessarily there are some occasions where tobacco and illicit drugs are also mentioned due
to pertinence or overlap.

The report is structured as follows:
      Section 2 summarises general methodological and data issues, in line with the
      structure of the report, where Sections 1, 2 and 3 cover such matters.
      Section 3 assesses the cost findings based on the methods and data, item by item and
      then in total, again following the structure of the original report – Sections 4, 5, 6 and 7.
      Section 4 evaluates the final Sections – 8, 9 and 10, suggesting alternative methods
      and parameters.

Access Economics was selected as the reviewer for this project due to extensive experience
and world expert status in cost of illness modelling. The firm’s health economics group has
conducted nearly forty cost of illness studies in the past seven years, across five continents,
many of which have led to peer-reviewed journal articles (eg, Taylor et al, 2006; Hillman et
al, 2006) and is currently drafting, with John Hopkins University, a paper on optimal methods
for undertaking cost of illness analysis, as a result of an international forum on the issue in
2007 in Vancouver. The group is also experienced in issues related to alcohol consumption
and health impacts including modelling the elasticity of demand for various alcoholic
beverages from industry data in a taxation context (not publicly available), a technical critique
of injury risk from excessive alcohol consumption (Access Economics, 2008a) and assessing
indicators of alcohol consumption among young people.




                                                                                                 1
                                                           Collins and Lapsley report review




2. GENERAL METHODOLOGICAL ISSUES
Section 1 of Collins and Lapsley (2008a) outlines the purpose of their costing exercise as
informing four policy objectives:
1     to give alcohol, tobacco and other drugs higher priority on the public policy agenda –
      ie, their starting point is that current priority is inadequate, although there is no
      evidence provided that this is in fact the case in relation to the marginal cost of more
      intervention relative to its marginal benefit or effectiveness;
2     to appropriately target policies in order to address specific problems – it is not clear
      from the rest of the report how this is actually achieved eg, in relation to identifying the
      best interventions to address current problems such as youth binge drinking;
3     to identify information gaps, research needs and desirable refinements to national
      statistical reporting systems – this is useful although in this case Collins and Lapsley
      have failed to identify and use some data that are already available; and
4     to develop baseline measures from which changes over time and international
      outcomes can be compared – this is also desirable, although in the case of alcohol it is
      noted that change in methods since their last costing mean that the two estimates are
      not comparable (page xii).

In summary, it is not clear that the report meets its fundamental aims, and indeed a second
report (Collins and Lapsley, 2008b) has now been produced in order to address the first two
objectives properly. The value of Collins and Lapsley (2008a) is thus not clear.

Section 2 of Collins and Lapsley (2008a) discusses a variety of methodological issues. For
ease of comparison, in this critique these issues are addressed in the sub-sections below in
the same order as in the original report.

2.1         COST DEFINITIONS
Costs are defined as ‘the value of the net resources which in a given year are unavailable to
the community for consumption or investment purposes as a result of the effects of past and
present drug abuse, plus the intangible costs imposed by this abuse’.

By using a retrospective concept, the authors note that their counterfactual is completely
hypothetical (the past cannot be changed). This means that the costing cannot be used as a
basis for policy in relation to present or future alcohol consumption, since the future is not the
same as the past. The reason for undertaking a retrospective rather than forward looking
approach is unclear.

In undertaking cost of illness analysis, Access Economics adopts either a ‘current
prevalence’ or ‘current incidence’ approach, depending on the research question that is
being addressed. This means the cost of prevalent or new (respectively) cases of alcohol
abuse are estimated, which is relevant for policy, since these and their impacts can be
changed.

Figure 2-1 helps depict the methodological differences, with 2008 as a base year (1990 and
2015 have been arbitrarily included as the start and end points of past and future impacts, for
the illustrative purposes of the diagram only). The figure considers three different cases:




                                                                                                2
                                                          Collins and Lapsley report review


      a, whose onset of alcohol abuse was in the past and who has incurred the associated
      costs up to the year in question (2008), with associated lifetime costs of A + A* (he may
      have ceased abusing some years ago, but his final year of costs are in 2008);
      b, whose onset of alcohol abuse was in the past and who has incurred the associated
      costs in 2008 as well as in the past and future, with associated lifetime costs of
      B + B* + B**; and
      c, whose onset of alcohol abuse occurred in 2008, with lifetime costs of C + C*.

Using an incidence approach, only cases like ‘c’ would be included, with the total cost
estimate equivalent to the sum of all the costs in the base year (ΣC) plus the present value of
all the future costs (ΣC*).

Using a prevalence approach, costs in 2008 relating to a, b and c would all be included, and
are measured as Σ(A, B, C):
     3     retrospectively looking at abuse cases that occurred in the period represented by
           A* and B* in the Collins and Lapsley approach; or
     3     the Access Economics suggested approach looking forward to impacts in the
           periods represented by B** and C*, useful for policy because all the costs relate
           to abuse that occurs in the year 2008 and can thus be ameliorated or prevented
           by new interventions starting now.

 FIGURE 2-1: INCIDENCE AND PREVALENCE APPROACHES TO MEASUREMENT OF ANNUAL COSTS
      1990                                   2008                                    2015
                        A*                     A
                        B*                     B                    B**
                                               C                     C*

Collins and Lapsley estimate costs for people who ever abused alcohol in the last 40 years.
Because they have estimated the prevalent costs in 2008 retrospectively rather than the
prevalent cases of abuse and their associated costs, it is not possible to differentiate whether
the costs today result from alcohol consumption patterns that could be changed, or if the
person stopped abusing years ago, but still incurs costs. This is not particularly useful from a
policy perspective, and represents an inferior definition and specification of cost incursion
relative to an alternative forward-looking definition. Their assumption of impacts for ‘at least
40 years’ appears arbitrary, with no reference or citation given for the selection of this
parameter, and with potential overestimation of the duration of harms.

The issue arises again in Section 2.4.

2.2          THE CONCEPT OF DRUG ABUSE
Collins and Lapsley note that there is no concept of alcohol abuse in the National Alcohol
Strategy (Department of Health and Aged Care, 2001), rather it is defined there as ‘misuse’.
They thus use a rather circular definition from a costing perspective: ‘Abuse occurs if the
community incurs net costs as a result of drug abuse’ (page 4). It is not clear how a costing is
undertaken when the thing being costed is defined in terms of its net costs.

Importantly, there is no discussion in this section of the beneficial effects of moderate alcohol
consumption (relative risks less than unity imply protective health effects), which is highly
relevant to this conceptual issue of ‘net costs’. Fortunately, Collins and Lapsley do include


                                                                                               3
                                                          Collins and Lapsley report review


these benefits but this is only discovered in Section 4 of their report, unlike the NHMRC draft
guidelines where these benefits are ignored (see our earlier critique of this matter – Access
Economics, 2008a).

Health benefits of alcohol consumption occur where moderate consumption has a protective
fact against particular conditions, relative to no consumption. The evidence suggests that
the protective effect comes from a regular pattern of drinking, preferably with meals (to help
reduce the high levels of blood lipids or fats found after eating) of a low to moderate amount
of alcohol. Examples included by Collins and Lapsley in their report are:
      for males: ischaemic heart disease, cholelithiasis and heart failure; and
      for females: ischaemic heart disease, cholelithiasis, ischaemic and haemorrhagic
      stroke, and hypertension.

Ischaemic heart disease, hypertension and heart failure are highly prevalent and high burden
conditions (Access Economics, 2005).

In addition, alcohol may have a number of other health benefits (NHMRC, 2001) and there is:
      strong evidence that alcohol consumption may be protective against gallstone
      production; and
      early evidence that low-level alcohol consumption may reduce the risk of Type 2
      diabetes, but further research is required to establish the nature of the gains.

Other evidence (examples of the numerous studies available in relation to the benefits) is
provided in Laslett et al (2002), English et al (1995), Single et al (1999) and Ridolfo and
Stevenson (2001).

2.3        ABUSIVE AND ADDICTIVE DRUG USE
Another major weakness of the methodology is the extension of Collins and Lapsley’s earlier
definition of alcohol addiction in order to determine the extent of alcohol abuse. Page 5 of
the report states: ‘In Collins and Lapsley (1996) it was concluded that 20% of alcohol
consumption was by addicted drinkers.’ ‘This does not address the issue of misused alcohol.’
‘Accordingly, the proportion of abusive alcohol consumption is assumed to be higher at 30%.’

The reference provided for Collins and Lapsley (1996) was a website address
http://www.health.gov.au/pubhlth/publicat/document/mono30.pdf The report was not
accessible at this address in September 2008 when attempts were made to access it. As
such, it has not been possible to verify the validity of the original 20% assumption, and a risk
is that it is high, when one contemplates the extent of alcohol that is consumed responsibly
and enjoyed as such by a majority of people. Of more concern is the leap to the 30%
parameter used in the 2008a report, which has no documented basis and no reference or
discussion of why it was selected rather than, say, 22% or 25%. Moreover, there is no
sensitivity analysis undertaken on a lower parameter, only on an even higher (possibly
extreme) parameter of 50%.

To base such a critical parameter value on an unpublished work by the same authors with no
reference to the extent of literature or attempt to quantify the parameter using data, and to
conduct no lower sensitivity analysis is potentially a serious flaw.




                                                                                              4
                                                                  Collins and Lapsley report review


2.4          DEMOGRAPHIC AND HUMAN CAPITAL APPROACHES
The so-called ‘demographic approach’ is not a commonly used approach compared to the
human capital approach, but the report simply references another report co-authored by
Collins and Lapsley (Single et al, 2003) as an explanation for using it, while ‘acknowledg[ing]
the legitimacy of both approaches’.

Searching briefly on-line for other references to such a ‘demographic approach’ revealed
other concepts using the same terminology, but no other literature was found which uses the
term the same way as Collins and Lapsley use it. Their ‘demographic approach’ appears
essentially to refer to comparing historical abuse patterns with the (impossible)
counterfactual described in Section 2.1 of no past abuse (ie, the sum of A, A*, B, B* and C).
Their description of the human capital approach is also unusual in that it is used specifically
to describe the discounting back of future cost streams (again as per Section 2.1 the sum of
B, B**, C and C*)2, whereas human capital concepts traditionally relate to equating the value
of lost productivity with lost wages in the context of having to replace workers in more than
frictional situations (either with retrospective or forward looking applications), according to
standard health economics texts. For example, Pizzi and Lofland (2005:68) summarise the
two standard approaches to productivity cost measurement as below (there is no mention of
the ‘demographic approach’).

      Two schools of thought exist for monetizing or placing a dollar value on lost
      productivity: the human capital approach (HCA) and the friction cost approach
      (FCA).

      Introduced in the 1960s, the HCA appears to be the more frequently adopted
      method based on current literature. Simplicity and ease of use are the HCA’s
      most notable characteristics. It is a function of an individual’s wage, with one hour
      of lost productivity equal to one hour of pay… Productivity losses are calculated
      from the beginning of the illness until the time of recovery or the end of the age of
      gainful employment.’

In our extensive global cost of illness work summarised in Section 1, Access Economics
uses Australian data on employment and wage rates by age and gender to estimate
productivity losses using a human capital approach and recommends this approach in cost of
illness analysis in developed countries (the frictional approach is more appropriate in
developing countries where unemployment rates are higher). Pizzi and Lofland (2005) make
the same recommendations, citing a variety of other well-established sources.

Productivity losses may arise when people work less than they otherwise would (either being
employed less, being absent more often or being less productive while at work) or they may
die prematurely. Collins and Lapsley appear to incorrectly ascribe productivity losses as
applying only to premature death (recall the quotation in footnote 2).

The authors correctly note that the two approaches yield different results and that,
specifically (page 6): ‘The human capital approach will place a higher emphasis on costs
which involve the discounting back to present day values.’ This appears to be a rather veiled
way of saying that the human capital approach, by discounting future costs, will likely yield a


2
  ‘The essential difference between the two approaches can be summarised in the following way. The human
capital approach calculates the present and future production costs of abuse-induced deaths which occur in the
present year. The demographic approach calculates the present production costs of abuse-induced deaths which
have occurred in past and present years.’ (Collins and Lapsley, 2008a:6)


                                                                                                            5
                                                          Collins and Lapsley report review


lower value than the demographic approach, which does not discount. In other words, the
demographic approach is likely to yield a much higher cost estimate than the human
capital approach. Again, it is not clear why this approach has been adopted.

Collins and Lapsley (bottom of page 6) note that: ‘the human capital approach is necessarily
always adopted in benefit-cost analysis (BCA)’ and ‘if abuse cost studies are to be extended
into the BCA of proposed drug programs, the human capital approach is indispensable.’ It is
thus curious why the authors adopted the demographic approach given their objectives (as
outlined in the preamble to this Section 2) and, importantly, the objectives of the Department
of Health and Ageing.

The final statement in this section is incorrect. Collins and Lapsley claim that the human
capital approach estimates the value of the loss of a life, while the demographic approach
estimates the value of the loss of a year’s living. The human capital approach can also be
used to estimate the latter.

2.5        GENERAL EQUILIBRIUM (GE) IMPACTS OF DRUG ABUSE
This section (page 7) argues that a GE approach to estimating costs is impossible and
inappropriate. Neither is the case. A GE approach is better if impacts are substantial and
are likely to affect the overall clearing of labour and other factor (and product) markets. In
GE, prices and wages adjust, which is not taken into account in the Collins and Lapsley
partial equilibrium approach. Access Economics has been careful in our cost of illness
analyses to ensure that the ‘small size’ assumption is observed; however, Collins and
Lapsley argue substantial impacts (eg, 50% of all alcohol consumption is abusive) yet do not
employ GE techniques to deal with the implications of these assumptions.

There are other more minor problems.
      The second paragraph claims that it is difficult/impossible to imagine that agricultural,
      manufacturing and distributive resources would have no alternative use, which may
      apply to illicit drugs and smoking but is not as relevant to alcohol when it is population
      consumption patterns which are the problem, not average consumption per capita per
      se (see also Section 2.7).
      The third paragraph is more relevant to the discussion of frictional and human capital
      approaches to measuring productivity losses than to this section.
      The fourth paragraph claims that no GE models exist that could estimate the impacts of
      eliminating drug abuse. Access Economics’ GE model could be used to do this, as
      could a number of other GE models, we understand, since their purpose is precisely to
      forecast robustly future production and consumption patterns and ‘shocks’ to current
      equilibrium. The real issue is whether it is worth doing the work to adequately enlarge
      the alcohol specifications in such a model and ‘shock’ it to remove the effects of abuse,
      when the impacts may be considerably less than what Collins and Lapsley speculate,
      in which case a GE model is possibly ‘modelling overkill’ and a partial equilibrium
      model would suffice.
      The last paragraph is somewhat controversial. It is not clear why it is speculative to
      model how government might allocate resources in the future, but not speculative to
      model how they might have allocated them in the past but didn’t.




