Overview and analysis of possible transitional Taxi Industry Inquiry by alicejenny


									Overview and analysis of possible transitional strategies:
Moving from a tightly restricted supply model to an open
entry taxi industry

Paper prepared for the Victorian Taxi Industry

                               October 2011

                                              Jaguar Consulting Pty Ltd
                                                    ABN: 56089 615636
Final draft: 18 October 2011

Where a move to an open-entry taxi market from a starting point of a high level of supply restriction
is contemplated, transition strategies are frequently advanced as constituting more feasible reform
paths than an immediate opening of entry. Feasibility in this context has several elements:
avoidance or minimisation of the disruptions and costs associated with very large market
adjustments, avoidance of economic hardship for existing licence-holders, due to the value of the
licence being suddenly eliminated, and avoidance of the political risks associated with these
adjustment issues.

Several types of transition strategy can be identified. All can be assessed in terms of two key
characteristics: the time that would elapse before an open entry market would be attained and the
amount, if any, paid to incumbent licence holders during the transitional period. These two
characteristics determine the allocation of the costs and benefits of reform.

Delay in achieving an open entry outcome prolongs the existence of deadweight economic losses
due to restricted supply. However, they also continue the transfer from consumers to licence-
holders that results from supply restrictions. Hence, longer transition periods involve both higher
economic costs and greater transfers from consumers to licence owners. The payment of
compensation, or adjustment assistance, to licence-holders similarly involves a transfer of the costs
of reform from those who have profited from monopoly rents accruing from supply restrictions to
consumers and/or taxpayers. The distribution of the costs of reform is determined by both the
amount of such payments and the sources of the funds used.

Consistent with this frame of reference, this paper groups transitional strategies into two types:
those that involve immediate moves to open entry, but differ according to the amount of
compensation or adjustment assistance paid and those that involve staged releases of licences over
a period of time. However, it is important to recognise that these two groups of strategies are not,
in practice, distinct. A true open-entry market equilibrium only exists when licences are freely
available at a cost no greater than that required to recover administrative and regulatory costs.
Thus, models that notionally provide for licences to be freely available but incorporate significant
annual licence fees as a means of recouping compensation payments cannot accurately be described
as involving immediate open entry. The distinction drawn between the two models centres on the
question of whether there is a formal limit on the number of available licences, but the practical
implications of two strategies that do and do not involve such limits may differ little in practice.

This paper also briefly discusses strategies that aim to manage the relative scarcity of taxi licences
without moving to an open-entry market equilibrium.

The discussion is illustrated, where possible, by practical examples of the use of the reform
strategies under discussion. However, most of the reform strategies that have been identified by
policy-makers or in the academic literature have never been adopted in practice. Hence, the
discussion of these options is necessarily conducted at a theoretical level. An appendix considers the

Final draft: 18 October 2011
experiences of five OECD countries that have moved from a restricted entry model to an open entry
model of taxi industry regulation in the past 20 years approximately.

Final draft: 18 October 2011

1.      Introduction ................................................................................................................................... 5
2.      Immediate opening of entry .......................................................................................................... 6
     2.1.     Open entry with full compensation for incumbents............................................................... 6
     2.2.     Immediate open entry with partial compensation ............................................................... 13
     2.3.     Immediate open entry with ex gratia payments only........................................................... 20
     2.4.     Immediate open entry with no payments to incumbents .................................................... 23
3.      Staged releases of licences........................................................................................................... 24
     3.1.     Issue of additional licences to existing licence-holders ........................................................ 24
     3.2.     Open market auction of additional licences ......................................................................... 29
4.      Formula-based issue of licences................................................................................................... 32
     4.1.     Description and rationale...................................................................................................... 32
     4.2.     Population ratio based formulae .......................................................................................... 34
     4.3.     Load factor based formula .................................................................................................... 35
     4.4.     Multi-criteria based formulae ............................................................................................... 36
     4.5.     Subjective criteria ................................................................................................................. 38
     4.6.     Discussion.............................................................................................................................. 38
5.      Other reform options ................................................................................................................... 39
6.      Easing entry to the hire car industry ............................................................................................ 42
7.      Conclusion .................................................................................................................................... 46
Appendix 1: Jurisprudence in relation to compensation ..................................................................... 47
Appendix 2: Summary of reform strategies pursued in major jurisdictions ......................................... 49
Appendix 3: Bibliography ..................................................................................................................... 55
Other readings ...................................................................................................................................... 56

Final draft: 18 October 2011
1.       Introduction
The paper identifies and assesses the merits of a wide range of options for managing transitions
from a taxi industry characterised by strong regulatory controls on entry to the taxi industry - such as
that currently found in Melbourne - to a regulatory environment free of quantitative restrictions on
licence numbers and characterised only by minimum quality standards. Such transitional
mechanisms are necessarily also be applicable in a context in which it was proposed to move to a
situation in which quantitative restrictions were retained, but substantially eased.

Each transitional strategy is assessed in terms of the following criteria:

        Cost to government;
        Impact on consumers and licence-owners;
        Time taken for full implementation (estimated);
        Feasibility assessment (based on practical experience and theoretical assessments, as
        Assessment of key risks.

Four groups of transitional strategy have been identified. The first involves the immediate removal
of overt supply restrictions, combined with various mechanisms to address the impact on incumbent
taxi licence-owners and, potentially, lessors. The second involves staged increases in the number of
taxi licences. This group of strategies is consistent with both an eventual removal of all supply
restrictions and an endpoint based on the retention of (less restrictive) supply constrols.

The third group of transitional strategies is a variant of the second, in that it involves moving to a
formula-based approach to licence issue and/or total licence numbers. Finally, a fourth group of
transitional strategies consists of indirect approaches to taxi reform, notably involving the
encouragement of competition in related markets.

Some strategies discussed have been implemented in the course of reforming taxi regulation in one
or more jurisdictions, while others have been adopted in transitions to deregulated markets in other
industries. A third group has been proposed in the literature dealing with regulatory reform in the
taxi industry but has not, for one reason or another, actually been implemented. Where practical
experience with a transitional strategy is available, this will be assessed as far as possible. Where a
strategy has not been implemented in practice, a theoretical discussion of its merits and likely
feasibility and outcomes is undertaken.

Final draft: 18 October 2011
2.       Immediate opening of entry
Immediate removal of restrictions on the number of taxi licences potentially provides the welfare
maximising approach to taxi reform, since the efficiency benefits of an open entry market are
obtained without delay. However, different variants of this model have substantially different
implications, both in terms of the expected rate of increase in taxi supply and in terms of the
incidence of the costs of reform. Thus, several variants of this model are discussed below.

2.1. Open entry with full compensation for incumbents

Description and rationale

Providing full compensation to incumbents is generally taken to mean that the government
undertakes to purchase any tradable licences at the ruling market price immediately prior to the
announcement of the regulatory reform. Given the lack of transparency found in most taxi markets,
there is generally some doubt as to the exact "market price" to be applied. Perhaps in consequence,
this option is sometimes formulated in terms of government undertaking to pay an amount equal to
the highest price for which a licence has previously traded.

While the translation of a "full compensation" policy into practice is simple in relation to perpetual,
freely tradable licences, significant difficulties necessarily arise in the increasingly common
circumstances in which various kinds of restricted licences have been issued by government. In
particular, where these licences are not tradable, there is no readily established "market price". A
plausible approach, in cases in which governments have sold these licences for significant sums is
that the licence-holder would be refunded the sum paid1. However, where governments have sold
licences for sums that, while substantial, fall short of the capitalised value of the monopoly rents
available from the exploitation of the licences, licence-holders will clearly be made worse-off by a
"refund-based" buyback2. To this extent, such an approach cannot be said to offer "full
compensation". Sales of licences undertaken in Melbourne since 2002 would all fall within this

Box 1: The rationale for compensation payments

Soon (1999) argues that the compensation issue is “both a moral and practical one”. Morally, it is possible to
argue that licence holders should be compensated for the loss of their licence values, since the government
policy change has disproportionately disadvantaged them as a group. However, a strong counter argument is

  Where annual sums are paid, any compensation would presumably be limited to the unexpired portion of the current
year's payment.
  For example, the 10 year unrestricted licences recently auctioned in Melbourne were sold for $180,000, to be paid in
instalments of $18,000 per annum, indexed to the CPI. By contrast, assignment fees are currently around $30,000 per
annum. Hence, the successful bidders will receive windfall gains in terms of potential monopoly rents from exploitation of
the licence of around $12,000 per annum. This group of licence owners would not, therefore, be fully compensated by a
refund-based buyback payment.

Final draft: 18 October 2011
that purchasers of an intangible asset that derives its value entirely from a particular government policy stance
must be presumed to be aware of the risk that changes in the government policy position may reduce, or even
eliminate, the value of the asset. Policy changes (or, perhaps more correctly, policy stasis) have substantially
increased the value of these assets in the past: in fact, licence holders have generally achieved high rates of
return which should, arguably, be seen as reflecting the inherently high level of risk associated with the
investment . Symmetry would seem to argue that, just as licence holders have been able to reap windfall
gains from past government policy choices without these being confiscated, so they must expect to bear
windfall losses due to other policy choices.

A particular Australian perspective on this issue can be derived from the experience of implementing
legislative reforms under the National Competition Policy. The approach to the losers from reform advocated
by the National Competition Council and generally adopted by State and Federal Governments was that there
would not be a presumed right to compensation for economic losses due to pro-competitive reform.
However, adjustment assistance was made available where it was concluded that there were net social
benefits in so doing. In virtually all cases in which it was paid, this adjustment assistance amounted to only a
relatively small proportion of the economic losses from reform. Had taxi reforms been widely implemented
under the auspices of the NCP process, there would have been a strong presumption in favour of the adoption
of this approach . A broadly similar approach was ultimately taken in Ireland, following the failure of legal
action seeking compensation for lost licence values (see below). The Irish Government ultimately provided for
payments to be made to incumbent licence-holders who could demonstrate that they were suffering financial
hardship as a result of the move to an open-entry taxi market.

A further argument against compensation is that payments based on the full, pre-reform market value of
licences would be so large in total that virtually the whole of the available benefits of reform would be
transferred to the former incumbent producers. Considered alternatively, the whole of the costs of reform
would be borne by the taxpayer. Some individuals who are taxpayers but who make little or no use of taxis
would almost certainly be worse off as a result of reforms in which full compensation is paid. Moreover, the
political difficulty of arguing for such payments suggests that paying compensation may reduce the feasibility
of reform, rather than enhancing it.

At a practical level, Soon argues that:

Reform is less likely to be disruptive if affected interests can be “bought off”. Though New Zealand succeeded in
deregulating its taxi industry without any compensation, it did so in an extraordinary period when many other
reforms took place.