                                                                                              6
                                                                   Collins and Lapsley report review


2.6            PRIVATE AND SOCIAL COSTS OF DRUG ABUSE AND
               THEIR POLICY SIGNIFICANCE
This is perhaps the most confused of the methodological sections, partly due (again) to
terminology issues.

In economics, private costs/benefits are borne by the individual producer/consumer and
social costs/benefits are these plus ‘externalities’ ie, those costs/benefits that spill over to
other entities. This concept is very well understood and used, for example even accurately
depicted      with     referencing    to     expert   sources    on    Wikipedia       –    see
http://en.wikipedia.org/wiki/Externality and http://en.wikipedia.org/wiki/Social_costs (the
definition below is from the latter source):

         In economics, social cost is defined as the sum of private and external costs…
         Rational choice theory assumes that individuals only consider their own private
         costs when making decisions, not the costs that may be borne by others.

Collins and Lapsley (2008a:8) adopt a different definition base on a source they cite: ‘What
Markandya and Pearce call social costs are often called negative externalities.’ In restricting
the analysis to the negative externalities, private costs and private benefits are ignored.
‘Since the objective of this study is to estimate the social cost of drug abuse, purely private
costs are not relevant to and are not counted in this study’ (page 9).

This is a major flaw, since social costs should include private costs and social benefits
should include private benefits. The title of the report does not reflect this distinction, the
stated objectives and Executive Summary do not indicate that a non-standard definition has
been used that in fact refers only to externalities, and the Markandya and Pearce reference
is not provided in the report. Worst of all, ‘private’ is not defined – it is unclear if it is
supposed to refer to individuals’ cost or to firms’.

There is no mention in the report of the very large private benefits of responsible alcohol
consumption in terms of utility value (choice, enjoyment). Yet clearly this is an important
factor to consider if an objective is formulating policy to reduce abuse and encourage healthy
levels and patterns of consumption. The benefits of moderate alcohol consumption are
summarised in Section 2.2. However, the health benefits are dwarfed by the consumption
and production benefits. In economic terminology, these are the ‘consumer surplus’ and
‘producer surplus’ associated with consumption in healthy quantities. As an industry, the
sector contributes to Australian production, exports and employment – exports in 2005-06 of
beer and malt were $220 million and of spirits were $52 million3, while exports of Australian
wine were worth $2.8 billion and the domestic market was worth $1.9 billion4. The utility and
hence economic value of the sector is rarely considered in the literature and ignored in the
Collins and Lapsley report.

The discussion is further confused by ascribing the different approach to assumptions about
rationality and perfect information. Issues of imperfect information and irrationality apply
throughout producer and consumer markets and do not inter alia imply that a special case
should be made for the alcohol market. Furthermore, Collins and Lapsley assume that even


3
    http://www.austrade.com/Overseas-Food-and-beverage-capability-overview/default.aspx
4

http://www.abs.gov.au/AUSSTATS/abs@.nsf/ProductsbyReleaseDate/98E5C8579331C47BCA2573DB0011D760
?OpenDocument


                                                                                                   7
                                                           Collins and Lapsley report review


before commencing abusive behaviours, less informed/rational people ‘self-select’ to these
behaviours, for which no detailed evidence is provided (the one 20-year old Courtwright
reference could not be found in the references – the arbitrary sub-classification style of
referencing in the report is opaque, such that electronic searching is required to try to locate
where references appear).

Rational (non-addicted) people are generally aware that too much alcohol consumption is not
good for their health although the theory of rational addiction is dismissed in the report.
While some rational addiction theorists suggest that all addiction can be explained, Collins
and Lapsley rightly point out that some addiction is acquired in teenage years when typically
behaviour is based on less perfect information and shorter time horizons than in later years.
However, most people grow out of these behaviour patterns and some acquire them later in
life, so this counterargument is not perfect either. Access Economics takes a ‘middle ground’
view between the rational addiction theorists and Collins and Lapsley – that there remains
some consumption utility to the abuser that is not limited to avoiding withdrawal effects
(which does not sit well with most forms of alcohol abuse), while also acknowledging that
addiction for an individual is likely to result in consumption in excess of levels optimal for that
individual’s long term utility maximisation, and indeed this is why the issue is of interest to
policy makers.

Collins and Lapsley engage in this discussion as though it applied to all the people in their
costing. In their costing, one third (and 60% in the high case) of abusive consumption is
undertaken by non-addicts, so the arguments in this section do not apply to this large group.

A final claim is that only social costs (read ‘externalities’) are relevant to public decision
making. In making decisions about sensitive health, industry and other related policy issues,
governments need to take into account the overall impact on consumers, producers and
others in society, in order to estimate who bears the costs and benefits and the extent to
which certain groups are made worse or better off by changes. To make the assumption that
Collins and Lapsley make could lead to some bizarre conclusions, such as privatising all
health and welfare system costs for alcoholics, smokers or illicit drug users, since then they
would take them into account and they would no longer be relevant to government.

Thresholds for public financing decision making, as well as rationality and information as they
underlie microeconomic decision making, are correctly outlined and discussed in Access
Economics (2008c).

Better terminology would use the traditional definition of social costs, by type of cost and by
who bears them. Access Economics adopts this approach in our cost of illness work, using a
matrix format as shown in Table 2-1. There is no potential for confusion using this
framework, and it aligns with common parlance. We would recommend this approach (in line
with the paper co-authored with John Hopkins University referred to in Section 1).




                                                                                                 8
                                                                     Collins and Lapsley report review


                                   TABLE 2-1: MATRIX OF COST ITEMS

                                                               Bearer of cost                             Total
                                                                                                         (social)
                                               Indiv-       Govern-      Employers       Others in        costs
Cost type                                      iduals        ments                        society
 Health system expenditure                        x              x              -             x              x
 Productivity losses (individuals and             x              x              x             -              x
 carers)
 Aids, home modifications and other               x              x              -             x              x
 financial costs
 Deadweight losses from transfer                  x              x              x             -              x
 payments
 Net burden of disease                            x              -              -             -              x
 Total                                            x              x              x             x              x
Note: x means a cost (can be negative) is usually borne. – means the entity rarely or never incurs this cost type.
 ‘Individuals’ includes families and carers in the household. ‘Others’ include private health insurance providers,
           workers compensation bodies and non-government assistance organisations, for example.

2.7          COST OF CONSUMPTION OF ABUSED DRUGS
As noted in Section 2.5, it is population consumption patterns which are the problem, not
average consumption per capita per se. Suppose a binge drinker, instead of drinking nothing
during the week and then ten standard drinks on Friday night (abusive and potentially an
addiction), instead drank two standard drinks per day with one alcohol-free day per week
(healthy consumption). Overall consumption of alcohol would increase, but abuse would
decline.

Alcohol is thus different from smoking and illicit drugs, where the cost of consumption should
be included in the costs. However, since Collins and Lapsley do not demonstrate that alcohol
abuse declines when consumption declines, this cost should not be included and its inclusion
means that total costs are over-estimated.

2.8          AVOIDABLE COSTS OF DRUG ABUSE’
We agree that it is more sensible to estimate costs in such a way that avoidable costs can be
discerned, and thus question the approach adopted in this report, as discussed earlier.

The conceptual framework of ‘avoidable costs’ is only applicable for incidence-based
costings or in forward-looking prevalence-based costings for those costs that are reversible.
As such, any potential applications in relation to cost benefit analysis (CBA) or cost
effectiveness analysis (CEA) of preventive health care initiatives to reduce abusive alcohol
consumption need to take account of the limited reversibility of impacts, the cost of the
interventions themselves and the relative efficacy of such measures. Emphasis should be
placed on well targeted interventions (ie, those that actually address problem drinking) rather
than blanket interventions that bluntly aim to reduce everyone’s consumption, whether or not
the extent of consumption is indeed a problem.

2.9          THE INCIDENCE OF ABUSE COSTS
This section states in a rather long-winded way, that Collins and Lapsley adopt a ‘who writes
the cheque’ approach (ie, who bears the first round net cost impacts). This means that flow-
on effects over time as, for example, changes in some prices may lead to changes in other



                                                                                                                    9
                                                            Collins and Lapsley report review


prices or in wages, are not considered. We agree with this approach for partial equilibrium
analysis.

2.10         ACTIVE AND INVOLUNTARY SMOKING
This section is not relevant to this critique in its current form, although it is noted that alcohol
abuse like smoking also imposes costs on non-drinkers (eg, in cases of foetal alcohol
syndrome, domestic violence and motor vehicle accidents). These are rightly captured in a
social costing exercise, but Collins and Lapsley note that ‘data deficiencies mean that we
have been unable to produce this disaggregation’.

2.11         VALUATION OF LIFE
This is a very weak discussion of the many issues involved, which are discussed more
comprehensively in Access Economics (2008c). A series of issues are summarised below.
       Premature death is not the only healthy life to be lost; there is also the loss of healthy
       life due to the morbidity burden as opposed to the mortality burden. Collins and
       Lapsley frame the whole discussion as though it only applies to premature deaths.
       Premature death results in more than two types of first-round costs; as well as
       productivity losses and lost quality of life for the individual and loved ones, there are
       also lost taxation revenues (borne by government), deadweight losses (borne by others
       in society), and the bring forward of funeral costs (borne by family/friends).
       Choosing an appropriate discount rate is not difficult. Access Economics (2008c)
       provides a detailed rationale for adopting the 3% real rate, also favoured by most
       others in the field (eg, the Australian Institute of Health and Welfare, AIHW, in relation
       to healthy life – see Begg et al, 2007). Sensitivity analysis can be provided at different
       real rates if desired.
       The human capital approach can be used to measure productivity losses, and
       willingness to pay (WTP) approaches can be used to measure loss of quality of life.
       The costs borne by the individual can then be netted out of the WTP estimates to avoid
       double counting, and the one adopted by Access Economics and others. This is a
       superior approach to the ‘either-or’ discussion provided by Collins and Lapsley.
       The report confuses the concepts of the value of a statistical life (VSL) and the value of
       a statistical life year (VSLY), claiming that it estimates the value of the loss of one
       year’s living, rather than the loss of many years. Consistent with forward looking
       prevalence-based costing, Access Economics and others adopt the approach that if
       someone dies in 2008 due to alcohol abuse, then the present value of their future
       streams of healthy life (including production) are ascribed to that year (estimated as the
       VSL or the sum of VSLYs for remaining years of life expectancy, discounted). If
       someone is not working but alive in 2008 albeit with reduced quality of life due to their
       abuse, the value of their reduced quality of life including productivity is included for just
       one year (based on a fraction of the VSLY and separately identifying the lost wage
       component). It is not particularly clear from the methodological chapter how Collins
       and Lapsley ascribe (in detail) the costs of people who died prior to 2008, yet would
       otherwise have been enjoying life and employment in 2008.
       The reference to Bureau of Transport Economics (2000) has been superseded
       subsequently. The value of $53,267 for a VSLY is extremely low as a result of the
       outdated methods. Access Economics (2008c), for which the now Bureau of Transport
       and Road Economics was one stakeholder, provides a superior and higher estimate of
       the VSLY based on sophisticated meta-analysis of the WTP literature.


                                                                                                  10
                                                        Collins and Lapsley report review


2.12       PAIN AND SUFFERING
The section claims that road accidents are the only area to which pain and suffering costs
are ascribable. Wherever DALYs are estimated as ascribable to risk factors (and this is done
for alcohol, smoking and illicit drugs in Begg et al, 2007, which Collins and Lapsley later
reference), these can be converted using VSLY and a cost of illness matrix to the net value
of the loss of quality of life (Access Economics, 2008c). Again, the use of Bureau of
Transport Economics (2000) is out-dated and more recent Bureau of Transport and Road
Economics estimates (specified in Access Economics, 2008c) are available on their website.

2.13       WELFARE
Collins and Lapsley rightly distinguish real from transfer costs, but unfortunately ignore the
transfer costs and their associated deadweight losses (DWLs), saying that it is not possible
to measure them. This represents a major flaw as the distribution of costs will be
substantially affected by transfers and the total real costs are also likely to be
underestimated.

DWLs refer to the costs of administering welfare pensions and raising additional taxation
revenues. Although invalid and sickness benefits and forgone taxation are transfers, not real
costs (so should not be included in the estimation of total costs), it is still worthwhile
estimating them as that helps us understand how the total costs of abuse are shared
between the government, the individual and other financiers. In addition, health system and
other real expenditures paid for by government must also be financed by taxation. If the act
of taxation did not create distortions and inefficiencies in the economy, then transfers could
be made without a net cost to society. However, through these distortions, taxation does
impose a DWL on the economy. Access Economics measures DWLs as follows.

DWL is the loss of consumer and producer surplus, as a result of the imposition of a
distortion to the equilibrium (society preferred) level of output and prices. Taxes alter the
price and quantity of goods sold compared to what they would be if the market were not
distorted, and thus lead to some diminution in the value of trade between buyers and sellers
that would otherwise be enjoyed (Figure 2-2).




                                                                                            11
                                                            Collins and Lapsley report review


                                FIGURE 2-2: DWL OF TAXATION

             Price ($)

                                                                             Supply
                                             Deadweight Loss (cost
                                             of raising taxation
                                             revenue)


         Price plus Tax

                          Taxation Revenue

                 Price




                                                                          Demand
                                                                               Output
                          Actual Quantity            Potential Quantity
                          Supplied                   Supplied


There are two sources of lost tax revenue – potential income tax forgone and potential
indirect (consumption) tax forgone. The latter is lost because, as income falls, so does
consumption of goods and services. The average personal income tax rate used by Access
Economics in 2008 is 21.2% and the average indirect taxation rate used is 12.0%, based on
parameters for 2008 from the Access Economics macroeconomic model.

The rate of DWL used by Access Economics in 2008 is 27.5 cents per $1 of tax revenue
raised plus 1.25 cents per $1 of tax revenue raised for Australian Taxation Office
administration, based on Productivity Commission (2003) in turn derived from Lattimore
(1997), ie, 28.75% overall.

2.14        REVENUE IMPACTS OF DRUG ABUSE
This section rightly notes that the taxation raised from smoking exceeds its costs to
Government, so cigarettes are budget positive and so ‘smokers pay their way’ (page 17).
However, the same is true for alcohol, although this is not spelled out in this section, but can
be deduced from the Executive Summary final table, which shows that alcohol consumption
benefits the government sector by some $1.4 billion (2004-05 prices) per annum. Alcohol
taxes thus more than pay for the social costs of alcohol abuse, by a considerable
margin, each year.

2.15        RESEARCH, EDUCATION AND DRUG PROGRAM COSTS
The correct approach is to exclude any cost of prevention programs from the cost of abuse,
which appears to be what Collins and Lapsley (page 20) have done, although they also
exclude research costs:

       The present study follows Collins and Lapsley (2002) in adopting the criterion
       that expenditures on prevention of drug abuse and on research concerning



                                                                                              12
                                                           Collins and Lapsley report review


       appropriate interventions can be seen as discretionary, but that expenditures on
       the interventions themselves should not treated as discretionary.

The section concludes that the authors were unable to estimate public expenditures on
research and education, and on drug programs. Access Economics treats research costs in
relation to prevention and amelioration as part of the costs of abuse, since if abuse could be
eliminated (the counterfactual), research costs would no longer be required (we would know
how to effectively prevent abuse). Prevention costs are different, because they would still
presumably need to be incurred in order to achieve ongoing effective prevention.

2.16        NEW OR REVISED METHODOLOGIES AND DATA
By structuring their report with Section 2 addressing general methodological issues, Section
3 addressing new or revised methodologies and data, and the remaining sections addressing
findings re individual cost items, it is not transparent how findings are calculated (and there is
no cross-referencing of sections in the report or consistency in ordering, which jumps back
and forward between concepts and costs). A more transparent approach would have been
to present the method, data and finding for each cost estimate together, or at least to cross-
reference. However, again, for ease of following the original report, this critique addresses
the sections in the order in which they appear.

2.16.1      GST IMPACTS
This provides an interesting historical summary, albeit without highlighting a key finding, that
the taxation of alcohol increased from $3.37 billion in 1998-99 to $5.11 billion in 2004-05, a
52% increase (8.6% per annum), greatly exceeding population or consumption growth.

2.16.2      PRODUCTION LOSSES IN THE PAID WORKFORCE
The section on reduced workforce size appears provides no relative risk of mortality data, no
morbidity data sources and only a passing reference to ‘national accounts data’ (not
specifying the precise source). As such there is inadequate detail to replicate or validate this
method. This is a serious flaw.

For absenteeism, the productivity cost estimates are reportedly based on Bush and Wooden
(1994) combined somehow with Pidd et al (1996) and unspecified Australian Bureau of
Statistics (ABS) data. The lack of detail again prevents replication or validation. It is unclear
why Bush and Wooden is used, which relies on the 1989-90 National Health Survey (NHS)
data, when there are 2004-05 ABS NHS data now available that could have been used
(Access Economics uses that Confidentialised Unit Record File data in order to estimate
absenteeism costs in cost of illness analysis). The Pidd et al (1996) data are also old, much
larger than other estimates and suggest that some 2.7% to 7.4% of Australian absenteeism
is due to alcohol abuse, which seems high relative to its disease burden as estimated in
Begg et al (2007). This cost item may thus be overstated.

Presenteeism costs were unable to be estimated by Collins and Lapsley.

2.16.3      PRODUCTION LOSSES IN THE HOUSEHOLD SECTOR
Collins and Lapsley appear to apply (unspecified) drug-attributable morbidity and mortality
estimates to unpaid work data. As well as not being transparent how these calculations are
made, it is also not clear that this is conceptually sound. For example, while someone may



                                                                                                13
                                                                 Collins and Lapsley report review


be absent from work due to alcohol-related morbidity, they may simply defer their household
tasks until they are feeling better.

Collins and Lapsley also use replacement cost valuation, which is higher than opportunity
cost valuation. Access Economics recommend use of opportunity cost valuation, which tends
to be more conservative as well as being more conceptually accurate (since it is unlikely that
a replacement worker is actually paid and brought in to undertake domestic tasks for
abusers). This cost item is thus likely to be overestimated.

2.16.4       HEALTH
No data sources are specified for hospital costs in the body of the text (Section 3.4 of Collins
and Lapsley 2008a) and there is no cross-reference to Appendix A, where there is more
detail on the casemix data used. It appears medical costs are based on a different data set
from hospital costs (AIHW, 2006) and unclear why the more recent AIHW (2007) data would
not be used or, indeed, the more appropriate AIHW (2005) dataset, which would have
permitted detailed breakdown of actual alcohol-attributable costs by alcohol-related condition
and by type of health system expenditure, using a more comprehensive top-down attribution.
This would have been far superior to the very arbitrary attributions used (eg, there is no
evidence basis provided for the assumption that ‘at least 15% of all nursing home admissions
have drug-related morbidities.’ This is also misrepresented in the text as a conclusion, rather
than an assumption. There is also insufficient detail provided for ambulance and
pharmaceutical costs to enable replication or validation, and pharmaceutical costs could
also, alternatively, have been based on the more reliable official top-down AIHW data. Had
this been done, it would also have been possible to incorporate estimates of over-the-counter
pharmaceuticals, as well as pathology and imaging, allied health and other health
expenditures.

Access Economics recommends use of the AIHW health data to estimate costs attributable
to risk factors, as estimated for example in Access Economics (2008b) in relation to the cost
of obesity.

Failing that, actual data sources should be clearly specified so that the reader can replicate
calculations, identify strengths and weaknesses/uncertainties and potentially have some
degree of confidence in the methods and results. This is not possible in relation to Collins
and Lapsley (2008a).

2.16.5       ROAD ACCIDENTS
This cost item is again based on the out-dated Bureau of Transport Economics (2000) report
(see Sections 2.11 and 2.12).5 It would be preferable to use the alcohol-attributable fractions
for motor vehicle accidents from Begg et al (2007) combined with top-down unit cost data for
road accidents from Austroads (2007, 2006, 2003).

2.16.6       FIRES CAUSED BY SMOKING
This section is not relevant to this critique.




5
       More     recent      publications    in     relation     to     road      safety  are   available   at
http://www.infrastructure.gov.au/roads/safety/publications/publications_list.aspx?page=3&meth=DESC&sort=date
&mode=Road#sorted


                                                                                                           14
                                                          Collins and Lapsley report review


2.16.7      RESOURCES USED IN THE CONSUMPTION OF ILLICIT DRUGS
This section is not relevant to this critique.

2.16.8      LITTER
The costs of litter are not estimated by Collins and Lapsley.




                                                                                          15
                                                            Collins and Lapsley report review




3. ANALYSIS OF COST ITEMS
This section critiques individual cost items in relation to common problems such as
omissions, understating, overstating, double counting, calculation errors, and appropriate use
of available data and literature – ie, whether they are the best, most recent and most
appropriate available. Underlying cost calculations are epidemiological estimates of
incidence, prevalence and mortality, and these estimates are also reviewed.

3.1         EPIDEMIOLOGICAL ESTIMATES
Epidemiological background is based on demographic work by JH Pollard at Macquarie
University (who has written the relevant Appendix A) and aetiological work by J Codde at the
university of Western Australia (who has written the relevant Appendix C). The work does not
appear to have been adequately quality assured for consistency, as there are
inconsistencies (eg, use of mid-year 2005 population data in Appendix A and mid-year 2004
population data in Appendix C).

A minor matter in Appendix A is that, while it is noted that no attempt was made to quantify
births that did not take place because of lives lost through drug usage, it is not noted that,
equally, no attempt is made to quantify births that do take place as a result of, say, alcohol
usage (which reduces inhibition).

Another minor matter in Appendix A is that the projection program method description on
page 97 only estimates (a) no illicit drugs (b) no illicit drugs or alcohol and (c) no illicit drugs
or alcohol or smoking. It would appear there is a need to run six not three simulations in
order to derive columns 3-8 of Tables 51 to 53.

A far more substantial concern in Appendix A is that flagged on page 98, that attributable
fractions are assumed unchanged between 1947 and 2005. Pollard describes this as a
‘serious limitation’, which it is. It is also unnecessary, since it would have been possible to
use a forward-looking rather than retrospective approach and use the Begg et al (2007)
forward projections instead.

In relation to Appendix C, a minor concern is the use of ABS data on deaths, which can
suffer from miscoding problems, rather than higher quality data from clinical studies. Another
minor concern is that the heading ‘morbidity data’ actually refers to hospital data, which is not
technically equivalent. Another concern is that it is unclear which NHMRC recommendations
on alcohol consumption are implied on page 115 (there is no reference and the 2008 version
differs from the more widely accepted 2001 version, see Access Economics 2008a).

A more major concern, however, is that even with the end-notes to the Appendix, it is unclear
exactly how or why the various data sources have been massaged in order to produce
Tables 61 through 67. Page 115 implies that in some unspecified cases (for unspecified
reasons) personal communications with authors of Begg et al (2007) may have been used
rather than published fractions. When summarising the attributable fractions in the main
body of the text (Table 8 on page 34), they are misrepresented as ‘conditions attributable to
drug abuse’ when a number (some quite major conditions) are in fact protected against
through alcohol use – ischaemic heart disease, supraventricular cardiac dysrhythmias
(omitted from Table 9 also), cholelithiasis, heart failure in males, stroke and hypertension in
females.




                                                                                                  16
                                                          Collins and Lapsley report review


Discounting to take account of causal interaction by 2.18% as described in Section 4.1 of
Collins and Lapsley (2008a) appears sensible.

Section 4.2 correctly notes that the appropriate comparator is abstinence, not low-risk
drinking (although it uses numbers estimated historically as being ‘preventable’, which is at
odds with elsewhere in the report where it is correctly noted that these findings cannot be
interpreted as representing ‘avoidable’ costs since they are historical, not forward-looking).

Section 4.3 notes the major changes in attributable fractions for alcohol, but this uncertainty
is not dealt with in the report, eg, through sensitivity analysis.

3.2         CRIME COSTS
In general, the methods used to estimate crime costs in Section 5 and Appendix B (written by
T Makkai of the Australian Institute of Criminology and J Temple of the Australian National
University) of Collins and Lapsley (2008a) suggest there is substantial risk of overestimation
for this cost item. However, this is one of the smallest alcohol cost components.

The calculation of attributable fractions for alcohol is particularly problematic for a number of
reasons. Appendix B notes that: ‘There are significant limitations in the current national
crime and justice collections which do not allow us to produce attributable fractions that could
be deemed as being ‘true’.’ The four limitations mentioned included that, due to data
problems, it was not possible to ascertain the extent to which offences were caused by
alcohol intoxication (although estimates were nonetheless made).

The self-report surveys use were the Drug Use Careers of Offenders (DUCO) and the Drug
Use Monitoring in Australia (DUMA), which asked detainees whether they were using
substances rather than whether they were intoxicated, and if the respondent notes use the
crime is assumed to be drug-related. As such the surrogate measure may substantially
overestimate the attributable fraction and Appendix B (page 105) states in relation to the
measures that: ‘In all likelihood they overestimate levels of intoxication, particularly for
alcohol.’

As well as there being ‘no measure or even approximation for causal behaviour for alcohol’ in
DUMA and taking ‘the most liberal estimate’ (page 105), in addition the sample is skewed
towards the more serious crimes, another source of upside bias. The authors summarise
(page 106) that ‘it is likely that our results, even the lower bound estimates, are liberal
estimates’. However, it is the DUMA attributable fractions on which the policing and criminal
court costs are based in Collins and Lapsley (2008a). Moreover, despite this statement in the
appendix and the one quoted in the paragraph above from page 105 re overestimation, the
main body of the text (page 47) states that there have been ‘conservative estimation
techniques adopted in the research’.

In addition, neither DUMA nor DUCA were nationally representative samples and it was
acknowledged ‘there is considerable variation in drug markets between NSW and the other
jurisdictions’ (page 108). Small sample size was also an issue for five of the eight crime
categories in the DUCO data (used to estimate prison costs) – confidence intervals for
‘driving under the influence’ and ‘disorder’ for females were thus ‘not available’ and the data
(Table 57) showed 100% being ‘high’ and none being ‘drunk’ (recalling that the measure in
fact is use/dependency not intoxication).
      Conversely, page 42 states that in the calculations, all drink-driving is assumed fully
      attributable to alcohol (this is also arguable, since if someone has had a lack of sleep
      and also a drink and falls asleep and has an accident, for example, it might be better


                                                                                               17
                                                           Collins and Lapsley report review


      methodologically to attribute some of the causation to sleep deprivation and some to
      alcohol consumption).
      Clearly data with such high standard errors and obvious anomalies are not reliable and
      table notes in Appendix B state that the estimates should be treated ‘with caution’. It is
      not clear how the uncertainty has been dealt with, however.

The attributable fractions that are used in Collins and Lapsley (2008a) in relation to crime
should thus be treated with considerable caution. It is a concern that there is no downside
sensitivity analysis around these parameters, even given the uncertainty, lack of
conservatism and cautions detailed by the contributing authors.

There are many steps involved in estimating the costs of $1.7 billion for crime due to alcohol
abuse presented in Table 16. The structure of the document is also quite difficult to follow,
with a whole Section (5) for crime costs, then another sub-section (6.1) in a chapter titled
‘Some disaggregated costs’, some elements also in Appendix B and others in other sections.
It is not possible to replicate the calculations since insufficient information is provided across
the various sections of the report in order to do so. Figures of $1.4 billion are also presented
described as ‘not elsewhere included’ in Table 16 and, in the Executive Summary page xii,
the figure provided for this item is $1.61 billion (just the tangible component of Table 16). It is
not clear why the upper bound estimate is lower than the lower bound estimate for police and
criminal court costs in Table 17 (page 50).