This argument may, however, be unduly pessimistic. New Zealand is not the only jurisdiction with high pre-
reform licence values to have removed supply restrictions immediately without paying compensation: as
noted elsewhere, Ireland also followed this path in 2000 despite Dublin licence values being little lower than
Melbourne licence values at the time of reform.

Does paying compensation remove political opposition to reform?

While incumbents would clearly prefer to obtain compensation, rather than not, it is less clear that the
promise of compensation would substantially eliminate opposition to reform from this quarter. From the
viewpoint of licence-holders, the maximum amount of compensation likely to be payable by a reformist
government will not exceed the immediately realisable value of the licence. In a context in which licence
values have historically risen rapidly, licence-holders can be expected to prefer the prospect of continued gains
to a compulsory buy-out at existing prices.

 The major risk involved inevitably being that of deregulation.
 As noted elsewhere, the Northern Territory – as the one jurisdiction to reform the taxi industry under the NCP program –
did not adopt this approach. However, this may be explained by the fact that the NT reforms occurred in the early days of
the NCP legislative reform program, prior to this general approach to compensation/adjustment assistance being
developed and clearly enunciated.

Final draft: 18 October 2011
Cost to government

The cost to government of full compensation based schemes is extremely large in jurisdictions, such
as Victoria, where persistent regulatory failure has led to very large imbalances between supply of,
and demand for, licences and, hence, very high licence values. At an average licence value of
$500,000, buyback of even the 3,100 unrestricted licences on issue in Melbourne would cost around
$1.55 billion, while additional expenditures would be required to buy back many or most of the
remaining licences - almost 2,000 in number.

The prohibitive nature of these costs has frequently led to consideration of options for recovery or
part or all of the buyback expenditure. Most commonly, it is proposed that a substantial licence fee
be imposed on an annual basis, either for a fixed period or until a certain proportion of the buyout
cost has been retrieved. This option was adopted in practice in the Northern Territory in the context
of its National Competition Policy-based reforms in 1999. In Darwin, where $95,000 had been paid per
licence in compensation, the annual fee imposed was $16,000. The government initially stated that this fee
was to be removed following recoupment of the compensation costs, estimated to occur in seven to eight

A plausible alternative approach to recouping buyout costs might be to impose a consumer levy, in
the form of a fare surcharge. Thus, for example, a 10% fare surcharge could be collected by the ATO
in parallel with the GST/BAS system. Two potential benefits of such an option are that it provides
greater transparency for consumers regarding the cost and financing of the payment of
compensation and that it may have a lesser impact in limiting entry to the industry in the early years
after supply restrictions are removed.

Impact on licence-owners

A full compensation largely insulates licence owners from the cost of reform, if considered in the
static sense. However, opposition to reform is likely to remain, even under such an option. A key
reason for this opposition lies in the opportunity costs involved: licence owners lose the opportunity
to continue with an investment that has historically generated high and stable annual returns5. In
addition, there may be concerns as to whether full compensation will actually be paid in a timely

The Northern Territory reforms constitute a practical example of a case where reform was strongly
opposed by licence-owners despite full compensation being paid at the outset. This opposition was
maintained after the opening of the entry and ultimately led to the reforms being reversed.

Impact on consumers

  ESC (2005) found that assignment values had been stable at 7-8% over the long term. In addition, the real value of taxi
licences rose by 76% in the 15 years to 2004, or an average of 4% per annum. Thus, total real returns to a passive
investment averaged 11-12% over the period, compared with a long-term average of around 9% for the stock market - a
market which demonstrates substantially higher volatility in terms of annual returns.

Final draft: 18 October 2011
Consumers theoretically obtain maximum benefit from any strategy based on immediate removal of
entry restrictions, since both the deadweight losses and the transfers to producers associated with
licence-owners exploiting monopoly rents available in a restricted-entry market are immediately

However, where governments seek to recover substantial costs incurred in making compensation
payments to incumbents via hypothecated charges, much lower consumer benefits will be obtained
in the medium term. Most obviously, where a substantial annual taxi licence fee is imposed in order
to recoup these compensation payments, the rate of entry can be expected to be lower than
otherwise, since the licence fee constitutes a substantial additional operating cost. The major
practical example of this dynamic is the case of the Northern Territory in 1999-2000, where entry
rates were substantially below those experienced in New Zealand following the opening of the
market due to the requirement for all licence-holders to pay a $10,000 annual fee. It was initially
anticipated that this fee would remain in place for around eight years, during which time entry rates
would necessarily remain substantially lower than would be the case in the absence of the fee.

Time taken for full implementation

Consistent with the above point, implementation of an open entry policy can only be judged to be
complete when any "clawback" mechanism - such as an annual licence fee or consumer surcharge -
has been removed. That is, it is only at this point that an equilibrium number of taxi licences will be
sought and obtained, as a result of the supply-inhibiting impact of the licence fee being eliminated.
Basic arithmetic indicates that attempts to recover fully the costs of a full compensation package will
necessarily involve the adoption of such charges over many years.

For example, in the Northern Territory case, it was initially anticipated that the fees were to be
levied for around eight years. In the event, the fee remained at the same level for eight years and
then, rather than being removed, was increased substantially by the Government: In 2007, the
Darwin fee rose from $10,000 to $16,000, with further increases to $18,400 in 2010 being
scheduled6,7. Thus, the fees continue to be charged some 13 years after the initial removal of supply

That said, the continuation of the high annual fee regime in the Northern Territory, in a context, in
which severe supply restrictions have subsequently been re-imposed, appears to constitute a
conscious (if not public) policy decision by government to appropriate the monopoly rents created
by its regulatory supply restrictions, rather than a view that there remains an actuarial shortfall in
recovering the cost of the 1999 buyback.

This, in turn, raises a further issue in relation to attempts to recover the cost of compensation
payments. As noted above, removing these payments is a necessary condition for the establishment
of a free-entry supply/demand equilibrium in the taxi market. However, establishing a process in
which the government receives a substantial annual stream of income from licence-holders


Final draft: 18 October 2011
necessarily creates a barrier to the further reform that removal of the licence fee implies. That is,
having achieved partial reform of the industry at the outset, there is a strong probability that a
government will fail to complete the reform process several years later, when this implies foregoing
a substantial annual stream of income. As an example, if the Victorian government were to levy a
$20,000 annual licence fee to defray the cost of paying compensation to Melbourne licence-holders,
the annual revenue from this fee could easily exceed $120 million.

Thus, governments that seek to recover the cost of compensation will not only delay the
achievement of the full benefits of reform, they will reduce the likelihood that these benefits will
ever be achieved.

Feasibility assessment

The use of compensation strategies is generally promoted on the basis that it renders a rapid or
immediate move from a highly restricted environment to an open entry one more feasible. As it is
generally accepted that there is no legal requirement to pay compensation8 the relevant notion of
feasibility here is clearly a political one. It is arguable that governments will be more inclined to risk
major reform, such as open entry to a highly restricted market, if they believe that payment of full
compensation will reduce or eliminate negative reactions from incumbents. However, as discussed
above, payment of compensation is likely, at best, to limit opposition, rather than removing it

Moreover, any "feasibility" benefit thus attained must be weighed against the question of the
"feasibility" of diverting large quantities of public funds to this purpose. In the current context,
payment of full compensation to existing Melbourne licence-holders would cost at least $1.5 billion9.
A budget allocation of this magnitude to be used as a transfer to a group that has both profited
greatly from government regulatory restrictions over a long period and provided a service widely
regarded as being of unacceptably low quality would clearly be politically controversial at best.

That the option of funding full compensation from general revenue is essentially infeasible is made
clear by reference to the history of compensation payments in the context of industry reform in

 See: Abelson (2010). Court actions in Ireland (three cases), the United States and the United Kingdom have all resulted in
decisions that the reduction or destruction of taxi licence values via regulatory change does not constitute a "taking" of
private property and therefore does not give rise to an obligation on governments to compensate licence-holders.
Economic analysis concurs: for example, Sidak & Spulber (1998) set out four conditions for the recovery of stranded costs
and explicitly exclude taxi licence-holders from the list of possible appellants for compensation, arguing that the
“irreversible investment cannot consist solely of a franchise right to receive supracompetitive returns.” See: Sidak, J. G., &
Spulber, D.F: Deregulatory Takings and the Regulatory Contract, Cambridge, 1998. Cited in Koehler, B (2004). Regulating
Supply in Taxi Markets. City University, London.
  $500,000 x 3,100 permanent licences = $1.55 billion. The question of whether any compensation would be payable to
the holders of the various classes of restricted licences that have been created would also need to be considered.
Compensation payments to at least some of this group would be a likely outcome and would necessarily further increase
the total costs involved.

Final draft: 18 October 2011
Australia. Where compensation to incumbents has been paid, this has usually been funded through
some form of industry levy, whether paid in the first instance by consumers or producers10.

Even if a model involving partial or total recovery of this amount from future licence fees were
promoted, significant controversy would remain likely, with the implications in terms of delayed
realisation of the full benefits of reform doubtless being prominent in the discussion.

A key perspective on the feasibility of a "full compensation" approach to opening entry to the taxi
market is found in the fact that the Northern Territory circa 1999 is the only currently known
example of this strategy having been adopted in practice anywhere in the world. The following box
provides details of the NT experience.

Box 2: Taxi reform gone wrong: the experience of the Northern Territory

The Northern Territory Government was the only Australian government to remove entry restrictions on its
taxi industry as part of the National Competition Policy reforms. It removed restrictions in January 1999 and
paid compensation to incumbent licence holders based on the highest licence sale price observed prior to the
reform. However, annual licence fees were implemented in an attempt to recoup the cost of the
compensation package. In Darwin, where $95,000 had been paid per licence in compensation, the annual fee
was $16,000. This fee was to be removed following recoupment of the compensation costs, estimated to
occur in seven to eight years.

The high annual fee restricted entry to levels substantially lower than those experienced in New Zealand and
Ireland. Nonetheless, the industry complained strongly of an overcrowded market and low rates of return to
both existing and new entrants. This, plus some conduct related complaints led, the government to respond
by re-imposing a “temporary” cap on licence numbers as soon as November 2001. The “temporary” cap was
twice extended and has now been made permanent, with a maximum of 11.1 taxis per 10,000 population
being allowed (NCC (2003), p 2-19) – a level similar to that ruling in other Australian capitals characterised by
major supply restrictions and very high licence prices and only about one third of that found in New Zealand’s
open taxi market. The annual licence fee continues to be paid, almost 13 years after the initial reforms and,
indeed, has been substantially increased in recent years. There is no indication that it is now intended to be
other than permanent.

Thus, the net impact of taxi reform in the Northern Territory appears to have been an initial transfer of
approximately A$100 million to incumbent licence holders, followed by the adoption of a system that will see
the monopoly rents from restricting entry to the taxi market accrue in perpetuity to the government, rather
than being appropriated by licence-holders. While the latter is a preferable outcome to the pre-reform
situation, the market position of consumers is now essentially unchanged from that obtaining before the
reform process began, with the exception that the use of an explicit formula to set taxi numbers reduces the
risk that relative supply will become even more constrained over time. The Northern Territory experience thus
serves as a clear example of the difficulty of embedding open entry reforms in the long-term.