3.3         HEALTH COSTS
Health system cost findings are presented in Section 6.2.

Data and methodological concerns in relation to the estimation of health costs were
presented in Section 2.16.4 – most seriously that a number of different partial and incomplete
data sources were used and blended, when a top-down data source was available (AIHW,
2005), which would have permitted detailed breakdown of actual alcohol-attributable costs by
alcohol-related condition and by type of health system expenditure including over-the-counter
pharmaceuticals, pathology and imaging, allied health and other health expenditures.

Access Economics recommends use of the AIHW health data to estimate costs attributable
to risk factors, as estimated for example in Access Economics (2008b) in relation to the cost
of obesity.

It is not possible to have confidence in the health cost results presented in Table 18 due to
the data/methodological flaws and to the fact that there is inadequate information presented
in the report to replicate the calculations.

3.4         PRODUCTIVITY COSTS
As with health system costs, our conclusion is that there are too many data and
methodological weaknesses and calculation gaps to have confidence in the productivity cost
estimates presented in Table 30. The text in Section 6.4 is quite inadequate given that this is
calculated by Collins and Lapsley to be the largest alcohol cost item at some $5 billion and
nearly half their estimate of total tangible costs. There is no cross-referencing to Sections
2.4, 3.2 and 3.3 where other aspects of the productivity cost estimation process can be
found, although in total these are inadequate (again) to replicate the calculations.




                                                                                                 18
                                                           Collins and Lapsley report review


3.5         ROAD ACCIDENT COSTS
As with crime, health and productivity costs, the road accident cost estimates presented in
Table 31 of Collins and Lapsley (2998a) are not replicable and suffer the serious flaw
observed earlier in this critique that they are based on old Bureau of Transport Economics
estimates when later estimates (Austroads, 2007; 2006; 2003) based on newer methods are
available (see Sections 2.11, 2.12 and 2.16.5).

3.6         FIRE COSTS
Fire costs are estimated in the report, but are not relevant to this critique, as they relate only
to smoking.




                                                                                                19
                                                                      Collins and Lapsley report review




4. CONCLUDING SECTIONS OF COLLINS AND LAPSLEY
   (2008A)
Section 7 (‘Aggregate results’) provides no more confidence in the cost estimation process,
but just cuts the findings different ways. Section 7 again reflects the cumbersome structure
of the report – total costs should be presented as a cross-tabulated matrix in one chapter
rather than in piecemeal fashion across two chapters (Sections 6 and Section 7).

There is no supporting data underlying Section 7.2 that provides evidence or clarity for the
attribution process used, so it is not possible to replicate the costs by bearer presented in
Table 36 (confusingly referred to as ‘incidence’ of tangible costs when in fact the costing
remains prevalence-based). Moreover, a main finding, that business bears over half the
tangible costs of alcohol abuse, is not mentioned in the discussion.

Similarly, in Section 7.3, the tables presented are not linked to the methods in a way that
enables replication or confidence in the calculations. It is not immediately clear why the net
impact on the federal and state budgets in Table 41 of $1.4 billion is different from the cost to
the Government sector of $2.9 billion presented in Table 36, for example.

Section 7.5 (the sensitivity ‘estimates’), however, is the most disappointing in the whole
analysis, containing not one quantitative finding. Of the six qualitative dot-points, three relate
to upside estimates for crime (already potentially too high, see Section 3.2 of this critique),
the tiny (almost irrelevant) exclusion of healthcare savings (only $74 million in $10.8 billion)
and the third being the problematic extreme parameter that 50% of alcohol consumption is
abusive (again, even the base case 30% is likely to be too high – see Section 2.3).

The penultimate dot-point point on absenteeism is not quantified in Section 3.2.2, to which
the reader is referred. The last dot-point relates to transfer costs, not real costs. Finally, there
is one presumably downside sensitivity estimate in the dot-point which refers the reader to
Table 33 and the possible exclusion of resources used in the production and distribution of
abusive consumption – since this cost is $1.69 billion for alcohol, the conclusion is that the
total estimate for alcohol is potentially overestimated by 18.4%, although this is not stated by
Collins and Lapsley (Section 2.7 of this critique concludes this cost item should be excluded
for alcohol).

We recommend that adequate sensitivity analysis is used in future estimates of the cost of
alcohol abuse that shows the variation in results dependent on variation in parameters, using
software tools such as @RISK to assess confidence intervals,6 and with due attention to
downside risks as well as upside risks.

Section 8 rightly concludes that trying to compare alcohol costs presented in this report with
those previously estimated by Collins and Lapsley is not sensible due to the different
methods employed in their different reports over time. However, the conclusion is based on
new cost methods being ‘better’ (as well as higher) for absenteeism, which is not clear (as


6
  @RISK undertakes monte carlo simulation, a well known technique, to determine the sensitivity of outputs to
changes in key inputs. It iteratively replaces numbers attached to key parameters (inputs) with random numbers
drawn from a specified distribution, where the type of distribution, the upper and lower bounds on the distribution,
and the number of iterations are chosen by the analyst. The monte carlo simulation provides a distribution around
chosen outputs (such as the financial and total costs) from which sensitivity of outputs to inputs can be
determined and graphically presented.


                                                                                                                  20
                                                          Collins and Lapsley report review


concluded in Section 2.16.2) due to reliance on the oldest of the ABS NHS data when
alternative recent data are available.

Section 9 summarises areas where future research would be desirable.

We agree with the recommendation to re-estimate the results using the human capital
approach (now done in Collins and Lapsley, 2008b), again bringing into question why using
the ‘demographic’ approach was not just a waste of resources and effort. We would also
recommend, however, that the spurious use of ‘human capital’ be avoided except in relation
to productivity costs and, instead, appropriate and accurate forward-looking terminology be
adopted to describe such prevalence costing.

It is not clear if the second recommendation is to re-estimate or not re-estimate the 1998-99
alcohol cost estimates using the methods in Collins and Lapsley (2008a). We would
recommend that, instead, a time series of comparable estimates is calculated based on a
sound methodology.

Section 9.3 highlights the problems with the data underlying the estimated cost of crime, and
the need to improve these. It is not clear why the estimates would be improved by more
detailed aetiological fractions (beer, wine, spirits) suggested in Section 9.7 – nor how such
data could be collected in a cost-effective, realistic way without enormous overlap between
the types. The problems identified in Sections 9.8, 9.9 and 9.11 would not exist if our
recommended approach to estimating health system costs were adopted. The costs of litter
(9.10) are likely to be tiny and we suggest not worth the effort of attempting to estimate them.
This may also be true in relation to 9.12, research and education expenditures. For 9.13, the
claim that the most recent data is for 1996 is not correct and, instead, the available
Austroads data should have been used.

We agree with the final suggestion (9.14) that the data needs are reviewed and an expert
advisory panel convened, also to review the Collins and Lapsley methods, given the
weaknesses identified in this critique.




                                                                                              21
                                                             Collins and Lapsley report review




REFERENCES
Access Economics (2008a) Injury risk and drinking: a technical critique, Report for the
     Distilled Spirits Industry Council of Australia, March, Canberra, available at
     http://www.accesseconomics.com.au/publicationsreports/showreport.php?id=160&searchfor=20
     08&searchby=year

Access Economics (2008b) The Growing Cost of Obesity in 2008: three years on, Report for
     Diabetes       Australia,     August,        Canberra,          available        at
     http://www.accesseconomics.com.au/publicationsreports/showreport.php?id=172&searchfor=2008&search
     by=year

Access Economics (2008c) The Health of Nations: The Value of a Statistical Life. Report for
     the Office of the Australian Safety and Compensation Council (OASCC), January,
     Canberra,                               available                                   at
     http://www.accesseconomics.com.au/publicationsreports/showreport.php?id=156&searchfor=2008&search
     by=year

Access Economics (2005) The shifting burden of cardiovascular disease Report for the
     National  Heart    Foundation    of    Australia,   Canberra,    available   at
     http://www.accesseconomics.com.au/publicationsreports/showreport.php?id=15&searchfor=2005&searchb
     y=year

Australian Institute of Health and Welfare (2007) Health Expenditure Australia, 2005-06,
     AIHW , Cat No HWE 37, October, Canberra.

Australian Institute of Health and Welfare (2006) Health Expenditure Australia, 2004-05,
     AIHW, Cat No HWE 35, September, Canberra.

Australian Institute of Health and Welfare (2005) Health system expenditure on disease and
     injury in Australia, 2000-01, AIHW, 2nd edition, Cat No HWE 28, April, Canberra.

Austroads (2007) ‘Guide to Program Evaluation Part 4: Project Evaluation Data’ (Second
     Edition), Austroads, Project No TP1050 Publication No AGPE04/07, May.

Austroads (2006) ‘Austroads Technical Report: Update of RUC Unit Values to June 2005’
     Austroads, Project No TP1166 Publication No AP-T70/06, September.

Austroads (2003) ‘Economic evaluation of road investment proposals: improved methods for
     estimating Australian unit crash costs’ Austroads, Project No BS.EN.511 Publication
     No AP-R238/03.

Begg S, Vos T, Barker B, Stevenson C, Stanley L, Lopez AD (2007) The burden of disease
     and injury in Australia 2003, AIHW PHE 82, April, Canberra.

Bureau of Transport Economics (2000) Road Crash Costs in Australia, BTE Report 102,
     Canberra.

Bush R, Wooden M (1994) Smoking, Alcohol and Absence in the Workplace: an analysis of
     the 1989/90 National Health Survey, National Institute of Labour Studies, Flinders
     University, South Australia, March.

Collins DJ, Lapsley HM (2008a) The costs of tobacco, alcohol and illicit drug abuse to
      Australian Society in 2004/05, Report for the Department of Health and Ageing.


                                                                                                     22
                                                          Collins and Lapsley report review


Collins DJ, Lapsley HM (2008b) The avoidable costs of alcohol abuse in Australia and the
      potential benefits of effective policies to reduce the social costs of alcohol, Report for
      the Department of Health and Ageing.

Collins DJ, Lapsley HM (1996) The social costs of drug abuse in Australia in 1988 and 1992,
      Report for the Department of Health and Aged Care.

Department of Health and Aged Care (2001) National Alcohol Strategy: A plan for action
     2001 to 2003-04, published for the National Drug Strategy.

English D, Homan C et al (1995) The Quantification of Drug Caused Morbidity and Mortality
      in Australia, Commonwealth Department of Human Services and Health, Canberra.

Hillman DR, Scott Murphy A, Antic R, Pezzullo ML (2006) ‘The Economic Cost of Sleep
      Disorders’ Sleep 29(3):299-305.

Laslett A, Donath S, Dietze P (2002) ‘Long term consequences of alcohol consumption’ in
      National Alcohol Research Agenda, Commonwealth Department of Health and Ageing,
      Canberra.

Lattimore R (1997) Research and Development Fiscal Incentives in Australia: Impacts and
      Policy Lessons, OECD Conference on Policy Evaluation in Innovation, 26-27 June,
      Paris, 81:574-7.

National Health and Medical Research Council (2001) Australian Alcohol Guidelines: Health
     Risks and Benefits, NHMRC, Canberra.

Pidd KJ, Berry JG, Roche AM, Harrison JE (2006) ‘Estimating the cost of alcohol-related
     absenteeism in the Australian workforce: the importance of consumption patterns’
     Medical Journal of Australia, 185(11/12).

Pizzi LT, Lofland JH (2005) Economic Evaluation in US Health Care Principles and
      Applications, Jones and Bartlett Publishers, ISBN 0763727466, 9780763727468.

Productivity Commission (2003) ‘Evaluation of the Pharmaceutical Industry Investment
     Program’ Research Report, AusInfo, Canberra.

Ridolfo B, Stevenson C (2001) The quantification of drug-caused mortality and morbidity in
      Australia, 1998, Australian Institute of Health and Welfare, Canberra.

Single E, Collins D, Easton B, Harwood H, Lapsley H, Kopp P, Wilson E (2003) International
      Guidelines for Estimating the Costs of Substance Abuse, World Health Organization.

Single E, Ashley M, Bondy S, Rankin J, Rehm J (1999) Evidence regarding the level of
      alcohol consumption considered to be low-risk for men and women, National Health
      and Medical Research Council, Canberra.

Taylor HR, Pezzullo ML, Keeffe JE (2006) ‘The economic impact and cost of visual
     impairment in Australia’ British Journal of Ophthalmology 90:272-275.




                                                                                              23
                                             23 December 2008




Collins and Lapsley report review




Report by Access Economics Pty Limited for

National Alcohol Beverage
Industry Council
                                                                                      Collins and Lapsley report review



CONTENTS
Executive summary...............................................................................................................i
1.      Background .................................................................................................................1
2.      The social costs of alcohol.........................................................................................2
3.      Methodological and data issues ................................................................................4
        3.1   Broad conceptual and methodological issues .....................................................................4
           3.1.1    Purpose of the analysis .............................................................................................4
           3.1.2    Concept of a feasible minimum .................................................................................5
        3.2 Estimating the feasible minimum .........................................................................................5
           3.2.1    The Arcadian Normal approach.................................................................................6
           3.2.2    Exposure based comparators....................................................................................6
           3.2.3    The distributional approach .......................................................................................7
           3.2.4    Evidence on the effectiveness of interventions .........................................................7
        3.3 Effectiveness of interventions ..............................................................................................8
           3.3.1    Method for deriving the interventions reviewed .........................................................8
           3.3.2    Ambit claims ..............................................................................................................9
           3.3.3    Other issues raised in Chapter 4 ...............................................................................9
4.      Interventions, conclusions and recommendations ................................................10
        4.1   Intervention choices ...........................................................................................................10
           4.1.1     Alcohol taxation .......................................................................................................10
           4.1.2     Bans on alcohol advertising.....................................................................................12
           4.1.3     Measures to reduce drink driving ............................................................................12
           4.1.4     ‘Brief’ interventions for reducing hazardous consumption.......................................13
        4.2 Other interventions (benefits not evaluated) ......................................................................14
           4.2.1     Control of drinking environments.............................................................................14
           4.2.2     Alcohol ignition locks ...............................................................................................14
           4.2.3     Guidelines for low-risk drinking................................................................................14
           4.2.4     Standard drinks labelling and health warnings ........................................................14
           4.2.5     Other interventions ..................................................................................................14
        4.3 Summary, conclusions and recommendations ..................................................................15
References..........................................................................................................................16

FIGURES
Figure 4-1: DWL of taxation for healthy drinkers                                                                                            11

TABLES
Table 4-1: Summary of estimated benefits (reduction in social costs), $m                                                                    15
While every effort has been made to ensure the accuracy of this document, the uncertain nature of economic data, forecasting
and analysis means that Access Economics Pty Limited is unable to make any warranties in relation to the information
contained herein. Access Economics Pty Limited, its employees and agents disclaim liability for any loss or damage which
may arise as a consequence of any person relying on the information contained in this document.
                                              Collins and Lapsley report review




GLOSSARY OF ACRONYMS
ABS           Australian Bureau of Statistics
AIHW          Australian Institute of Health and Welfare
BAC           blood alcohol concentration
B/C (ratio)   benefit/cost (ratio)
CBA           cost benefit analysis
CEA           cost effectiveness analysis
DALY          disability adjusted life year
DWL           deadweight loss
ICER          incremental cost effectiveness ratio
GDP           gross domestic product
NHMRC         National Health and Medical Research Council
NHS           National Health Survey
NPV           net present value
OBPR          Office of Best Practice Regulation
QALY          quality adjusted life year
SMR           standardised mortality rate
                                                          Collins and Lapsley report review




EXECUTIVE SUMMARY
This report reviews the methodology, assumptions and conclusions in the monograph:
Collins DJ, Lapsley HM (2008b) The avoidable costs of alcohol abuse in Australia and the
potential benefits of effective policies to reduce the social costs of alcohol, Report for the
Department of Health and Ageing.

The critique has focused on three main areas: (1) whether the estimate of social costs of
alcohol abuse that can be reduced is in fact reasonable; (2) whether methods and data
represent a correct approach to the primary research question; and (3) whether the
intervention choices, conclusions and recommendations are appropriate in relation to the
potential cost reductions (benefits) from a policy perspective.

Is the estimate of the social costs of alcohol abuse reasonable?

Chapter 2 of Collins and Lapsley (2008b) is based on Collins and Lapsley (2008a)
comprising an estimated $10.8 billion in ‘tangible’ (financial) costs and $4.5 billion in loss of
healthy life (totalling $15.3 billion). Our previous critique of Collins and Lapsley (2008a)
raised serious concerns in relation to major conceptual and methodological problems with
this estimate (see Section 2 of this report for a summary of these problems).

      We concluded that the costing analysis displayed a paucity of supporting
      evidence, incomplete references for claims, lack of transparency in calculation
      processes, and singularly poor methodology and data use. There was apparent
      upward bias in most of the cost estimates and uncertainty was not dealt with due
      to inadequate sensitivity analysis. Using a best possible approach would exclude
      one cost item, making the overall cost of alcohol abuse at least 18% lower. The
      other findings of the report should be viewed with substantial scepticism.

      We agree with the final suggestion (Section 9.14 of Collins and Lapsley, 2008a)
      that data needs are reviewed and an expert advisory panel convened, also to
      review methods, given the serious weaknesses identified.

Are methods and data appropriate to the purpose?

The title of the report suggests that the purpose of the analysis is to identify the potential
benefits of effective policies to reduce alcohol abuse. The most common approach to
achieve such an objective is the use of evaluation techniques such as cost benefit analysis
(CBA) and cost effectiveness analysis (CEA). However, neither of these techniques is
adopted in the report and, notably, the costs of the interventions are not estimated. To
estimate gross benefits without estimating costs is likely to provide spurious
conclusions – if costs outweigh benefits there may be no net benefit in any of the
interventions reviewed. It is thus dubious how the information can assist in developing
evidence-based strategies.

The concept of a ‘feasible minimum’ is unusual in economics. Optimal social outcomes are
achieved when marginal social costs are equated to marginal social benefits, and a more
traditional incidence or prevalence based CBA/CEA would be more useful for policy makers.
It is quite conceivable that Australian society is currently operating at the social optimum for
harm minimisation from alcohol abuse – or has even intervened too much – yet these
possibilities are not considered in the report.


                                                                                                i
                                                                       Collins and Lapsley report review


Rather, a five-step process (requiring three more stages of analysis) is suggested by Collins
and Lapsley. It is not clear why one would undertake such a prolonged process rather than
just to do conventional CBA/CEAs in the first instance, as recommended by the CBA Unit at
the Australian Government’s Office of Best Practice Regulation (OBPR), which provides
guidance on how to undertake studies to provide economic evidence for policy inputs – see
http://www.finance.gov.au/obpr/cost-benefit-analysis.html Notably, this site references
Access Economics (2008c) in CBA best practice for valuing potential healthy life gains from
interventions.

The ‘feasible minimum’ concept, in contrast, is ascribed only to references where Collins and
Lapsley are co-authors (Single et al, 2003; Collins and Lapsley, 20061) in the context of work
they did for Health Canada ie, there is no other literature cited that uses this approach. The
‘four methods’ they devised as possible mechanisms for estimating the ‘feasible minimum’,
while having interesting names, have fundamental flaws. The so-called ‘Arcadian Normal’
and ‘exposure based comparator’ methods selectively include/exclude countries and ignore
demographic/ethnic, climate, income, cultural, structural (policy, regulatory), pricing or other
potential factors that might result in different average alcohol consumption patterns across
countries or differences in their age-standardised mortality rates for alcohol-attributable
conditions. There is no information provided to ascertain why the poorly-summarised
‘distributional approach’ was discarded by Collins and Lapsley.

The fourth and final approach, which looks at the effectiveness of interventions, represents
the conventional means where efficacy parameters are derived from the literature, as key
parameter inputs into CBA and CEA studies. However, Collins and Lapsley (2008b) instead
use the parameters indirectly to estimate ‘avoidable costs’, thereby creating a problem of
time lags (which they suggest may be addressed in their next report – third in the series of
five). However, well-constructed conventional CBAs and CEAs estimate net present values
(NPV) over future timeframes (eg, 10 years, 25 years or longer), bypassing such problems
(Access Economics, 2008c).

         The methods (‘feasible minimum’, ‘Arcadian normal’, ‘exposure-based
         comparators’) selected are not appropriate to the purpose. Conventional CBA
         and/or CEA would have been preferable, as recommended by the OBPR for
         example, for evaluating economic evidence in relation to proposed interventions.
         The methods and numbers selected to derive the conclusion that it would be
         possible to reduce the social costs of alcohol over time by about a half, have no
         valid evidence basis, in our view. Conceptually, by preventing all incident cases
         of alcohol abuse, over time one could eliminate all its social costs.

How were the interventions chosen?

There is no discussion of the literature review process, search protocols used or the basis for
inclusion or exclusion of particular studies, although this is standard practice. Certainly the
literature is much broader than the eight studies mentioned. Even then, only one study is
used to provide evidence, with no good reason given for its use. Moreover, this particular
study does not provide objective comparisons of interventions such as a league table of
ICERs2 or benefit/cost (B/C) ratios (as would be expected), but instead provides a subjective


1
    The latter paper as referenced by Collins and Lapsley (2008b) relates to smoking, not alcohol.
2
  CEAs use the incremental cost effectiveness ratio (ICER) – a common and well-accepted evaluation technique
in health economics worldwide and across the Australian public and private sectors. The ICER metric measures
dollars per disability adjusted life year (DALY) averted or per quality adjusted life year (QALY) gained.


                                                                                                           ii
                                                                   Collins and Lapsley report review


1/2/3-star rating system. Included interventions have been tried historically with some degree
of gross (not net) benefit, an approach likely to embed historical mistakes in future policy.

      Methods used to select so-called effective interventions are problematic and out
      of line with standard practice. Unreferenced claims are made that certain
      interventions are ‘world best practice’, ‘proven-effective’, or ‘very cost-effective’,
      without evidence. Important issues are raised such as the treatment of
      combination policies and the risk that some policies will reduce benefits as well
      as costs of consumption – yet these are ignored, thus overstating the potential
      benefits of the selected interventions.

Are the intervention conclusions appropriate?

Excise taxation is a blunt instrument that causes welfare losses for healthy consumers while
at the same time not adequately preventing harmful use. Collins and Lapsley (2008b) ignore
the deadweight losses (DWLs) of higher taxation on healthy drinkers, and conclude that
differential alcohol taxes provide scope for targeting tax increases on hazardous and harmful
consumption. Collins and Lapsley provide no evidence for this claim. Theory and evidence
suggest that increasing taxation is unlikely to be cost-effective when large adverse impacts
(some 28.75% of the tax value) on a large population (the vast majority of consumers are
healthy drinkers) from such a tax are considered (eg, very large DWLs and smaller addiction
swapping impacts). The comparison with Norway as the lowest per capita consumer of
alcohol, and the US and Italy as wine producers, on which to base a target for Australia is
fairly random, and suffers the plethora of issues raised in relation to the ‘Arcadian Normal’
and ‘exposure based comparator’ approaches to measuring the ‘Feasible Minimum’. A
seminal low point of the report is the unreferenced claim that reducing alcohol consumption
will proportionately reduce alcohol abuse. There is thus no sound basis for the conclusion
that unspecified levels of tax increase would generate social benefits of $2.2 to $5.9 billion.

Bans on alcohol advertising are recommended as effective despite conflicting evidence,
on the basis of a few selected studies (bereft of meta-analysis techniques). Highly simplistic
(‘back of envelope’) methods are used to estimate efficacy from these limited findings,
lacking distributional or sensitivity analysis. No confidence can be placed in the ‘savings’ so
estimated of $2.5 billion (partial) to $1.7 billion (full bans, ‘additional’). There is no discussion
of what a partial ban might consist of in Australia.

Two measures to reduce drink driving – more intensive enforcement of current blood
alcohol concentration (BAC) levels and lowering the legal BAC level – are presented as
efficacious based on one experience from Sweden, ignoring the body of evidence. There is a
problematic association with accident reduction for the first measure and the second
measure has only a 9% impact on alcohol-attributable accidents for a 60% reduction in BAC
across the whole Australian population. No confidence can be placed on the finding that
both measures (whose benefits are added together simply) is in fact $1.2 billion.

Brief interventions have better evidence for efficacy, but Collins and Lapsley make errors in
estimating the percentage reduction in deaths and errors in applying the reduced mortality
rate to total costs. As such, the estimated benefits of $5.8 billion are overstated.



Benchmarks for cost effectiveness have been developed in different sectors by different organisations – for
example the World Health Organization, defining cost effective interventions as those that costs less than three
times GDP per capita per QALY gained or DALY averted and very cost effective interventions as those that costs
less than GDP per capita.


                                                                                                               iii
                                                        Collins and Lapsley report review


     No confidence can be placed in any of the estimates regarding the supposed
     cost savings from interventions, given the numerous methodological flaws.

Five other interventions are presented that apparently are also deemed ‘likely to be cost-
effective’ by Collins and Lapsley (2008b:30), although it is not clear on what basis,
particularly given they say the benefits are not quantifiable. Quantification is a logical
necessity to demonstrate cost effectiveness. The interventions are:
     control of drinking environments (only evidence in favour of greater controls is
     presented);
     alcohol ignition locks (evidence presented is again uni-directional and claims of net
     benefit are not validated);
     guidelines for low-risk drinking (information such as the National Health and Medical
     Research Council (NHMRC) draft 2008 guidelines is only ‘useful’ if accurate, and it is
     unlikely that these guidelines are – Access Economics, 2008a);
     standard drinks labelling and health warnings on drinks containers (no evidence is
     referenced in relation to any claim made in this section); and
     other interventions – we agree further evidence is required in order to evaluate the
     relative cost effectiveness of programs such as intensive alcohol detoxification, school-
     based education and workplace interventions. Prospective CBA and CEA studies
     would be very valuable in relation to these and all the other interventions considered in
     the report.

Are the recommendations appropriate?

In our view, there are fundamental flaws in the methods used in this Collins and Lapsley
(2008b) report and hence in its recommendations. There is also no disclaimer on the report
that the findings do not represent the views of the Australian Government. We recommend
the matter is reviewed by the Department of Health and Ageing and that genuine evidence
only informs policy consideration of this important issue.


Access Economics
December 2008




                                                                                             iv
                                                                     Collins and Lapsley report review




1. BACKGROUND
The purpose of this report is to undertake a critique of the Collins and Lapsley (2008b) work:
‘The avoidable costs of alcohol abuse in Australia and the potential benefits of effective
policies to reduce the social costs of alcohol’ (hereafter the Avoidable Costs report) as a
follow up to our earlier analysis of the Collins and Lapsley (2008a) work ‘Costs of tobacco,
alcohol and illicit drug use to Australian society in 2004/05’3 (the Social Costs exercise).

While the earlier study (Collins and Lapsley, 2008a) estimated the cost of alcohol abuse in
Australia in 2004-05 as over $15 billion, this monograph focuses on ‘the extent to which
these costs could be reduced by the implementation of appropriate public policy interventions
[and] indicates that it should be possible, over a period of time, to reduce these costs by
approximately half.’ (Collins and Lapsley, 2008b:ix)

This report is structured as follows:
         Section 2 briefly and summarily critiques the social costs in Chapter 3 of Collins and
         Lapsley (2008b), with reference to the costing issues raised in the Social Costs
         exercise.
         Section 3 summarises general methodological and data issues, where Sections 1, 2
         and 4 cover such matters, together with Appendix A (there are no other appendices).
         Section 4 reviews the intervention choices, conclusions and recommendations
         (Sections 5-9) with special comment in relation to:
         3     the potential dollar reduction in Australian social costs (based on per capita
               consumption rates as being the achievable parameter in Norway, USA or Italy);
         3     the potential benefits of specific policy interventions in terms of reductions in
               social costs; and
         3     the overall policy implications concluded as a result of the Collins and Lapsley
               Avoidable Costs work.

Access Economics was selected as the reviewer for this project due to extensive experience
and world expert status in cost of illness modelling and health evaluation, particularly cost
benefit and cost effectiveness analysis. The firm’s health economics group has conducted
nearly forty cost of illness studies across five continents, many of which have led to peer-
reviewed journal articles (eg, Taylor et al, 2006; Hillman et al, 2006). The group is also
experienced in issues related to alcohol consumption and health impacts including modelling
the elasticity of demand for various alcoholic beverages from industry data in a taxation
context, a technical critique of injury risk from excessive alcohol consumption (Access
Economics, 2008a) and assessing indicators of alcohol consumption among young people.
Finally, Access Economics is a panel provider of services in health economics evaluation
and health and social policy analysis for the Australian Government (including the
Department of Health and Ageing, the Department of Families, Housing, Community
Services and Indigenous Affairs, the Department of Veterans’ Affairs and the Department of
Education, Employment and Workplace Relations), as well as numerous State Government
bodies, and is well regarded for the quality and accuracy of our health and social policy
modelling methods and advice.