 For example, the $1.29 billion dairy industry adjustment scheme was funded by a consumer levy. See:

Final draft: 18 October 2011
Finally, the position of assignees of taxi licences must be considered. Assignment contracts typically
have a three year duration and involve annual payments of at least $24,000 to the licence owner.
Immediately moving to open entry would eliminate the monopoly rents that effectively fund these
payments. This implies that there is a case for action to be taken to prevent assignees suffering
hardship as a result of the reform. One option would be to pre-announce the reform, to take effect,
say, three years hence. A second would be to legislate to declare all assignment contracts void as
from the date on which the open entry takes effect.

Assessment of key risks.

A key risk from any variant of the immediate open entry strategy derives from the size of the
adjustment that would be required. The Northern Territory provides an example - albeit one in
which the starting point is one of a lesser degree of supply/demand imbalance than that found in
Melbourne at present - of the risk of a reform commitment being rapidly reversed due to the
political fall-out from a market in a state of adjustment toward a new equilibrium in the medium-

Addressing this risk would require a clear understanding on the part of decision-makers of the
expected post-reform trajectory of the market and a sustained attempt to communicate this to
consumers and the public. The use of legal mechanisms to make early reversal of the reforms
difficult to undertake would also merit consideration.


Payment of full compensation to incumbents as a corollary of an immediate move to an open entry
market would involve very substantial expenditures and would only be partially effective in
removing opposition to reform. The high cost of such payments gives rise to significant questions
regarding the feasibility of such an approach. The merits of such an option, in a context in which
case law indicates there is no legal requirement to pay compensation, are open to significant

Only one jurisdiction (the Northern Territory in 1999) is known to have adopted this approach to
reform. It did so in a context of much smaller licence values than those ruling in Melbourne at
present. While the reform initially proceeded, it was later reversed. However, this policy reversal is
unlikely to have been related to the payment of compensation.

Final draft: 18 October 2011
2.2. Immediate open entry with partial compensation

Description and rationale

This transitional strategy seeks to tread a compromise path between a full compensation and a zero
compensation model. Clearly, numerous variants of a partial compensation model can be
developed. At the simplest level, this could involve a determination that a given percentage of the
pre-reform market value of the licence, or a given dollar amount per licence, would be paid to all
licence-owners. However, a more sophisticated variant of this model would seek to differentiate
between owners according to identified criteria and would yield different outcomes for different
groups of licence owners. Deighton-Smith (2000) identified several general criteria that can be used
to assess claims for compensation in a context in which governments choose to reform regulated
industries in ways that will financially disadvantage incumbents. These were:

       Whether the owner paid the government for the licence or permit to operate and, if so, how
        much was paid;
       Whether significant additional investments have been made in the asset (or related assets)
        over time;
       What time has elapsed since acquisition of the licence and, by implication, to what extent
        any investments have been amortised; and
       The inherent risk involved in investing in an intangible asset reliant on the continuation of
        govt policy.

The application of these criteria clearly leads to a conclusion that holders of most taxi licences -
including all permanent, transferable licences - have little claim to compensation payments. Those
who have recently paid substantial sums to government for restricted licences would, on the other
hand, have some claims.

Importantly, Deighton-Smith (2000) also points out that, in the context of the conduct of the
National Competition Policy's Legislative Review Program in the 1990s and 2000s, the approach
generally recommended by the National Competition Council and adopted by the Federal
Government is to reject a "compensation" perspective for incumbents in an industry being reformed
in favour of an "adjustment assistance" perspective.

Here too, the arguments for the provision of assistance in the taxi industry context are somewhat
limited, particularly since there are virtually no "stranded assets" involved, other than the licence,
which is rendered worthless if an immediate open entry policy is adopted. Deighton-Smith (2000)

        "...in some cases significant assistance has been provided, while in others it has been much
        more limited, or even non-existent.

Final draft: 18 October 2011
         One key consideration in terms of adjustment assistance is that of whether substantial
         hardship would be likely to result in the absence of such payments. As noted above, for many
         owners, reductions in taxi-plate values will constitute the reversal of large “paper gains”
         made on the basis of relatively modest initial investments. However, for others, the
         possibility of hardship resulting from regulatory decisions that remove taxi-plate value is
         certainly real. It is most apparent in the case of small investors who have recently allocated a
         large proportion of their portfolios to the purchase of taxi plates. This “hardship” perspective
         tends to suggest consideration of the possibility of adopting different approaches to the
         payment of adjustment assistance, based on recognition of individual circumstances.

         It can be expected that arguments would be made that any such discrimination would violate
         equity principles. However, an understanding of the payments as being “adjustment
         assistance” suggests that what is acknowledged is not a right of redress against a change in
         government regulatory policy but, rather, an obligation on the part of government to
         address particular economic circumstances that could arise as a result. Thus, payments
         should be regarded as “ex gratia” in nature, rather than as representing compensation in the
         specific sense.11

Costs to government

Modelling conducted by Deighton-Smith (2000) concludes that the adoption of such a differentiated
approach could substantially reduce the cost to government of compensation payments (vis-a-vis
the adoption of an across the board full compensation model), while preventing hardship for
licence-holders. A number of scenarios were modelled, with the following results:

        Purchasing licences for amounts equal to the real value of the initial purchase price would
         yield a total cost 40% lower than that of a full compensation based approach;
        Limiting this buyback to those who had purchased their licences within the previous ten
         years12 would yield a total cost 55% lower than that of a full compensation based approach;
        Applying a "sliding scale" in which 100% of licence purchase price was paid to those who had
         purchased a licence in the past year, reducing to 50% of licence purchase price for those
         who had held licences for 10 years or more would also yield a total cost 55% lower than that
         of a full compensation based approach;
        Applying a sliding scale moving from 100% of purchase price in year 1 to 50% in year 6 would
         yield a total cost slightly more than 60% lower than that of a full compensation based

While these costs are clearly much lower than the cost of paying full compensation in all cases, the
amounts involved remain substantial. Applying the above percentages to the currently estimated
cost of a full compensation based buyout suggests that costs of at least $620 million and as much as
$930 million could still be incurred under such proposals. This inevitably raises the question of
whether incurring such costs is likely to be significantly more politically feasible, in practical terms,

  Deighton-Smith (2000), pp 15-16.
  This model assumed that no compensation would be paid to those who had held licences for longer than ten years, on
the grounds that their initial investments were largely amortised.

Final draft: 18 October 2011
than the option of paying full compensation. Conversely, this approach implies a broader
distribution of the costs of reform, as noted below.

Impact on licence owners

The different scenarios highlighted above have significantly different impacts on licence owners. In
the first scenario, all licence owners receive the inflation adjusted value of their initial investment.
Thus, while they incur significant paper losses, the net result of their investment in the taxi industry
remains strongly positive, given that assignment prices have consistently averaged 7-8% of the ruling
licence value at any given time13. Thus, for example, an investor purchasing a licence in 1997, for
approximately half its current nominal value, would have received annual returns initially averaging
7-8% and rising to around 15% per annum (nominal) in recent years.

In the second scenario, long-term licence holders incur substantial paper losses, since they receive
no compensation. However, this scenario is predicated on the view that, for those who have held
licences for more than a decade, the initial investment in the licence has been substantially
amortised via the high rates of return enjoyed in the long period since purchase of the licence.

Scenarios 3 and 4 would imply that many or most licence holders would face losses of some
proportion of the value of their initial investment. However, the proportionate losses would be no
greater than 50%, while the higher proportionate losses would apply only to long-term licence
holders who had had the opportunity to substantially amortise their initial investments. Hence,
hardship outcomes would be extremely unlikely, albeit that "paper losses" - i.e. the value of
compensation paid compared with the pre-reform value of the licence - would be very large in many

Impact on consumers

The same essential dynamics identified above in relation to full compensation based strategies are
likely to apply to these partial compensation approaches. That is, given the need to fund
compensation payments that could potentially exceed $1 billion, it is highly probable that
governments would seek to recover these expenditures via an industry levy, leading to significant
increases in licence-holders' cost functions and lower entry rates than would otherwise occur.

As the amount of compensation paid would be substantially smaller, consumers would clearly bear a
significantly lower proportion of the costs of reform. Put alternatively, they would obtain more
significant benefits in the short to medium term (if a lower annual fee were charged than under a
full compensation option) or would obtain the full benefit of open entry conditions sooner (if the
higher fee were charged over a shorter period).

   This long-term historical relationship appears to have broken down more recently, with lower rates of return being
earned. However, there is some indication that the increasing use of covert payments may mean that the apparent decline
in returns over-estimates the actual position.

Final draft: 18 October 2011
Time taken for full implementation

Were the government to choose to levy a licence fee to recoup the partial compensation payments
made, plausible assumptions suggest that a four to ten year period would elapse before the fee
could be abolished. For example, were a $20,000 fee to be set - an amount that is somewhat below
current assignment rates and thus allows for some degree of entry - annual revenue of $120 - $150
million would be generated14. As noted above, compensation costs under the various scenarios
would likely total between $600 million and $1.2 billion. Thus, four to ten years would be required
to fully recoup the initial compensation period in nominal terms and slightly longer were the
government to seek to recover the present value of the initial payment.

Clearly, this period could be reduced if the government were to choose to fund a proportion of the
compensation cost via consolidated revenue, rather than through a hypothecated charge on the
industry or its consumers.

Feasibility assessment

The above suggests that it is possible to fund partial compensation payments from a licence fee
while still allowing significant consumer benefits to be obtained from year 1, as well as providing for
the elimination of the licence fee over the medium term. However, the implications of different
partial compensation scenarios for the length of the transition period - and hence the ultimate
feasibility of the reform - are significant. Thus, there is a need to focus on limiting compensation
expenditure as far as possible and to consider partial funding of the compensation paid from
consolidated revenue in order to complete the transition phase as rapidly as possible.

The following is presented as an indicative scenario for a reform path based on partial

Box 3: Partial compensation scenario analysis

The basis for the payment of compensation is that set out in the third scenario above - i.e. that a "sliding scale"
approach to compensation would be adopted, in which 100% of licence purchase price was paid to those who
had purchased a licence in the past year, reducing to 50% of licence purchase price for those who had held
licences for 10 years or more. As noted above, this would be expected to yield a total cost 55% lower than that
of a full compensation based approach . On the assumption that compensation would only be paid to the
owners of the 3,100 perpetual, transferable licences, this implies a total cost of approximately $700 million.