3
    http://www.nationaldrugstrategy.gov.au/internet/drugstrategy/publishing.nsf/Content/mono64


                                                                                                     1
                                                           Collins and Lapsley report review




2. THE SOCIAL COSTS OF ALCOHOL
Chapter 2 of Collins and Lapsley (2008b) is based on Collins and Lapsley (2008a)
comprising an estimated $10.8 billion in ‘tangible’ (financial) costs and $4.5 billion in loss of
healthy life (totalling $15.3 billion).

Our previous critique of Collins and Lapsley (2008a) raised serious concerns in relation to
major conceptual and methodological problems in particular the following.
      A ‘demographic’ (retrospective) approach is used, which is likely to yield a much higher
      estimate than the more common human capital approach.
      Terminology is confused in relation to human capital methods and there is a failure to
      distinguish between incidence and prevalence based costings.
      Alcohol abuse is not clearly defined, with definitions implicit in different cost elements
      varying throughout the report and data on criminality impacts relating to consumption
      rather than abuse.
      Key parameters are not validated – eg, that 20% of alcohol consumption is addictive
      and 30% is abusive, with sensitivity analysis at a high 50% only.
      3    Lack of such validation – eg, through primary data collection – is considered a
           serious flaw, as is the lack of lower sensitivity analysis.
      Social costs are defined as negative externalities with private costs and benefits
      ignored, rather than a superior matrix approach (ie, cost type by who bears the cost).
      3     There is no mention in the report of the very large private benefits of responsible
            alcohol consumption in terms of utility value (choice, enjoyment).
      3     It is a major flaw to exclude private costs from total social costs and private
            benefits from total social benefits and it is incorrect to claim that costs borne by
            individuals are irrelevant to public decision-making.
      There are a number of data and methodological problems in relation to key
      parameters.
      3    For the valuation of healthy life, Access Economics (2008c) provides an
           alternative, most recent and superior approach.
      3    For deadweight losses (affecting the magnitude and distribution of the Collins
           and Lapsley estimates), Access Economics suggests a method and values
           based on Lattimore (1997) and Productivity Commission (2003) (Section 2.13).
      There is a failure to identify important corollaries of their analysis – eg, that taxation
      raised from alcohol abuse exceeds its cost to Government by a considerable margin, or
      that taxation increases from 1998-99 to 2004-05 greatly exceed population or
      consumption growth.
      The lack of epidemiological information is a weakness – the number or demographic
      characteristics of the estimated abusers are not estimated or provided, and attributable
      fractions are assumed unchanged historically, which the Appendix A author (Pollard)
      describes as a ‘serious limitation’.
      Appendix B author (Makai) cautions against use of the data and the substantial risk of
      over-estimation of crime costs ($1.4 billion), yet there is no sensitivity analysis, despite
      the upside bias.




                                                                                                2
                                                        Collins and Lapsley report review


     Estimates of productivity losses ($5.1 billion) have four sources of upward bias:
     (1) estimates of absenteeism are very high and rely on the oldest (1990) Australian
     Bureau of Statistics (ABS) National Health Survey (NHS) data when alternative 2004-
     05 data are available; (2) unpaid work is estimated based on the (high) work
     absenteeism estimates, although someone may simply defer household work till they
     are feeling better; (3) the ‘demographic’ approach is used; and (4) replacement cost
     valuation is used when opportunity cost valuation is lower and preferable.
     3     There was also inadequate detail regarding mortality, morbidity and national
           accounts data to replicate or validate the method.
     Health expenditure ($2.0 billion) is estimated from a number of different data sources,
     and suffers from omissions and incompleteness that is identified by Collins and Lapsley
     as an area for improvement in Sections 9.8, 9.9 and 9.11.
     3    These weaknesses would not exist if the more appropriate Australian Institute of
          Health and Welfare (AIHW, 2005) dataset had been adopted, which would have
          permitted detailed breakdown of actual alcohol-attributable costs by alcohol-
          related condition and by type of health system expenditure, using a more
          comprehensive and methodologically superior top-down attribution and enabling
          estimates of omitted items such as over-the-counter pharmaceuticals, pathology
          and imaging, allied health and other health expenditures.
     3    An example of how this can be done is provided in Access Economics (2008b) in
          relation to the cost of obesity.
     Road accident costs ($2.2 billion) are based on an out-dated Bureau of Transport
     Economics (2000) report when it would be preferable to use unit cost data for road
     accidents from Austroads (2007, 2006, 2003).
     Resources used in abusive consumption ($1.7 billion) are included and, while this is
     appropriate for smoking and illicit drugs, where any consumption is harmful, it is
     inappropriate for alcohol since the problem is the pattern of consumption, not average
     consumption and hence misallocation of resources in production.
     3    Hence, Collins and Lapsley over-estimate the total costs of alcohol consumption
          by 18.4% and this item should be completely excluded.

Our conclusions in relation to the report, as per the box below, also relate to the costs as
they are used in Collins and Lapsley (2008b).

     Overall the report is poorly structured with inadequate cross-referencing, there is
     a paucity of supporting evidence and references for claims, the referencing is
     incomplete and there is a very serious lack of transparency in calculation
     processes. Methodology and data use are singularly poor.

     There is apparent upward bias in most of the cost estimates and uncertainty is
     not dealt with due to grossly inadequate sensitivity analysis. Using a best
     possible approach would exclude one cost item, making the overall cost of
     alcohol abuse at least 18% lower. The other findings of the report should be
     viewed with substantial scepticism.

     We agree with the final suggestion (Section 9.14) that the data needs are
     reviewed and an expert advisory panel convened, also to review methods, given
     the serious weaknesses identified in this critique.




                                                                                           3
                                                          Collins and Lapsley report review




3. METHODOLOGICAL AND DATA ISSUES
This section critiques the appropriate use of concepts, methods, data and literature – ie,
whether they are the best, most recent and most appropriate available, and relates mainly to
Sections 1, 2 and 4 of Collins and Lapsley (2008b), and Appendix A.

3.1         BROAD CONCEPTUAL AND METHODOLOGICAL ISSUES

 3.1.1       PURPOSE OF THE ANALYSIS
An over-arching issue is the actual purpose of this report and what it claims to do. The title of
the report suggests that the purpose of the analysis is to identify the potential benefits of
effective policies to reduce alcohol abuse. However, the Executive Summary (p xii) states:

      The objective of the present study is not to recommend the adoption of a
      particular set of alcohol policies. It is to consider from an economic perspective a
      range of policies which have been shown to be effective and to indicate, as far as
      the data allow, the economic benefits likely to flow from the implementation of
      these policies. The information is intended to assist in the development of
      evidence-based strategies for the reduction in the social costs of alcohol abuse in
      Australia.

      The most common approach to achieve such an objective is the use of
      evaluation techniques such as cost benefit analysis (CBA) and cost effectiveness
      analysis (CEA). However, neither of these techniques are adopted in the report
      and, notably, the costs of the interventions are not estimated.

      To estimate gross benefits without estimating costs is likely to provide spurious
      conclusions – if costs outweigh benefits there may be no net benefit in any of the
      interventions reviewed. It is thus dubious how the information can assist in
      developing evidence-based strategies.

Less importantly, there is confusion in the report as to whether the purpose is in fact to
provide policy inputs. The citation above starts by saying that the purpose is not to
recommend particular policies. Yet the flavour of its conclusion is that it is intended to be a
policy building block. Page 1 then says that the ‘object of the present study is the provision of
economic information which will assist in the design of rational and effective public policies
towards alcohol’. Page 2 though, says it is ‘not the intention of the report’s authors to provide
specific policy recommendations’ and the ‘evidence presented here is essentially economic
in nature’. There is an entire Section (8.1) titled ‘Policy Implications’ and the flavour of the
Executive Summary is very much: ‘here are the policies that are effective and here are their
benefits.’ The appearance is thus that the purpose is to gain traction for a certain set of
policies, although in fact no economic evidence (CBA or CEA) is provided for any of
the identified policies.

Appendix A elaborates by suggesting that: Collins and Lapsley (2008a) represented an
estimate of the total social costs of alcohol abuse; Collins and Lapsley (2008b) represents an
estimate of what proportion overall are avoidable; and there is a need for three more follow-
up studies – (1) the time period over which the reduction can occur; (2) the nature of
programs that could achieve the reduction; and (3) calculating the social rate of return from



                                                                                               4
                                                         Collins and Lapsley report review


such interventions. It is not clear why one would undertake this five-step process when
it would be possible just to do realistic CBAs in the first instance.

A good starting point is the CBA Unit at the Australian Government’s Office of Best Practice
Regulation (OBPR), which provides guidance on how to undertake studies to provide
economic evidence for policy inputs – see http://www.finance.gov.au/obpr/cost-benefit-
analysis.html. Notably, this site references Access Economics (2008c) in CBA best practice
for valuing potential healthy life gains.

 3.1.2       CONCEPT OF A FEASIBLE MINIMUM
The concept of a ‘feasible minimum’ is unusual in economics. Optimal social outcomes are
achieved when marginal social costs are equated to marginal social benefits. As such, the
conceptual approach that Collins and Lapsley (2008b) adopt is highly irregular and overly
simplistic: essentially, ‘we know we can’t eliminate all the social costs of alcohol abuse but
let’s define a level that we think is achievable and call that the Feasible Minimum’ (with
capital letters). The concept is ascribed only to references where Collins and Lapsley are co-
authors (Single et al, 2003; Collins and Lapsley, 2006) in the context of work they did for
Health Canada ie, there is no other literature cited that uses this approach.

Moreover, if an incidence based costing had been adopted, marginal costs could then be
compared to marginal benefits. Moreover, it is conceptually feasible to eliminate all social
costs of alcohol abuse over time by eliminating all incident cases over time.

Of note is that it is quite conceivable that Australian society is currently operating at the
social optimum for harm minimisation from alcohol abuse – or has even intervened too much
– yet these possibilities are not considered in the report. Appendix A (Collins and Lapsley,
2008b:44) mentions the economic concept of optimal use and its superiority over the
avoidable costs approach:

      ‘The estimation of avoidable costs takes no account of the costs which would be
      incurred in achieving the required reduction in substance abuse and so of itself,
      the concept yields no information on rates of return. The concept of the optimal
      level on the other hand, takes explicit account of the costs involved in reducing
      substance abuse.’

However, after this important statement, there is no further discussion, instead switching
back to how to estimate a ‘feasible minimum’. Notably, the decision to adopt a ‘demographic’
(historical) approach, although a traditional prevalence or incidence-based costing would be
more appropriate, largely creates the problem of then having to trying to identify a feasible
minimum, which will necessarily change over time if policies are in fact effective.

The ‘four reasons for estimating avoidable costs’ (p3 and Appendix A) – government
expenditure prioritisation, targeting specific problems, identifying research gaps, determining
policy efficiency – would all be far better addressed using conventional CBA and CEA
techniques that are used across other Australian Government programs (eg, the
Pharmaceutical Benefits Scheme, program evaluation for cancer screening and other health
and social programs, and so on), as well as in the private sector.

3.2        ESTIMATING THE FEASIBLE MINIMUM
The four Health Canada approaches are summarised in the next section, with discussion of
the appropriateness of each approach.



                                                                                             5
                                                                 Collins and Lapsley report review


    3.2.1      THE ARCADIAN NORMAL APPROACH
The ‘Arcadian Normal’ approach takes the lowest age-standardised mortality rate (SMR) for
alcohol-attributable conditions in a group of 14 countries4 as reported in an unpublished study
(referred to in Collins and Lapsley, 2008b, as ‘Rehm et al, forthcoming’).

Since the benchmark is the current SMR for these various alcohol-attributable conditions in
another country, the implication is that such a country cannot do better on alcohol abuse
policies, even if the reason for its low SMR is because it has excellent breast cancer
screening programs or a high degree of enforcement of low speed limits for example (breast
cancer and road traffic accidents are both alcohol-attributable conditions). There appears to
be no econometric or regression analysis to explain why the SMRs might be lower or higher
in each country, and no control for such confounding factors. The only factors taken into
account are age, gender and GDP per capita (within 10% of Canada’s). A major weakness
of this approach is, thus, that structural (policy, regulatory) and cultural differences between
countries are not taken into account and the approach is fundamentally flawed as a
consequence. Collins and Lapsley (2008b:8-9) mention this serious flaw almost in passing:

       Countries with low overall road traffic accident death rates may have relatively
       high death rates from alcohol attributable accidents. In fact, a case could be
       made out that the Arcadian Normal should be determined by the size of the
       country-specific attributable fraction, rather than the country-specific attributable
       mortality, though it is in practice unlikely that such data would be available at the
       necessary level of country disaggregation.

Collins and Lapsley (2008b) also self-identify other major problems with the approach:
       it can only be applied to mortality/morbidity costs, not crime or productivity costs;
       it ignores the benefits of protective effects of responsible alcohol consumption;
       it has not proved possible in practice to use the approach so far; and
       there is no analysis of the policies that have enabled the Arcadian Normal countries to
       achieve that status.

These are such serious limitations that it is curious why the approach is presented at all, let
alone that its conclusion of ‘48% feasible minimum’ reduction is utilised.

    3.2.2      EXPOSURE BASED COMPARATORS
The exposure-based approach compares Australian average alcohol consumption with that
in selected other OECD countries5 for 1960, 1970, 1980, 1990 and each year 2000 to 2003.
Norway’s average consumption over this period was lowest, and comparing Australia’s
average relative to Norway’s gives a ‘potential reduction’ of 63.8% in 1960 falling to 38.8% by
2003. Collins and Lapsley (2008b) courageously conclude that: ‘it would appear that in broad
terms there is the potential to reduce Australian per capita consumption by about 40%’.