This capital cost is assumed to be recovered by government via an annual licence fee to be payable by all taxi
licence-holders. This fee is to be a fixed term measure, with legislation setting a specific date for its abolition.
The fee would be set at a level that would allow for significant entry to the industry in the short to medium

  Assumes approximately a 20-50% increase in the current number of licences due to new entry.
  This is based on the assumptions set out in Deighton-Smith (2000). In particular this relates to the rate of turnover of
taxi licences and the rate of increase in taxi licence values. Clearly, should these variables have changed, there will be
implications for the cost of the buyout package and the time horizon for the completion of the reform process.

Final draft: 18 October 2011
term. Assignment fees are currently of the order of $24,000 . Hence, it is necessary for the annual fee to be
set at a level significantly below this level to achieve this outcome.
For simplicity, it is assumed that each of the 4,450 taxi licences yields a monopoly rent of $24,000 per annum.
In practice, peak service taxis and, to a lesser extent, WAT will yield smaller rents than perpetual licences, since
their average revenues are lower. Conversely, it is likely that the presence of hidden payments means that the
figure of $24,000 understates true assignment fees to some extent.

On this basis, the total monopoly rent earned in the market annually is $24,000 x 4,450 = $106.8 million. This
declines to zero in as the number of taxis licensed reaches 12,000 (i.e. 3 per 1,000, given a population of 4
million). For the sake of simplicity, it is assumed that the value of the monopoly rents available would decline in
a linear fashion as taxi numbers increased .

If taxi numbers were to increase to 5,800, available rents would diminish to approximately $87.7 million.
Dividing the former figure into the latter, yields an average monopoly rent per licence of $15,121. If the
government were to make taxi licences available at an annual fee of $15,000, approximately 5,800 licences
would be demanded in a steady state. Total revenue to the government would be (5,800 x $15,000) = $87
million. If government levied fees at this level for 8 years, the capital value of the initial partial compensation
package (i.e. $700 million) would be recouped.

This rough calculation represents only one possible option: higher fees would be associated with lower rates of
entry and shorter payback periods, while lower fees would be associated with higher entry rates and longer
payback periods. For example, at a fee of $11,000 should be associated with 6,300 licensees, an annual
revenue of $70 million and a 10 year pay back period.

It should be emphasised that these calculations are general and imprecise. Two limitations should be noted,
but tend to be offsetting in terms of their impact. First, payback periods have been calculated on the nominal
value of the compensation package and do not take into account the reduced value of the stream of licence
payments over several years (as a result of discounting). Conversely, the analysis is statically based. In
practice, at a set licence fee, the number of licences demanded at a given licence fee will increase over time
with taxi demand. This is a potentially important factor working in favour of this model, given the relatively
high rates of taxi demand observed in recent years and the increase in demand which the initial increase in
supply (and hence reduction in waiting times) would bring forth.

On the above analysis, the adoption of a $15,000 fee would lead to an initial increase of at least 1,350, or 30%,
in the number of Melbourne taxi licences, while government could recoup the cost of the partial compensation
package within eight years.

As noted above, multiple variants of this package are feasible. For example, if government were willing to fund
part of the compensation package from consolidated revenue, the delay until open entry was achieved could be
reduced. For example, if government were prepared to fund a cost of $260 million from consolidated revenue,
the open entry end-point could be achieved within five years.


   In practice, the decline in rents per additional licence may diminish over a wide range, due to the increases in taxi
demand which would be brought forth by additional supply and consequent reductions in waiting time.

Final draft: 18 October 2011
The above indicates that a transitional strategy based on a partial compensation payment being
recouped through annual licence fees over a transitional period prior to open entry being adopted is,
at least in the arithmetic sense, feasible. Importantly, the above scenario is based on a total
compensation bill that is less than half of the cost of paying full compensation at current market
value, while the minimum feasible transition period is approximately eight years. This implies that
a full compensation based transition would not be feasible, in the sense that it would not be possible
to fully recover the cost of the compensation package over any time horizon18.

Identification of key risks

The key risk identified above in respect of the levying of licence fees is that the government will
choose to retain this rich source of revenue in perpetuity (estimated at $87 million per annum in the
above "base case"), rather than allowing for open entry (i.e. licences priced at administrative cost)
after having recouped their initial costs. The prospect of this outcome occurring in practice is
heightened by the fact that licence-holders are likely to continue to lobby against policy changes that
would lead to further entry to, and competition within, the industry.

Moreover, the risk of the reform transition not being completed as planned must inevitably increase
with the length of the transition period: reneging on "old" promises is likely to have lower political
cost, given the enhanced potential to argue that circumstances have changed and point to a less
restricted industry with, presumably, higher levels of consumer satisfaction than were previously

Shorter adjustment periods will be obtained if there is a combination of a rigorous approach to the
design of a partial compensation scheme, which would minimise its cost, and the inclusion of a
component of funding of the scheme from general revenue, rather than recouping through licence
payments. These measures are therefore likely to be critical to the prospects of the reform being
completed in the medium term and a true open-entry outcome being achieved.


A key benefit of this option is that it would prevent the adoption of open entry reforms leading to
hardship for incumbents. This is likely to be seen as reducing the political risks of reform
substantially, relative to an immediate opening of entry with no compensation. The fact that most
incumbents would receive a high proportion of their initial investments in government payments
would mean that the majority of the public would be likely to see this outcome as generally
equitable, where this is unlikely to be the case in respect of a no-compensation approach.

Moreover, the partial compensation option is more feasible than a full compensation option in the
sense of its ability to be internally funded - at least to a substantial degree - by the industry within

     Given the impact of the discount rate.

Final draft: 18 October 2011
the medium term and be consistent with a move to fully open entry (i.e. licences at administrative
cost) within the medium term.

That said, this option would be opposed strongly by incumbents and would, in this respect, yield
similar political challenges to a no-compensation based option.

Final draft: 18 October 2011
2.3. Immediate open entry with ex gratia payments only

Description and rationale

As noted above, it is generally accepted that governments face no legal obligation to pay
consultation. Consequently, any decision to make compensation payments, whether full or partial,
is a political one. Equally, governments may choose not to compensate members of an industry who
have, for the most part, reaped substantial profits over an extended period simply via ownership of
a paper asset whose value is supported by no more than government regulatory decisions.

This option was employed in Ireland at the time of the major reform of the taxi industry carried out
in the late 1990s. While the initial intention was to make no payments at all to incumbent licence
holders, it was subsequently decided to make ex gratia payments available to those who could
demonstrate that personal hardship had resulted from the destruction of the former value of their
taxi licences. An independent Taxi Hardship Panel was established to review individual applications
for ex gratia hardship based payments and make recommendations.

Costs to government

The costs to government of a scheme based on the provision of ex gratia payments to some or all
licence holders may vary widely according to both the characteristics of the specific taxi market
being opened to competition and the criteria adopted. However, in the Irish case, the costs of the
hardship based payment regime adopted were extremely modest.

Payment of full compensation would have been extremely expensive: immediately prior to
deregulation there were 2,722 Dublin taxi licences with a value of I£90,000 each. Thus, the cost of
buying these licences alone at market prices would have been I£245 million. In addition, there were
almost 1,200 licences spread across the rest of Ireland. If the average value of these licences was
even one quarter of those of the Dublin licences, additional compensation costs of I£27 million
would have been incurred, making a total cost of full compensation of over I£270 million19.

In the event, the Taxi Hardship Panel recommended payments totalling €12.6 million (£9.9 million).
In answer to a 2005 Parliamentary question, the Irish Minister for Transport stated that individual
payments of between €3,000 and €15,000 had been made - i.e. a maximum of 16.7% of the pre-
deregulation value of a Dublin taxi licence. A total of €16,349,000 in hardship payments had been
made to 1,412 persons by 200520. The latter figure represents around 36% of the total of around
3,900 licences on issue prior to reform. Although the actual number of pre-reform licence-holders is

   Barrett, S.D., (2003). Regulatory Capture, Property Rights and Taxi Deregulation: A Case Study. Economic Affairs,
December 2003, pp 35-40.
   http://www.transport.ie/viewitem.asp?id=6424&lang=ENG&loc=1858. Note that these figures appear to be final, as the
closing date for applications was 30 April 2004.

Final draft: 18 October 2011
unknown, published discussion of the industry suggests that there were relatively few owners of
multiple licences.

The total of €16,349,000 (I£12.9 million approximately) in hardship payments represents less than
5% of the total cost of paying full compensation, as estimated above.

Impact on licence-holders

As Barrett (2003) points out, numerous harms were claimed to have been suffered by licence-
holders as a result of the Irish reforms:

            "The panel21 reported that ‘amongst the many and varied medical conditions cited as being
            directly attributable to, or intensified by, liberalisation, were strokes, hypertension, high
            cholesterol and heart problems, stress, anxiety and panic attacks, depression, asthma, back
            problems and fatigue’ and that, ‘this by no means purports to be an exhaustive list of
            ailments.’ (p 38)."

As noted, the Irish context was one in which no compensation at all was initially proposed by
Government, with the Taxi Hardship Panel being an ex post response by governments about
concerns raised over hardship resulting for some licence-holders. The above perhaps argues in
favour of a strategy of making an ex ante announcement of such arrangements. However, it is also
notable that the Victorian Essential Services Commission received warnings of similar outcomes and
worse (including suicide) in response to a 2005 draft report recommending that the existing hire car
licence fee be halved from $66,000 to $33,000.

In financial terms, as noted above, the hardship payments amounted in total to only around 5% of
the pre-reform value of licences. Thus, around 95% of the cost of the reform was borne by licence-
holders in this case. Individual licence-holders would, in many cases, not only have borne substantial
paper losses, but have experienced significant losses when any payment made was measured
against the historical cost of their investment.

Impact on consumers

From the consumer perspective, the implications of this option are identical to those of the
immediate entry with no compensation option. That is, the deadweight loss to economic welfare
associated with the presence of supply restrictions is removed immediately.

Time taken to full implementation

In economic terms, this option is consistent with the immediate achievement of the full benefits of
moving to open entry. The completion of the process of determining and distributing hardship
payments could, on the Irish experience, take a number of years.

     i.e. the Taxi Hardship Panel

Final draft: 18 October 2011
Feasibility assessment

This option arguably ranks significantly more highly than the closely related alternative of
immediately moving to an open entry outcome without any payments to incumbents, given that
discussion of the impediments to reform frequently focuses on the position of those that have paid
the market price for a licence in the recent past. Moreover, the Irish experience suggests that the
costs of a carefully designed hardship based approach could be very small in relation to the current
licence value, so that it does not create a practical impediment to reform based on the need to
divert substantial funds from the budget.

Identification of key risks

From the political perspective, this option entails risks in relation to the ability of incumbents to
argue that any hardship based scheme will, in practice, be inadequate to prevent substantial loss
and hardship occurring in practice.


This option arguably conceptually belongs at the end of the continuum of "partial compensation"
based approaches considered above. That is, while a conceptual distinction is made in this context
between "compensation" and "adjustment assistance" made as an ex gratia payment, the common
link is that some payment is made to former incumbent licence holders.

Final draft: 18 October 2011
2.4. Immediate open entry with no payments to incumbents

This option is included for the sake of completion, although it arguably does not constitute a
"transitional strategy". It's implications are essentially identical to the above option of moving to
immediate open entry with only hardship-based payments being made. The key differences are:

       Relatively small costs to consolidated revenue of funding hardship based payments to
        former incumbent licence-owners are avoided; and
       The government faces an additional political risk due to the fact that licence-owners are able
        to make hardship based arguments against reform proceeding, as well as arguments based
        on economic loss/uncompensated government "takings".

Final draft: 18 October 2011
3.      Staged releases of licences
A staged move to open entry has been proposed in a number of contexts, primarily as a means of
providing a more gradual transition to open market conditions and thereby ameliorating costs for
incumbent licence-holders and/or governments. In particular, it is often argued that such
approaches have the potential to avoid the imposition of hardship. This option is normally
formulated in terms of a program to issue a defined number of licences annually over a set number
of years, with licences to become freely available at the end of the transition period. It is generally
assumed that, at the completion of the transition period, the pre-existing excess demand for
licences will have been largely removed and, with it, the value of licences. Thus, the move to open
entry represents a small further adjustment, in practical terms.

From the perspective of the achievement of the efficiency benefits of reducing the extent of supply
restrictions, the fundamental consideration in relation to a staged program of licence release is that
of the rate and timing of licence releases. From the perspective of the allocation of the costs of
reform, the key questions relate to the allocation of the additional licences and the amount, if any,
paid to incumbents as compensation or adjustment assistance.

No practical example of the successful adoption of transitional strategies of this type have been
identified. Victoria's post-2002 reforms might be regarded as an example of this approach, except
that there was no clear statement made, either at the time of the adoption of the strategy or
subsequently, to the effect that the ultimate objective was to achieve an open entry taxi-market.
That said, it is noted below that objectives of reducing the extent of supply restrictions over time
were explicitly stated in 2003. Thus, Victoria itself can be regarded, in part, as a case study.

Given the lack of practical experience with the strategies, the following discussion is based on the
findings of researchers who have sought to model its expected impacts. A number of different
variants of this model can be identified and are considered in turn below. These are:

Strategies based on the staged release of licences can be divided into two types: those that are
based on all additional licences being issued to incumbent licence-holders and those based on
making additional licences available in the open market. A further distinction in respect of the latter
variant lies in the question of whether the proceeds from auctioning licences are used to partially
compensate incumbents for losses of monopoly rents or are retained by governments.

3.1. Issue of additional licences to existing licence-holders

The general approach of confining licence issues to incumbents is proposed on the basis that, while
the total monopoly rents available across the taxi industry will decline substantially due to the
licence issue, providing the new licences only to incumbents has the effect of ensuring that they
continue to profit from the whole of these rents. Thus, this option minimises the losses borne by

Final draft: 18 October 2011
incumbents, for any given level of licence issue. This is seen as a means of reducing the likelihood
that compensation will need to be paid. Indeed, discussions of this option generally presume that
compensation will not be paid, albeit that there is no conceptual reason for this to be the case.

Two broad variants of this option can be identified. The first is extremely simple, and involves simply
issuing an additional licence to each holder of an existing licence, on a "one for one" basis. This is
generally envisaged as a "one off" licence issue. It is usually expected to be followed after a pre-
announced period by a move to complete open entry, albeit that it is also conceivable that this
approach would be taken without embodying a commitment to an open entry market as the
ultimate goal.

The second variant of this option involves issuing an additional "partial licence right" in relation to
each licence. This would imply that incumbent licence-holders would be issued a right to, for
example, 0.2 licences per licence held. Holders of these rights would be entitled to trade them
freely, allowing them to be aggregated into new licence rights. This variant implies that there would
be regular - perhaps annual - licence issues. Again, such issues are generally expected to continue
until the value of a licence is near zero and an open entry policy can be adopted with little further
adjustment to market dynamics.

The implications of these two variants of the model are considered in turn below.

3.1.1. "One for one" issues

This option necessarily implies immediately doubling the number of taxi licences on issue. It is
therefore only feasible in a context in which the number of licences currently on issue is significantly
less than half of the estimated open market equilibrium number. That is, unless a significant
measure of monopoly rent continues to be available to incumbents after the issue of the additional
licences, no "staged transition" to an open market is possible.

Given that Melbourne currently has approximately 1.0 - 1.1 licences per 1,000 population, while
most open entry markets have more than 3 licences per 1,000 population, the adoption of this
option would be consistent with the maintenance of some monopoly rents and, hence, licence
value. In this context, it can be assumed that the "one for one" issue would apply only to owners of
the 3,100 perpetual, transferable, licences currently on issue. Thus, the impact of adopting this
policy would be to increase total licence numbers by around three quarters, from approximately
4,200 to around 7,300.

This means that a one-for-one issue is at least theoretically feasible in the current Melbourne
context. If it is assumed that the open-entry equilibrium ratio of taxis to population is approximately
equal to 3 per 1,000, the adoption of this policy would be expected to move Melbourne's taxi
industry approximately slightly less than half-way toward this open-entry equilibrium.

Final draft: 18 October 2011
The impact of a "one for one" issue on licence values is, in fundamental terms, determined by two
factors. The first is the size of the move toward an open entry equilibrium position that the "one for
one" issue would represent. The second is the timing of the announced move to open entry22.

Open entry equilibrium number of taxis

While no open-entry taxi markets are known to have significantly lower taxi ratios than 3 per 1,000
population, several have much higher ratios. Thus, it is possible that the relative size of such a move
in the Melbourne context would be smaller than the above estimate implies. The larger the
remaining "gap" between the actual number of licences and the open entry equilibrium number, the
higher will be the value of licences immediately after the licence issue. The high level of demand
increase in the taxi industry relative to population growth in Melbourne in recent years suggests that
the true equilibrium ratio may be greater than 3 per 1,000, so that the impact of an initial "one for
one" issue may be to move the industry less than half way toward the open entry position.

Timing of the move to open entry

Because the licence value is equal to the (risk adjusted) capitalised value of the expected flow of
future monopoly rents available from its exploitation, there is a direct relationship between the
number of years delay until the implementation of open entry and the value of the licences. A
second factor determining the value of licences is necessarily the credibility of the announcement
that an open entry policy will be pursued at the stated time. If this is less than 100%, the licence
value will be increased, ceteris paribus.

Assessment of policy merits

The adoption of this policy would, in the short term, lead to a substantial disruption of the market,
as a large number of new taxis would be added to the fleet in the short term. In this respect, the
initial impact of the policy would be broadly similar to that of an "immediate open entry" option: the
number of licences released initially would be broadly similar to that expected in the event of an
open entry option that included a significant annual licence fee as a transitional element, although
evidently smaller than the number of licences that would be demanded in the absence of such a fee.

The policy can clearly be criticised on equity grounds, in that it provides no opportunity for those
outside the industry to benefit from the issue of new licences. However, the primary justification
advanced for this strategy is that it will minimise the loss of licence value experienced by incumbents
for a given level of new licence issue. Thus, it is promoted with reference to a different conception
of equity.

A key benefit of this approach is that, by minimising the loss of monopoly rents experienced by
incumbents, it is expected that the need to pay compensation may be averted. However, as

  A third factor is whether any additional licence issues would occur in the interim. However, it has been assumed above
that none would occur.

Final draft: 18 October 2011
discussed above, the main determinant of the "one off" loss of licence value experienced by
incumbents is the timing of the announced future move to an open entry licensing regime.

The 1994 analysis undertaken by Gaunt and Black23 suggests that, were the anticipated move to
open entry to occur in fewer than seven years, the impact of this announced change would be at
least as great as the impact of the initial licence release. That is, if it is assumed that the equilibrium
taxi ratio for Melbourne is around 3 per 1,000 population, a "one for one" issue resulting in an
increase in licence numbers from the current 1.0 per thousand to around 1.75 per thousand would,
of itself, be expected to reduce licence values to around $300,000. However, if this were
accompanied by an announcement that an open entry regime would commence, say, five years
hence, the licence value would be expected to fall below $100,00024.

Thus, a government concerned to avoid major windfall losses to incumbents would be expected to
find this approach to reform unappealing: a move to open entry that was sufficiently distant as to
preserve a substantial proportion of the existing licence value would also be too far distant from the
present to be credible. This option therefore seems to have little merit, in practice, in terms of
weakening any argument for the payment of compensation.

The option of implementing the initial "one for one" policy, without committing to a future open
entry policy might be considered more feasible in this context, as it could see 60% or more of the
existing licence value preserved. However, the corollary of this is that the consumer benefit
delivered would also be far smaller than that available under open entry. Indeed, the adoption of
this policy would do no more than return the taxi licence to population ratio to a level equivalent to
that found in several Australian cities in the 1960s. For example, Gaunt and Black find that the
licence ratio in Brisbane at this time was around 1.9 per 1,000, compared with the above estimate
under the application of this policy in Melbourne of 1.75 per 1,000.

Practical experience with the reform strategy

No example of this option being adopted in practice has been identified. The initial reform proposal
developed by the Irish Government was structured broadly along the lines of a "one for one" policy:
It was proposed that each licence-holder would receive one additional licence, with a further 500
licences to be made available to all parties in Dublin.

In the event, however, this policy was abandoned prior to implementation. This was the result of
the position of strong opposition to the policy adopted by hire-car drivers, who sought access to any
new taxi licences to be issued and successfully challenged the legality of the government's policy
proposal in the courts25.

   Gaunt, C. & Black, T. (1994). The Unanticipated Effects of the Industry Commission's Recommendations on the
Regulation of the Taxicab Industry. Economic Analysis and Policy, Vol. 24, No. 2, pp 151-162.
   The licence value in this case would be equal to the discounted value of the assignment fees that could be obtained over
five years, since it would have no value after this time. Given that licence numbers would have increased by 75%, the
assignment price would be expected to fall substantially from the current level of around $24,000.
   Barrett, S.D., (2003). Regulatory Capture, Property Rights and Taxi Deregulation: A Case Study. Economic Affairs,
December 2003, pp 35-40.

Final draft: 18 October 2011

This alternative has been shown to be unlikely to significantly lessen the opposition of incumbent
licence-holders to major reform and to have little practicability ability to cushion incumbents from
experiencing major losses as part of a move to adopt open entry within a feasible time-period.
Moreover, significant opposition to such an option from non-licence-owning parties who seek access
to taxi licences would also be anticipated. This would likely include a large proportion of both taxi
operators and taxi drivers.

3.1.2. Partial rights issues

Abelson (2010) and others have also highlighted the possibility of using the basic mechanism of
issuing additional licences to incumbents to expand supply more slowly, by creating "partial" rights.
This would imply that incumbent licence-holders would be issued a right to, for example, 0.2 licences
per licence held. Holders of these rights would be entitled to trade them freely, allowing them to be
aggregated into new licence rights. The benefit of this option is that it allows for a more flexible
approach to be taken to the issue of additional licences.