4
 Australia, Austria, Belgium, Denmark, Finland, France, Germany, Iceland, Italy, Japan, Netherlands, Sweden,
UK and US. It is unclear why the approach is called ‘Arcadian Normal’.
5
  Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Ireland, Italy, Netherlands, New
Zealand, Norway, Spain, Sweden, Switzerland, UK and US. Bold countries are additional to those in the
Arcadian approach, although Iceland and Japan are excluded. Reasons for the country selection are not clear.


                                                                                                          6
                                                             Collins and Lapsley report review


This approach, like the previous one, is seriously flawed. One policy implication might be to
encourage Norwegians to drink more, and then Australia would not have a policy problem.
And indeed they have been, doubling consumption over the period for a whole range of
unexplored reasons. One such reason might well be relative growth in incomes – Norway’s
increased 34-fold over the period while Australia’s increased only 15-fold.6 Major problems
with this approach are listed below.
       There is no comparison of income, cultural, climate, demographic/ethnic or pricing
       differences, among a plethora of other potential factors that might result in different
       alcohol consumption patterns across countries.
       Average alcohol consumption is not a proxy for rates of alcohol abuse, so the
       comparison says nothing about rates of abuse.
       If Norway (or other countries) perchance have policies designed to reduce alcohol
       abuse and these have been effective, this measure provides no information about such
       policies, let alone their economic rationale for application in Australia.

Again, these are such serious limitations that the inclusion of this as a ‘method’ in a technical
report is hard to understand.

    3.2.3     THE DISTRIBUTIONAL APPROACH
This approach is summarised in one paragraph (Collins and Lapsley, 2008b:45) as
conceptualising the avoidable burden by specifying alternative scenarios for risk-factor
distributions, including potential future distributions. There is no information provided to
ascertain why this approach was discarded by Collins and Lapsley, except a general
reference to the choice of method being ‘a pragmatic issue’ depending ‘on data availability’
(p47). Moreover, it seemingly contradicts the summary on page 38 that this approach is
called the ‘epidemiological model’ and was rejected due to data demands.

    3.2.4     EVIDENCE ON THE EFFECTIVENESS OF INTERVENTIONS
This approach appears to represent the conventional means where efficacy parameters are
derived from the literature, as key parameter inputs into CBA and CEA studies.

       It may at times be useful also to utilise research evidence on the effectiveness of
       interventions designed to reduce or alleviate the effects of substance abuse.
       (Collins and Lapsley, 2008b:46)

Such studies are directly useful in answering the key research question regarding the
economic evidence for useful policy interventions to cost-effectively reduce abuse. However,
Collins and Lapsley (2008b) use the parameters indirectly to estimate ‘avoidable costs’,
rather than moving directly to CBAs/CEAs that would straightforwardly address the issues of
most interest.

Moreover, well-constructed CBAs and CEAs estimate net present values (NPV) over future
timeframes (eg, 10 years, 25 years or longer), addressing the issues raised in relation to
‘time lags’ (Section A4 of Appendix A in Collins and Lapsley, 2008b). Selection of
timeframes and discount rates are addressed in Access Economics (2008c) as
recommended by OBPR for best practice. Conventional NPV analysis would avoid cautions


6
       Income and other comparisons can use online data and tools such                  as   that   at
http://www.nationmaster.com/red/graph/eco_gdp_percap-economy-gdp-per-capita&date=2003


                                                                                                    7
                                                                     Collins and Lapsley report review


made by Collins and Lapsley such as (p47): ‘Our estimates are expressed in 2004-05 dollars
but these benefits could not have fully accrued in 2004-05.’

The statement at the conclusion of the appendix (that CEA inevitably results in reduced
usefulness since only policies with the same output can be compared) seems unduly
negative. CEA using the incremental cost effectiveness ratio (ICER) is a common and well-
accepted evaluation technique in health economics worldwide and across the Australian
public and private sectors. The ICER metric measures dollars per disability adjusted life year
(DALY) averted or per quality adjusted life year (QALY) gained. Benchmarks for cost
effectiveness have been developed in different sectors by different organisations – for
example the World Health Organization.7
         WHO defines a cost effective intervention as one that costs less than three times GDP
         per capita per QALY gained or DALY averted.
         WHO defines a very cost effective intervention as one that costs less than GDP per
         capita per QALY gained or DALY averted.

         The methods (‘feasible minimum’, ‘Arcadian normal’, ‘exposure-based
         comparators’) selected are not appropriate to the purpose. Conventional CBA
         and/or CEA would have been preferable, as recommended by the OBPR for
         example, for evaluating economic evidence in relation to proposed interventions.
         The methods and numbers selected to derive the conclusion that it would be
         possible to reduce the social costs of alcohol over time by about a half, have no
         valid evidence basis, in our view. Conceptually, by preventing all incident cases
         of alcohol abuse, over time one could eliminate all its social costs.


3.3             EFFECTIVENESS OF INTERVENTIONS

    3.3.1         METHOD FOR DERIVING THE INTERVENTIONS REVIEWED
The basis for the choice of effective policies is not convincing. There is no discussion of the
literature review process, search protocols used or the basis for inclusion or exclusion of
particular studies. Such methods are standard protocols for evidence-based research, yet
are not adopted in Collins and Lapsley (2008b). Certainly the literature is much broader than
the eight studies mentioned on p11.

Even then, only one (European) study is used to provide evidence (Anderson and Baumberg,
2006), with no reason given for its use. Moreover, this particular study does not provide
objective comparisons of interventions such as a league table of ICERs or benefit/cost (B/C)
ratios (as would be expected), but instead provides a subjective 1/2/3-star rating system
including qualitative summary terminology such as ‘relatively high’, ‘moderate’ or ‘low’ cost to
implement and sustain (with no mention of the benefits relative to such costs).

The main criterion for inclusion appears to be that the interventions have been tried at some
point historically and found to have some degree of gross (not net) benefit. Such an
approach is likely to embed historical mistakes in future policy.




7
    See for example, http://www.who.int/choice/costs/CER_levels/en/index.html


                                                                                                     8
                                                          Collins and Lapsley report review


 3.3.2       AMBIT CLAIMS
Claims are made regarding the status of interventions for which no evidence or reference is
provided. For example (Collins and Lapsley, 2008b:12): ‘it may well be that a particular
Australian intervention policy is already close to world best practice (for example, driver
blood alcohol testing)’. No evidence is provided that this intervention is ‘world best practice’
(no definition is given for the term), nor whether the benefits outweigh the costs of the
intervention (although it is also described, unreferenced, as ‘proven-effective’). Such testing
might be cost effective when road traffic accident rates from alcohol abuse are high, but
economic theory suggests there will come a point at which they are no longer cost effective.
A key issue is whether we have yet reached that point in Australia, and the Collins and
Lapsley framework, by departing from traditional frameworks, is not able to address this.

Similarly, the suggestion that a policy ‘such as RBT or alcohol excise taxation’ ‘may be very
cost-effective’ because they ‘may be able to be implemented at very low … additional cost’ is
an unsubstantiated statement that ignores the denominator – effectiveness. Furthermore, for
whom additional costs are ‘very low’ is not mentioned, and it is not clear that costs of either
intervention are low to individuals. Rather it appears that only costs to government may have
been considered in this claim.

 3.3.3       OTHER ISSUES RAISED IN CHAPTER 4
The point is well made in Section 4.3.2 of Collins and Lapsley (2008b) that it is difficult to
aggregate the economic benefits of individual policies, and the authors do not attempt to do
this overall. The risk is that potential benefits are overestimated if summed.
     However, in quality CBA/CEA studies, the effect of combination policies are evaluated,
     using efficacy parameters for each intervention separately and for both interventions
     together.    Access Economics has adopted such an approach in evaluating
     combinations of options for occupational health and safety codes in Regulatory Impact
     Statements.

The point is also well made in Section 4.3.3 of Collins and Lapsley (2008b) that some
interventions (eg, excise) run the risk of reducing the benefits as well as the costs of alcohol
consumption. Unfortunately, not enough is made of this very important point, which is in fact
the raison d’être of undertaking economic evaluations for policy inputs. Moreover, the point
is only made in respect to health benefits, ignoring consumption (utility) benefits.

     Methods used to select so-called effective interventions are problematic and out
     of line with standard practice. Unreferenced claims are made that certain
     interventions are ‘world best practice’, ‘proven-effective’, or ‘very cost-effective’,
     without evidence. Important issues are raised such as the treatment of
     combination policies and the risk that some policies will reduce benefits as well
     as costs of consumption – yet these are ignored, thus overstating the potential
     benefits of the selected interventions.




                                                                                              9
                                                                     Collins and Lapsley report review




4. INTERVENTIONS, CONCLUSIONS AND
   RECOMMENDATIONS
This section reviews the intervention choices, conclusions and recommendations (Sections
5-9 of Collins and Lapsley, 2008b) with special comment in relation to:
         the potential dollar reduction in Australian social costs (based on per capita
         consumption rates as being the achievable parameter in Norway, USA or Italy);
         the potential benefits of specific policy interventions in terms of reductions in social
         costs; and
         the overall policy implications concluded as a result of the Collins and Lapsley
         Avoidable Costs work.

4.1             INTERVENTION CHOICES
Four interventions are presented:
         alcohol taxation;
         bans on alcohol advertising;
         measures to reduce drink driving; and
         ‘brief’ interventions for reducing hazardous alcohol consumption.

    4.1.1         ALCOHOL TAXATION
Collins and Lapsley (2008b:14) note there has been no Australian research on the price
elasticity of demand for alcohol in over a decade. Anderson and Baumberg (the one study
used in drawing conclusions), however, give three stars to alcohol tax as a means to reduce
the quantity of alcohol demanded and alcohol harms. It is not clear what the exact evidence
is that abusive consumption is reduced or what the efficacy, net benefits or cost-
effectiveness is likely to be in Australia from a given percentage increase in GST, customs
duties, excise duties or the Wine Equalisation Tax.

Collins and Lapsley (2008b:15) exclude GST and customs duties as appropriate
mechanisms, due to lack of feasibility for differential GST and the narrow base for imported
alcohol. Although noting that WET is ad valorum, they suggest it is still appropriate to use to
target abusive consumption (the validity of this conclusion is not clear). A major flaw is the
unreferenced claim (called a ‘fact’, although controversial8) that:

         ‘alcohol excise taxes are capable of being designed explicitly to target the types
         of alcohol known to be the subject of abuse (for example, high strength beer and
         alcopops) and to discriminate in favour of types of which is known to be relatively
         low (eg, low strength beer).’

There is no evidence provided that high strength beer and alcopops are exclusively the
subject of abuse, that low strength beer is not, or the status of other drinks eg, bottled spirits
and wine, relative to proportions consumed in healthy ways.



8
    See for example, Parliament of Australia (2008), Tobin (2008).


                                                                                                     10
                                                               Collins and Lapsley report review


Moreover, taxation imposes a deadweight efficiency loss (DWL) on the economy. DWL is the
loss of consumer and producer surplus, as a result of the imposition of a distortion to the
equilibrium (society preferred) level of output and prices. Taxes alter the price and quantity
of goods sold compared to what they would be if the market were not distorted, and thus lead
to some diminution in the value of trade between buyers and sellers that would otherwise be
enjoyed (Figure 4-1).

                     FIGURE 4-1: DWL OF TAXATION FOR HEALTHY DRINKERS
               Price ($)

                                                                              Supply
                                               Deadweight Loss (cost
                                               of raising taxation
                                               revenue)


           Price plus Tax

                            Taxation Revenue

                   Price




                                                                            Demand
                                                                                Output
                            Actual Quantity            Potential Quantity
                            Supplied                   Supplied


The rate of DWL used by Access Economics in 2008 is 27.5 cents per $1 of tax revenue
raised plus 1.25 cents per $1 of tax revenue raised for Australian Taxation Office
administration, based on Productivity Commission (2003) in turn derived from Lattimore
(1997), ie, 28.75% overall. For every dollar of excise tax imposed on healthy drinkers, there
are 28.75 cents of economic loss imposed on the Australian economy that are ignored by
Collins and Lapsley (2008b).

These points are also raised in Clarke (2006), who additionally discusses potentially harmful
addiction-swapping as a result of higher alcohol taxes (eg, towards recreational drugs) and
concludes that a better approach than high taxes is to more intensively penalise the socially
costly actions that are associated with alcohol abuse.

Collins and Lapsley (2008b) go on to spuriously conclude that an effective means of
addressing alcohol abuse is to determine the target rate of overall alcohol consumption
(including that of healthy drinkers, where there are no harms occurring), ‘and then design a
tax system to achieve that target’. The inflationary impact of the tax is not estimated, nor are
any possible government revenue gains or efficiencies relative to overall losses. Indeed, no
quantitative economic evidence is presented at all, nor any modelling.

     The comparison with Norway as the lowest per capita consumer of alcohol, and
     the US and Italy as wine producers, on which to base a target for Australia is
     fairly random, and suffers the plethora of issues raised in relation to the ‘Arcadian
     Normal’ and ‘exposure based comparator’ approaches to measuring the
     ‘Feasible Minimum’.



                                                                                               11
                                                             Collins and Lapsley report review


The heroic assumption (with no reference or evidence) is then made that ‘it is assumed that
a given percentage reduction in per capita alcohol consumption would lead to a similar
percentage reduction in the social costs of alcohol abuse.’ This is a fundamental issue, and
cannot be assumed away.

      Excise taxation is a blunt instrument that causes welfare losses for healthy
      consumers while at the same time not adequately preventing harmful use.
      Collins and Lapsley (2008b) ignore the deadweight losses (DWLs) of higher
      taxation on healthy drinkers, and conclude that differential alcohol taxes provide
      scope for targeting tax increases on hazardous and harmful consumption.
      Collins and Lapsley provide no evidence for this claim. Theory and evidence
      suggest that increasing taxation is unlikely to be cost-effective when adverse
      impacts of such a tax are considered (eg, very large DWLs and smaller addiction
      swapping impacts). A seminal low point of the report is the unreferenced claim
      that reducing alcohol consumption will proportionately reduce alcohol abuse.
      There is thus no basis for the conclusion that unspecified levels of tax increase
      would generate social benefits (cost savings) of $2.2 to $5.9 billion.


 4.1.2        BANS ON ALCOHOL ADVERTISING
A tiny amount of highly selective literature, devoid of sophisticated meta-analysis techniques
or even presentation of the range of efficacy parameters, is used to justify reductions in
consumption (not abuse) of alcohol as a result of partial and complete advertising bans.
Simple averages are used to estimate ‘best estimates’ (with no distributional or sensitivity
analysis), such as could have been modelled using a software program such as @Risk.