Thus, whereas the adoption of the "one for one" model would only allow for a major, one-off
increase in licence numbers, prior to a possible delayed adoption of fully open entry, this variant
allows for smaller numbers of licences to be issued, presumably on a more frequent schedule. For
example, holders of "pre-reform" licences might receive a right to 0.2 licences annually until licence
values are exhausted - likely to be a period of ten years or more in the Melbourne context.
Evidently, this variant of the model more closely resembles the alternative considered below, of
auctioning licences on the open market, in terms of the expected trajectory of licence increases over
time. In theoretical terms, there is little difference between this option and one involving regular
auctions of licences with the proceeds paid to former incumbents: the result in both cases is that
incumbents receive partial compensation for their loss of monopoly rents, while these losses are
extended over a substantial period of time.

However, as demonstrated above, a key driver of declines in licence values is the timing of the move
to open entry. To the extent that this is so, this variant of the option demonstrates the same
weakness as that identified above: any relatively short transition to open entry would create a
situation in which large losses were incurred over a short period of time and would, as a result, do
little to weaken incumbents' "moral" case for compensation.

It is theoretically possible that this option could be combined with a partial compensation payment
in respect of the licence value lost, however, this does not appear to have been proposed by
advocates of this model.

Practical experience with the reform strategy

No jurisdiction is known to have implemented this transitional strategy.

Final draft: 18 October 2011

In practical terms, this option is broadly equivalent to the option of auctioning set numbers of
licences on a regular basis and using the proceeds to pay partial compensation. Incumbents will
receive little benefit from the adoption of this option unless a very long time horizon is adopted.

3.2. Open market auction of additional licences

An alternative to issuing licences without charge to incumbent licence-holders is to auction a pre-
determined and clearly advertised number of additional licences on the open market. It is generally
assumed that a given number of licences would be auctioned annually, with the long-term evolution
of the auction program being established in advance, with the volume of licence issues being
calculated to enable a seamless move to an open-entry market within a defined timeframe.. The
credibility of this program is essential to maximisation of the revenue from licence sales, since any
fear on the part of potential purchasers that actual licence releases could be greater than scheduled
will clearly reduce auction receipts substantially.

No example of this model being adopted in practice has been identified. However, Gaunt and Black
(1994)26 modelled the application of this model to the Brisbane taxi market, using an assumed seven
year transition to an open-entry market. This paper was published in response to the then Industry
Commission having recommended such an approach as a feasible reform option in a 1993 report on
urban transport.

 Importantly, Gaunt and Black concluded that, even were the total value of licence auction receipts
to be returned to incumbent licence-holders, the total amount of compensation that could be
financed by these means would amount to less than one third of the pre-reform licence value.
Gaunt and Black concluded that this model represented a poorly performing approach to reform,
since it substantially delayed the achievement of much of the efficiency benefit of reform while
delivering only a relatively small amount of compensation to incumbents.

It should be noted, however, that the major determinant of the revenue obtained from licence
auctions (and potentially available for redistribution by way of compensation to incumbents) is the
time horizon for the move to open entry market conditions. As noted above, Gaunt and Black's
calculations were based on a seven year time horizon. A longer time horizon would increase
receipts significantly. However, it would simultaneously postpone the capture of the benefits of
reform and greatly increase the risk of the reform program being derailed prior to completion due to
successful lobbying from incumbents.

Box 4: Montreal taxi licence buy-back27

  Gaunt, C. & Black, T. (1994). Ibid.
  Trudel, M (1995). The Fundamentals of Taxi Regulation and the Quebec Experience. Paper presented to the 7th
Congress of the European Taxi Confederation, Donostia - San Sebastian, Spain, February 1995

Final draft: 18 October 2011
The experience of the Montreal licence-buyback, organised and funded by licence-holders in that city between
1985 and 1990 provides a real example that underlines Gaunt and Black's conclusions, albeit that the direction
of change is reversed. In that case, licence holders collectively repurchased 25% of the licences on issue in the
Montreal taxi market over a 5 year period and retired them from use.

The buybacks occurred at market value and this value rose from $10,000 at the commencement of the program
in 1985 to $30,000 at its finish in 1990. By 1995, licence values had increased to $55,000. Thus, industry
participants had taken a rational decision to purchase a higher degree of monopoly rents for their existing
licences. Remarkably, the buyback was supported by the regulatory authorities, rather than being the subject
of legal action by the competition authority, as might be expected.

Arguably, were a "one for one" licence issue to be put into practice, there would be a significant risk of such
collective action being taken by incumbents to defeat the purpose of the policy. As demonstrated above, it
would be in the interests of incumbents to agree collectively not to either exploit the additional licences granted
or to make them available for sale. The apparently highly concentrated nature of the taxi industry, which is
characterised by numerous cross-ownership relationships and other interdependencies, suggests a higher
likelihood that such action would be feasible than would be the case in most industry contexts.

This risk of action by incumbents to defeat the purposes of a licence issue program theoretically attaches to all
variants of progressive licence issue, other than those that are formula based. However, governments would
arguably be better placed to take action in response in the context of a progressive licence issue policy, as
distinct from a "one off" licence release policy.

Melbourne licence issues - post 2002

The regular licence issue program commenced in Melbourne from 2002 arguably constitutes a case
study of this reform option. As noted above, no explicit statement was made to the effect that an
open-entry market was the desired endpoint of the reforms. However, a number of official
statements identified objectives involving a substantial improvement to existing demand/supply
balance. Thus, in 2003 the Victorian Government28 reported to NCC that the program would:

        break the link between the provision of taxi services and the value of the taxi licence as an
         improve the current demand/supply balance;
        lead to a gradual decline in the value of taxi licences over time; and
        prevent future "windfall gains" being obtained by licence-holders.

The report stated that the announced program did not constitute the endpoint of reform and that,
should the licence issue program not achieve these stated objectives, additional 24 hour licences
would be issued.

   Department of Treasury and Finance (2003). National Competition Policy: Report for the 2003 Assessment
on Victoria’s Implementation of National Competition Policy, pp 91-4. See:

Final draft: 18 October 2011
Comparison of these identified objectives with the actual performance of the post 2002 reforms
highlights the practical difficulties involved in implementing a progressive licence release option.
The first of these is a systematic tendency, particularly in the presence of strong lobbying by
incumbents, to systematically underestimate the extent of licence release required in order to
achieve a given policy objective. Despite advice to the contrary by the NCC, the Victorian
government apparently believed that the issue of 100 licences annually from 2003 to 2008 and 150
licences annually from 2009 to 2014 would be sufficient to achieve the above objectives29.
Experience has demonstrated that this is not so. However, a more instructive calculation is based on
determining what level of licence release would have been required over the period to achieve the
desired goals.

As a starting point, the number of licences required to achieve an open entry market equilibrium can
be estimated. The average taxi/population ratio in New Zealand cities, 20 years following
deregulation, is in the vicinity of 3.0 and 3.5 taxis per 1000 population, while in Ireland the
equivalent ratios are as high as 6.0 per 1,000. Taking the New Zealand ratios as a reasonably
conservative estimate and applying this to the current Melbourne population of around 4.0 million
suggests an open entry equilibrium number of taxis of between 12,000 and 13,500. Given that
there were fewer than 3,500 licences on issue in 2002 (including 3,100 transferable general
licences), this implies that between 8,500 and 10,000 licences would need to have been issued over
a 12 year period in order to approximate an open entry equilibrium over that time.

This is equivalent to the issue of between 700 and 830 licences per annum, approximately, over a
period exceeding a decade. By comparison, programmed issues were to average 125 per annum
over the period.

These estimates suggest that, even if a move to an open entry market was never in contemplation,
in order to achieve the stated objectives of an improvement in supply/demand balance and a
gradual reduction in licence values over the reform period30, it would have been necessary to release
around four times as many licences as were called for under the program.

These notional calculations highlight a fundamental feasibility issue in relation to the design and
implementation of staged licence issue schemes. Estimates of the number of licences required at
the given "endpoint" tend to be based on static analyses or unduly conservative demand growth
estimates, while it is politically difficult to convince stakeholders that a tripling in the taxi/population
ratio is needed, especially in the face of concerted lobbying to the contrary from stakeholders. As an
example of the former dynamic, in the lead-up to the Irish taxi reforms, the Oscar Faber Report
prepared for the Irish government in 1998 argued that a fleet of 5900 taxis would be needed after

   While not generally publicly acknowledged, the documentation provided to the NCC stated that at least half of the peak
services licences issued in the first six years would, on the 6th anniversary of their issue, be converted to 24 hour licences
and replaced by additional peak service licences. This was expected to lead to a total increase of 46 per cent in actual
licence numbers. See Department of Treasury and Finance (2003), op. cit., p 7.
   For example, even a modest $10,000 annual reduction in nominal licence values from their $280,000 starting point in
2002 would imply that the 2014 licence value would have been no more than $160,000, or less than one third of its current
value. This would appear to imply a taxi fleet of at least 9,500 on the above calculations.

Final draft: 18 October 2011
10 years in order to service the Dublin market. In the event, two years after deregulation, 9000 taxis
were operating in Dublin, while these numbers have subsequently been maintained.


Staged licence issue programs provide a theoretically feasible means of reducing the impact of
reform on licence-holders, by both allowing revenue to be attained by government from licence
auctions and redistributed to licence holders as partial compensation and by allowing some
monopoly rents to continue to be earned over a number of years. However, the cost of following
this path lies in the substantial delay in achieving the majority of the benefits of reform.

More importantly, two systemic issues suggest that the final objective of this strategy is unlikely to
be met. First, the licence issue program is likely to be miscalculated, leading to far too few licences
being issued, as in the Victorian case post 2002. Second, if the transition period is long enough to
substantially smooth the adjustment path for incumbents, it will also lead to a high risk that
successful lobbying will lead to its abandonment prior to completion. Another Victorian example,
that of the post-Foletta report reforms of the late 1980s, provides a practical example of this risk
translating into reality.

4.       Formula-based issue of licences

4.1. Description and rationale

The fundamental concern with mechanisms such as those discussed above, which involve
determining ex ante the number of licences to be issued annually over a given period, is that they
imply a high risk that the underlying objectives of reform will not be achieved. This will occur if the
annual licence issued required to achieve the reform objective over the identified period is under-
estimated by policy makers during the project design phase. Such under-estimates are highly likely
to occur, for several reasons:

        A static analysis of the taxi market may be employed, which does not account for increases
         in demand due to service quality improvements and/or price falls resulting from reform;
        Ongoing increases in taxi demand due to factors exogenous to the taxi market are under-
         estimated (frequently due to inadequate data);
        Strong lobbying from incumbents to minimise the extent of additional licence issue.

Moreover, there are impediments to policy-makers correcting these initial under-estimates
following initial program implementation. In particular, there is likely to be ongoing lobbying from
incumbents to “honour commitments” made to a particular program of licence issue at the outset
(i.e. in preference to increasing issues), together with a perceived need to maintain the predictability
of the reform program in the interests of “market stability”.

Final draft: 18 October 2011
The experience of the current (post 2002) Victorian taxi reform program provides a clear example of
this dynamic. Advice was provided by the National Competition Council prior to the announcement
of the policy indicating that the proposed quantum of licence issues would be insufficient to improve
supply/demand balance over time. This was based on analysis of historical demand growth in the
market. However, no change was made to the initial proposals in response.

In 2003 the Victorian Government31 reported to NCC on the objectives of the reform program and
the “flexibility” said to be contained in it. It indicated that a key objective of the reform was to break
the link between the provision of taxi services and the value of the taxi licence as an asset and that
the reforms were expected to improve the current demand/supply balance. It also stated that it was
expected that taxi licences would gradually decline in value over time as a result and that it would no
longer be possible for licence-holders to obtain “windfall gains”. In particular, the report stated that
the announced program did not constitute the endpoint of reform and that, should the licence issue
program not achieve these stated objectives, additional 24 hour licences would be issued.

In the event, while none of the above objectives was achieved, with licence values almost doubling
over the ensuing nine years, no further licences were issued32.

A formula-based licence issue program provides a plausible means of overcoming these systemic
problems. The key decision to be made in this context is what criterion or criteria will be used to
identify the appropriate number of licences to issue. Several possibilities can be identified, based on
both practical experience and theoretical analysis. In terms of practical experience, three OECD
countries currently adopt formula-based licence issue programs, as do New South Wales and the
Northern Territory. Four of these five jurisdictions adopt a formula based on a single criterion, while
the fifth adopts a formula based on multiple criteria. In addition, Bruce Schaller, a former taxi
regulator, has adopted an alternative, multi-criteria formula. These options are discussed in turn.

The purpose of a formula-based approach is clearly to provide a decision-rule that consistently
provides an appropriate matching of taxi supply and demand. The key benefit of a formula based
approach is that the existence of such a rule removes decision-making on taxi licence issue form the
arena of subjective political or administrative decision-making and, hence, reduces or eliminates the
risk of regulatory capture. However, the formula-based approach will only provide positive
outcomes if two conditions hold:

         The criterion or criteria used constitute a good proxy for taxi demand (or demand growth);
         the "benchmark" value(s) of the criterion or criteria are appropriately set.

   Department of Treasury and Finance (2003). National Competition Policy: Report for the 2003 Assessment
on Victoria’s Implementation of National Competition Policy, pp 91-4. See:
   The “one off” issue of 200 10 year licences and 330 WAT licences that occurred in 2010 appears, in effect, to have largely
substituted for the programmed releases of 200 peak period licences during 2010 and 2011, together with the anticipated
conversion of 100 peak period licences to full licences over the same period and their replacement with a further 100 peak
period licences, as envisaged under the original reform program. When other failures to carry out aspects of the initially
announced reforms are taken into account, the total number of licences issued from 2002 to 2011 remains less than that
originally envisaged.

Final draft: 18 October 2011
4.2. Population ratio based formulae

The most commonly used criterion is that of population ratio. This is used in Flanders and Romania,
as well as the Northern Territory and some German cities33. Flanders uses a ratio of 1:1,000, in
common with the Northern Territory34. However, Romania uses a ratio of 4:1,000.

No information is available as to the reasoning behind the adoption of the 1:1,000 ratio in Flanders.
However, it is notable that, when the Northern Territory adopted this ratio in the course of reversing
its recently adopted open entry reforms, it chose a ratio that represented the actual pre-reform ratio
of taxis to population. This suggests that decision-makers opted to try to revisit the pre-reform state
of the industry. Alternatively, the fact that the government chose to continue to capture the
available monopoly rents through maintenance of a large annual licence fee can also be seen as
providing a strong incentive for it to choose a relatively low ratio, since this would maximise the
available rents and, hence, the size of the licence fee that can be charged. Recent significant
increases in the fee provide some support for this hypothesis.

In practice, Romania's 4: 1,000 ratio is a legislated national maximum, with local authorities being
permitted to impose tighter restrictions35. Notably, the 4: 1,000 ratio is broadly in line with the taxi
ratios found in cities characterised by open entry markets. Conversely, the 1: 1,000 ratio adopted in
Flanders and in the Northern Territory is broadly equivalent to the minimum ratios found in cities
with ad hoc licence issuing arrangements, such as most Australian capitals.

While only a small number of observations are available, it is clear that the outcomes of the
adoption of a population ratio based formula can, and do, vary widely. On one view, observation of
actual ratios found in open-entry markets should provide a benchmark which would allow decision-
makers to determine an appropriate ratio. Observation across many markets with tradeable
licences of a clear inverse relationship between licence values and taxi: population ratios might also
be thought to provide useful guidance to decision-makers charged with selecting an appropraite
ratio. However, the above observations provide no evidence of a "clustering" of outcomes and no
real evidence that the actual ratio of taxis per head of population is systematically higher in
jurisdictions adopting this criterion as part of a formula based approach than is the case for cities
with ad hoc approaches to licence issue.

Thus, the practical performance of this type of formula based approach cannot be judged to be
particularly positive. At a theoretical level, the problem may be that the administrative mechanisms
put in place have not incorporated any specific decision-rules to be used in determining an
appropriate ratio.

   Bekken, JT. (2007). Experiences with (de)Regulation in the European Taxi Industry. In OECD/ECMT (2007) (De)-
Regulation of the Taxi Industry. See pp 40-41. No information is available on the ratios used in German cities or the
rationales for their adoption.
   Strictly speaking, the NT ratio is 1.11 per 1,000.
   In practice, however, there seems to have been relatively little use made of this power by local authorities: according to
Barrett, taxi numbers in Romania more than doubled since 1989. See: OECD/ECMT (2007), op. cit., p 136.

Final draft: 18 October 2011
More broadly, it is clear that a city's population size provides only a very loose approximation for taxi
demand and that the "optimal" ratio of taxis to population is likely to change over time, perhaps
substantially. The recent history of Melbourne provides an obvious example, with demand growth
having substantially exceeded population growth over a long period36. Moreover, regression
analysis conducted by Schaller (see below) finds that population is not a significant determinant of
taxi demand, once other key variables have been controlled.

In sum, there is no single "optimum" ratio of taxis to population and, consequently, little reason to
believe that the adoption of a formula-based approach using this criterion will necessarily yield a
positive outcome. While the general rationale for a ratio-based approach is that it can function as a
preventative against declining relative supply due to ad hoc decisions being made over time by
decision-makers who demonstrate degrees of capture by the industry, the practical experience with
this approach, albeit limited, provides little support for it.

4.3. Load factor based formula

One OECD country - South Korea - uses a formula based on the utilisation rate of taxis to determine
licence issue decisions. This formula requires that licences be issued where load factors fall below
55% (for cities of more than 500,000), or 50% for smaller cities. In addition, licences will be issued if
operation rates of the taxi fleet falls below 80%.

Such criteria, in particular the former, arguably better reflect taxi demand than does a simple
population ratio. To the extent that this is so, this option potentially provides a superior decision-
rule. In practice, the adoption of this approach does seem to have coincided with a slight decline in
the real value of Korean taxi licences and, by implication, an improvement in supply/demand
balance. Kang reports that taxi licences were valued at around $65,000 in 1998, whereas OED
(2007) reports that the then current value of a licence was in the range A$25 - 65,000 approximately.
This broadly static nominal licence value over a decade suggests a significant decline in nominal
values over the period, contrary to the experience of most taxi markets with controlled entry.

Given that there is only a single observation of the use of this criterion in practice, it clearly cannot
be concluded that this represents a superior approach or that it will systematically perform well in
practice. That said, it is clear that load factors constitute a more direct and reliable measure of
supply/demand balance than does the taxi/population ratio.

On the other hand, there is clearly room for substantial disagreement as to the specific occupancy
rate that should trigger taxi licence issues. From a theoretical perspective, the "optimum" rate is a
function of consumer preferences measured on a price/quality continuum. That is, lower occupancy
rates are necessarily associated (cet. par.) with lower waiting time, which is a key element of service

  For example, the Victorian NCP review of taxi legislation found that demand had increased at an average of 3% per
annum between the early 1980s and the late 1990s - substantially higher than the rate of population growth. Similarly, the
strong post-2002 inflation of taxi licence values, occurring in a context of licence numbers being increased broadly in line
with population, implies that this trend has continued.

Final draft: 18 October 2011
quality, while also implying higher cost structures. Thus, different populations with different
valuations of their waiting time will implicitly prefer different load factors.

Again, it is not clear how a regulatory authority would go about systematically assessing these
consumer preferences and determining an optimum load factor to maintain as a target over time.
Unless such a mechanism can be posited, there is no basis for believing that a licence issue formula
based on this criterion will systematically achieve good outcomes over time.

4.4. Multi-criteria based formulae

Sydney model

Sydney constitutes the only currently known example of a multi-criteria based formula for licence
issue. The Sydney model was adopted very recently, so that there is little possibility of assessing its
practical performance over time. However, it can be assessed in terms of a theoretical consideration
of the criteria used.

The Sydney model uses ten criteria, as follows:

Table 1: Sydney taxi licence issue formula - criteria used

Key demand driver                                      Components
Likely taxi demand (including latent demand)           State final demand
                                                       Sydney population size
                                                       Unemployment rate
                                                       Sydney airport passenger numbers
                                                       Total network bookings
Industry viability and sustainability and demand       Value of licences
for new licences                                       Plate lease costs
Performance of existing taxi services                  Annual average pick up time (mins)
                                                       Percentage of pickups within 15 mins
                                                       Percentage of "no cars available"

This formula clearly includes elements of both the "population ratio" and "load factor" based
formulae discussed above, while also taking a broader range of factors into account. Some bear a
very indirect, or even tenuous relationship to taxi demand: the unemployment rate and Sydney
airport passenger numbers arguably fall within this category. Conversely, the "performance of
existing taxi services" criteria are clearly directly related to actual service quality, while the "industry
viability and sustainability" criteria are direct indicators of relative supply/demand balance, since taxi
plate values and assignment rates move directly with expected monopoly rents.

Final draft: 18 October 2011
Despite the varying relevance, or importance, of the different criteria, they receive a relatively even
weighting in the model: while there are some changes from year to year, eight criteria received the
same (10%) weighting in 2011/12, while the unemployment rate received a zero weighting and state
final demand a 20% rating.

A likely impact of adopting the multi-criteria approach appears to be to reduce the likelihood of
large increases in plate numbers occurring in any given year. As an example, licence releases based
solely on the values of the performance indicators in 2011/12 would have totalled 5.4% of the fleet,
compared with the actual 3.6%.