Although the method for deriving Table 12 from Tables 10 and 11 is not stated, it is in fact
derivable as ‘partial’ plus ‘full’ minus ‘partial’*‘full’, implicitly assuming independence of the
policies despite the problems highlighted in Section 4.3.2 of Collins and Lapsley (2008b).

The tables in Sections 5.2.2 and 5.2.3 are then presented with no discussion of the
weaknesses nor mention that Tables 16-18 ‘savings’ are subsumed in Tables 13-15. The
relativities in the ‘savings’ from road accidents to totals contradict the relativities in the Social
Costs report, since different efficacy parameters were used. The reason for this is not
intuitive or discussed. Inadequate evidence is also presented in relation to the view that
pressure for relaxation of any bans should be resisted.

      Bans on alcohol advertising are recommended as effective despite conflicting
      evidence, on the basis of a few selected studies (bereft of meta-analysis
      techniques). Highly simplistic (‘back of envelope’) methods are used to estimate
      efficacy from these limited findings, lacking distributional or sensitivity analysis.
      No confidence can be placed in the ‘savings’ so estimated. There is no
      discussion of what a partial ban might consist of in Australia.


 4.1.3        MEASURES TO REDUCE DRINK DRIVING
Two measures are put forward – more intensive enforcement of current blood alcohol
concentration (BAC) levels and lowering the legal level. The analysis is so simplistic that
differential levels for learner and probationary drivers, or differences across jurisdictions, are
not even mentioned.




                                                                                                   12
                                                           Collins and Lapsley report review


A single study over a decade old (Henstridge et al, 1997) is used to justify the efficacy
conclusions, with the assumption that the test rate per licensed driver is the same across
Australia now as it was in 1995 for the four states reviewed in that one study. The study
recommended increasing tests by 90% to one per licensed driver per annum, which Collins
and Lapsley claim would reduce serious accidents by 30%.                         They ‘adopt this
recommendation as being an externally validated and feasible target’. There is no evidence
presented for its feasibility, external validation, net benefits or cost effectiveness. Sensitivity
analysis is provided at 25% and 35%, with no reason provided for this confidence bound and
concern regarding whether it is wide enough given the uncertainty.

While acknowledging there is little international evidence for reducing BAC levels further,
Collins and Lapsley (2008b:24) nonetheless adopt some results from Sweden that reducing
from 0.05 to 0.02 (ie a 60% reduction) would lead to a 9% (8-10%) reduction in alcohol-
attributable road accidents. The benefits of the two policies are then considered additive (in
contrast to Section 4.3.2).

      Two measures to reduce drink driving - more intensive enforcement of current
      blood alcohol concentration (BAC) levels and lowering the legal BAC level – are
      presented as efficacious based on one experience from Sweden, ignoring the
      body of evidence. There is a problematic association with accident reduction for
      the first measure. No confidence can be placed on the finding that both
      measures (whose benefits are added together simply) is in fact $1.2 billion.


 4.1.4       ‘BRIEF’ INTERVENTIONS FOR REDUCING HAZARDOUS CONSUMPTION
The evidence provided in Wutzke et al (2001) and Gomel et al (1998) is also selective and
for the year 1996, but it is Australian and it does attempt CEA of GP-delivered interventions
among heavy drinkers, with a control group in three-arm analysis, which is an improvement
on the previous ‘evidence’. The evaluation metric is dollars per life year saved (an orthodox
CEA metric). However, the finding that there were 674 lives saved in the control group
suggested there might be sample size issues and, indeed, the ‘Australian’ (Sydney) sample
was 273 men and 124 women, suggesting only around n=132 in each arm at baseline.
Collins and Lapsley ignore the control and take the maximal support estimate of 1,972 lives
saved (the correct technique is to take 1,972-674) and compare it with alcohol attributable
deaths (in a later year) of 3,494 to estimate 56% reduction in alcohol-attributable deaths from
the brief intervention (this should be 37% if the control is deducted and different again if the
same years are used). As such the ‘best estimate’ of 38% is too high and the mortality
impact is spuriously applied to all cost impacts.

Chapter 5 of Collins and Lapsley (2008b:29) positively concludes: ‘It is very clear that, even
on the most pessimistic assumption, the cost effectiveness of brief interventions is very high’.
This is a valiant statement, given that they estimate an avoidable cost, not cost effectiveness,
and it is based on one small sample study.

      Brief interventions have better evidence for efficacy, but Collins and Lapsley
      make errors in estimating the percentage reduction in deaths and errors in
      applying the reduced mortality rate to total costs. As such, the estimated benefits
      of $5.8 billion are overstated.




                                                                                                 13
                                                            Collins and Lapsley report review


4.2         OTHER INTERVENTIONS (BENEFITS NOT EVALUATED)
Five interventions are presented:
      control of drinking environments;
      alcohol ignition locks;
      guidelines for low-risk drinking;
      standard drinks labelling and health warnings on drinks containers; and
      other interventions.

These interventions apparently are also deemed ‘likely to be cost-effective’ by Collins and
Lapsley (2008b:30), although it is not clear on what basis, particularly given they say the
benefits are not quantifiable. Quantification is a logical necessity to demonstrate cost
effectiveness.

 4.2.1       CONTROL OF DRINKING ENVIRONMENTS
Only evidence in favour of greater controls is presented.

 4.2.2       ALCOHOL IGNITION LOCKS
Again the evidence presented is uni-directional. It is also not clear how random retests can
definitively prevent someone else blowing into the alcohol ignition lock (a second time).
Collins and Lapsley (2008b:31) report that:

      Anderson and Baumberg (2006) categorise alcohol locks as being ‘relatively high
      cost to implement and maintain’, although their European estimates indicate
      benefits much higher than costs.

If this is indeed the case, why are the B/C ratios not reported?

 4.2.3       GUIDELINES FOR LOW-RISK DRINKING
It is unclear on what basis Collins and Lapsley (2008b) conclude that the NHMRC draft
guidelines are ‘useful’. Information is only useful to consumers if it is accurate, and it is
unlikely that much of the information provided in the draft is (Access Economics, 2008a).
Moreover, the NHMRC draft Guidelines are explicitly not targeted at the general population,
noting for example on p17 of the draft Guidelines document that “…it is a technical
document, it is not primarily aimed at the general public.”

 4.2.4       STANDARD DRINKS LABELLING AND HEALTH WARNINGS
No evidence is referenced in relation to any claim made in this paragraph.

 4.2.5       OTHER INTERVENTIONS
We agree further evidence is required in order to evaluate the relative cost effectiveness of
programs such as intensive alcohol detoxification, school-based education and workplace
interventions. Prospective CBA and CEA studies would be very valuable in relation to these
and all the other interventions considered in the report.



                                                                                            14
                                                           Collins and Lapsley report review


4.3        SUMMARY, CONCLUSIONS AND RECOMMENDATIONS
Chapter 7 of Collins and Lapsley (2008b) presents no new information. It is not clear why the
excise results are presented in a different table from the other interventions. The chapter
notes some (unspecified) limitations of the method and why the results should be interpreted
with ‘great caution’ (p34). Normally, precise limitations are enumerated with confidence
intervals around estimates.

Various additional puzzling claims are made such as: ‘All the interventions considered above
have been shown in the literature to be cost effective’.

The cost savings presented are over an unknown period, according to Collins and Lapsley.
This is not very helpful and falls well short of conventional CBAs where annual costs and the
NPV of net benefits over a particular defined period are presented as a minimum standard.

Table 4-1 shows the Collins and Lapsley results, with their conclusion being that the
maximum cost savings in the base case is higher than their estimate of the social costs and
over 1.5 times higher in the high case. Even allowing for the ‘summing’ issue (see Section
3.3.3), ‘savings’ of this order of magnitude might be considered potentially misleading.

       TABLE 4-1: SUMMARY OF ESTIMATED BENEFITS (REDUCTION IN SOCIAL COSTS), $M

                                                   Base           Low        High
                 Tax increase (?%)               2,810          2,810      5,940
                 Brief interventions*            5,830          3,490      8,160
                 90% increase in RBTs              940            780      1,090
                 BAC reduced to 0.02               280            250        310
                 Partial ad ban*                 2,450          1,680      3,210
                 Complete ad ban                 3,860          2,500      5,150
                 Total if additive              16,170        11,510      23,860
                 * Exact specification in an Australian setting unknown

It is unclear why the estimate of $15.3 billion (p36 and derived in Collins and Lapsley, 2008a)
is referred to as a ‘minimum’ social cost.

The discussion of marginal costs on p36 (second paragraph) seems to imply that the
‘Feasible Minimum’ is about socially optimal allocation. It is not. It was supposed to estimate
the proportion of historical costs that could feasibly be ‘avoided’. This is clarified in the
following paragraph, but one is left wondering why good economics was abandoned in favour
of a sub-optimal and flawed approach. We disagree with the claim that there are ‘extremely
useful policy conclusions’ from the paper, and consider it potentially misleading.

The research recommendations in Chapter 9 are not considered to be priorities. Rather,
detailed, robust CBA and CEA studies would be more useful, in line with normal practice.

      In our view, there are fundamental flaws in the methods used in this Collins and
      Lapsley (2008b) report and hence in its recommendations. There is also no
      disclaimer on the report that the findings do not represent the views of the
      Australian Government. We recommend the matter is reviewed by the
      Department of Health and Ageing and that genuine evidence only informs policy
      consideration of this important issue.



                                                                                             15
                                                             Collins and Lapsley report review




REFERENCES
Access Economics (2008a) Injury risk and drinking: a technical critique, Report for the
     Distilled Spirits Industry Council of Australia, March, Canberra, available at
     http://www.accesseconomics.com.au/publicationsreports/showreport.php?id=160&searchfor=20
     08&searchby=year

Access Economics (2008b) The Growing Cost of Obesity in 2008: three years on, Report for
     Diabetes       Australia,     August,        Canberra,          available        at
     http://www.accesseconomics.com.au/publicationsreports/showreport.php?id=172&searchfor=2008&search
     by=year

Access Economics (2008c) The Health of Nations: The Value of a Statistical Life. Report for
     the Office of the Australian Safety and Compensation Council (OASCC), January,
     Canberra,                               available                                   at
     http://www.accesseconomics.com.au/publicationsreports/showreport.php?id=156&searchfor=2008&search
     by=year

Anderson P, Baumberg B (2006) Alcohol in Europe: a public health perspective. A report for
     the European Commission. Institute of Alcohol Studies, UK, June.

Australian Institute of Health and Welfare (2005) Health system expenditure on disease and
     injury in Australia, 2000-01, AIHW, 2nd edition, Cat No HWE 28, April, Canberra.

Austroads (2007) ‘Guide to Program Evaluation Part 4: Project Evaluation Data’ (Second
     Edition), Austroads, Project No TP1050 Publication No AGPE04/07, May.

Austroads (2006) ‘Austroads Technical Report: Update of RUC Unit Values to June 2005’
     Austroads, Project No TP1166 Publication No AP-T70/06, September.

Austroads (2003) ‘Economic evaluation of road investment proposals: improved methods for
     estimating Australian unit crash costs’ Austroads, Project No BS.EN.511 Publication
     No AP-R238/03.

Bureau of Transport Economics (2000) Road Crash Costs in Australia, BTE Report 102,
     Canberra.

Clarke H (2006) Thinking About Alcohol Policy, Department of Economics and Finance, La
     Trobe              University,           October,             available            at
     http://www.latrobe.edu.au/business/assets/downloads/profiles/clarke/Alcohol%20policy
     %20paper.pdf

Collins DJ, Lapsley HM (2008a) The costs of tobacco, alcohol and illicit drug abuse to
      Australian Society in 2004/05, Report for the Department of Health and Ageing.

Collins DJ, Lapsley HM (2008b) The avoidable costs of alcohol abuse in Australia and the
      potential benefits of effective policies to reduce the social costs of alcohol, Report for
      the Department of Health and Ageing.

Collins DJ, Lapsley HM (2006) Counting the costs of tobacco and the benefits of reducing
      smoking prevalence in Victoria, Report prepared for the Victorian Department of
      Human Services.



                                                                                                     16
                                                      Collins and Lapsley report review


Gomel MK, Wutzke SE, Hardcastle DM, Lapsley H, Reznik RB (1998) ‘Cost effectiveness of
    strategies to market and train primary health care physicians in brief intervention
    techniques for hazardous alcohol use’ Social Science and Medicine 47:203-211.

Henstridge J, Homel R, Mackay P (1997) The long term effects of random breath testing in
     four Australian states: a time series analysis, Federal Office of Road Safety, April.

Hillman DR, Scott Murphy A, Antic R, Pezzullo ML (2006) ‘The Economic Cost of Sleep
      Disorders’ Sleep 29(3):299-305.

Lattimore R (1997) Research and Development Fiscal Incentives in Australia: Impacts and
      Policy Lessons, OECD Conference on Policy Evaluation in Innovation, 26-27 June,
      Paris, 81:574-7.

Parliament of Australia (2008) Dissenting report of Liberal Senators available at
      http://www.aph.gov.au/SENATE/COMMITTEE/clac_ctte/alcohol_beverages/report/d01.
      htm

Productivity Commission (2003) ‘Evaluation of the Pharmaceutical Industry Investment
     Program’ Research Report, AusInfo, Canberra.

Single E, Collins D, Easton B, Harwood H, Lapsley H, Kopp P, Wilson E (2003) International
      Guidelines for Estimating the Costs of Substance Abuse, World Health Organization.

Taylor HR, Pezzullo ML, Keeffe JE (2006) ‘The economic impact and cost of visual
     impairment in Australia’ British Journal of Ophthalmology 90:272-275.

Tobin H (2008) ‘How long until our pubs have no beer?’ IPA Review, July, available at
     http://www.ipa.org.au/library/publication/1219714336_document_60-3_tobin.pdf

Wutzke SE, Shiell A, Gomel MK, Conigrave KM (2001) ‘Cost effectiveness of brief
     interventions for reducing alcohol consumption’ Social Science and Medicine 52:863-
     870.




                                                                                         17

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:102
posted:1/31/2012
language:French
pages:103