However, the fundamental issue with the Sydney model is that its starting assumption is that the
current number of taxi licences is in some sense optimal and that the purpose of the formula is to
maintain this existing level of supply/demand balance over time. Given the clear evidence of excess
demand - demonstrated by the $400,000 value of a Sydney licence - this is not a supportable

In sum, the Sydney formula is, as demonstrated above, potentially consistent with further declines in
service quality but, when its recommendations are considered in the light of the historical record of
licence releases in Sydney, it is plausible that it will result in a greater degree of maintenance of
current levels of supply/demand balance than would have occurred under a continuation of the
former, ad hoc approach to new licence issue.

Schaller37 model

Schaller (2005) modelled taxi demand by conducting regression analysis on six variables, using data
from 118 US cities. Schaller argues for the use of three variables to determine taxi demand and,
hence, required taxi numbers. These are the number of workers commuting by subway, the number
of households with no private vehicles and the number of airport taxi trips.

This conclusion is based on regression analysis conducted on six variables, of which the above three
are found to be closely correlated with the observed number of taxis in a range of cities studied. He
reports that the results show all three variables as being statistically significant at the 95%
confidence interval38. However this, apparently impressive, result is rendered substantially less
credible by both a lack of transparency as to his methodology and results and circularity in his
argument. He states that:

         The predicted number of cabs closely matches the actual number in cities that, based on
         separate information, appear to have an appropriate number of cabs. (Schaller (2005 ) p
         73).[emphasis added]

   Bruce Schaller is a former New York City taxi regulator and proponent of restricted entry to taxi markets. See: Schaller,
B. (2005), A Regression Model of the Number of Taxicabs in US Cities. Journal of Public Transportation, Vol 8, No 5, 2005.
   Interestingly, Schaller concludes that population is not a determinant of taxi demand, once these three significant
variables have been controlled for.

Final draft: 18 October 2011
This statement strongly suggests that the statistically significant results relate only to a handpicked
subset of the cities studied. Moreover, Schaller provides little detail on what separate information
was used to determine which cities were judged as having “an appropriate number of cabs” and
which were judged to have “undersupply” or “oversupply”. Arguably, the “appropriate” number of
cabs may have been defined in a way that would maximize the apparent predictive power of the
model. This suspicion is heightened by the fact that Schaller does not even report the number of
cities in respect of which the model’s predictions do and do not accord with observed reality, instead
relying on a small number of partial observations39.

In sum, there is strong doubt as to the actual predictive performance of Schaller's model, while no
city is, to date, known to have adopted it as the basis for determining its licence release program.

4.5. Subjective criteria

Bekken states that:

         Subjective criteria are the most common criteria for issuing new licences. In such cases, an
         authority assesses the need for an increase or decrease in the number of licences. Often
         these criteria are related to terms such as public need, excess demand, excess driver profits
         and other societal reasons. The real differences between objective and subjective criteria
         may be small. To make the decision more “objective”, different studies or investigations may
         be conducted. Nonetheless, the authority or the politicians will have the last word.40

As this quote suggests, the use of such criteria provides, at best, only minimal guidance to, or
constraint on the discretion of, decision-makers in determining what numbers of taxi licences should
be issued from time to time. Review of the Transport Act 1983 shows that it sets out a large number
of such criteria as the required basis for Ministerial decision-making in relation to licence
applications. The dismal performance of the Victorian legislation in this regard underlines the lack of
practical utility of such subjective criteria.

4.6. Discussion

Formula-based approaches to licence issue do not constitute "transitional strategies" as such. In
fact, they share the characteristic of being based on the maintenance of some form of status quo,
whether specified in terms of population ratios, service standards or other criteria. There are
therefore 2 concerns with this approach:

        While the formula might be used to specify a desired end-point, a transitional strategy is
         still required if this desired end-point is a significant distance from the current position and

   These are that: The predictions are within 10% of observed numbers in “Denver, Los Angeles, San Diego, St Louis and
several smaller cities or counties”; “model predictions are within 7% of the actual number of taxicabs in several
jurisdictions that do not regulate the number of taxicabs..[including] Orange County (California), Phoenix, Newark and New
York City (the latter three cities including open-entry livery vehicles in the vehicle count”; but “the model predicts 29 to
55%fewer cabs than are actually licensed in Dallas, Houston and Washington DC”; and “the model predicts 79 to 128%
more cabs in Boston, Montgomery County (Maryland) and San Francisco”.
   Bekken (2007), op cit, p 42.

Final draft: 18 October 2011
        The formula based approach contains an implicit rejection of open entry as the appropriate
         regulatory position. Hence it is necessarily a 2nd best option that can only (in the best
         cases) approximate the benefits of that alternative.

None of the above practical examples of the use of the formula appears to be predicated on a desire
to move toward an approximation of an an open entry taxi market outcome. Rather, they are
explicitly based on an objective of maintaining a certain supply/demand balance. Thus, a formula
based approach to licence issue that was adopted as a transitional strategy toward achievement of
an open-entry market would appear to be unique if implemented.

That said, such an option is clearly feasible. One formula based approach to achieving open entry
would involve estimation of the ratio that approximates an open-entry equilibrium, with licence
issues programmed so as to achieve steady progress toward achievement of this ratio after a set
number of years. Such a model would not necessarily lead to the achievement of a seamless
transition to open entry, since the open-entry equilibrium ratio of taxis to population may be under-
estimated at the outset. Thus, further significant entry may occur at the end of the transitional
phase. However, this does not pose a significant problem for the reform program unless a very
substantial under-estimate occurs.

Alternatively, the formula could be based on achieving a certain price-path for taxi licences. As
noted above, the Victorian government indicated in 2003 that its intention was to achieve a gradual
reduction in licence values over the course of its 12 year reform program. A potential alternative
program design would have seen an unlimited number of licences made available each year at a pre-
determined price, with this licence price being reduced each year. Thus, given a pre-reform price of
around $300,000 and a 12 year time horizon, a price of $275,000 could have been set in year 1,
declining by $25,000 annually, until licences were available at administrative cost from year 12
onward. The number of licences to be issued each year would not be pre-determined, but would be
dependent on the number of purchasers willing to pay the set price. This approach ensures that
sufficient licences are issued to compensate for increases in taxi demand, as well as achieving a
degree of progress toward the objective of rebalancing supply and demand.

5.       Other reform options

Licence tenure and transferability


A significant regulatory change made in many jurisdictions in the 1980s and 1990s, as high taxi
licence values arose, was to move to make licences readily saleable. More recently, some
commentators have argued that this change should be reversed as a key element in a reform

Final draft: 18 October 2011
In general, economics recommends that assets should be freely tradable, since this ensures that
they flow to their most productive use. In the taxi licence context, the argument for transferability
was that, in the presence of supply restrictions, ensuring that those who can exploit licences most
efficiently can obtain them will improve service standards. This argument is sound within its own
frame of reference, but is arguably rendered nonsensical when it is observed that the basic premise -
that taxi licences are in finite supply - is indefensible.

Conversely, it is clear that transferability has contributed to the observed inflation in licence values,
for two reasons. First, those who can most efficiently exploit a taxi licence (or other asset) will reap a
higher return from it and, as a corollary, will bid a higher price for it. Second, the tradability of
licences has meant that expectations of future increases in the capital value of the licence have
become a potentially important driver of the market value of the licence. The recently observed
decline in the rate of return to licence assignments (i.e. annual assignment price as a percentage of
capital value of the licence) provides evidence of this dynamic: the existence of a 12 year reform
program with specified rates of licence issue allowed industry participants to calculate the expected
impact on relative scarcity over the period, as well as lowering the "risk premium" attaching to
licences, since the likelihood of additional, unanticipated licence issues was much diminished by the
existence of a long-term program.

Calls to remove licence transferability are generally accompanied by proposals that only licensed
drivers should be able to hold licences. These proposals have at least two rationales. First, some
believe that the removal of "passive investors" from the industry and a return to active participation
by licence-owners is likely to improve service standards. Second, it is argued that such a move could
facilitate further reform. Announcing a requirement that only full-time drivers can own licences
would, on its own, be expected to lead to a significant one-off reduction in licence values, if the
requirement were able to be enforced reliably. It is also expected to lead to a greater focus on the
licence as a "tool of trade" rather than a financial asset. If combined with a delayed move to non-
transferability, the value of the licence in the short term would decline further.

These two factors are seen as being likely to reduce significantly the extent of opposition to further
reforms in the shape of substantial additional licence issues. However, even if these arguments are
accepted, removing licence transferability can also have negative implications for reform. In
particular it obscures the need for new licence issues, since the market price of the licence is the
most reliable gauge of the size of the monopoly rents being earned and, hence, of the degree of
relative scarcity. This means that, in the absence of a transparent gauge of relative scarcity, taxi
regulators may, in fact, be less likely to expand supply over time, rather than more likely to do so.


A significant shift in policy adopted at the time of the 2002 Victorian reforms is the move from
issuing permanent, transferable licences to issuing time-limited, non-transferable licences.

It can be speculated that a part of the intention in issuing non-transferable licences was to
progressively neutralize licence holders as anti-reform agents, as an increasing proportion of licence-

Final draft: 18 October 2011
holders became holders of licences with no exchange value. However, any such erosion of the “anti-
reform lobby” would:

       Only occur in the very long term; and
       Be limited in extent, since PS licence holders still have substantial incentives to oppose the
        issue of larger numbers of new licences.

In the context of the 2002 reform package, full implementation would have meant that, even after
12 years, only around 35% of licences would be non-transferable and time limited. If this approach
were maintained, the transferable licences would continue to become progressively scarcer (and
more valuable, for a given taxi/population ratio) over time. However, the owners of permanent
licences would continue to be a powerful lobby for a very long period.

In sum, the issue of whether licences should be perpetual and transferable or time-limited and non-
transferable is only relevant in a context in which moving to a free-entry regulatory system has been
rejected as a short or medium term policy option. It is therefore a strategy that has to be assessed
on second best grounds. Such an assessment reveals that the use of non-transferable licences is as
likely to inhibit moves to increase licence numbers in the medium to long term as to favour them.

Final draft: 18 October 2011
6.        Easing entry to the hire car industry
The NCC noted in 2002 that one possible strategy for addressing the undersupply of taxis is to act to
favour the expansion of the hire car industry. The distinction between a taxi and a hire car is
exclusively the product of regulatory arrangements: the main additional restriction applicable to hire
cars, vis-a-vis taxis, is that they are unable to operate in the rank and hail market41. However, in
practical terms, hire cars and taxis provide the same service within the pre-booked market, which
accounts for approximately 60 per cent of demand in the Melbourne market (Firecone Ventures
2008). Moreover, increasing the supply of hire cars would be expected to lead to some diversion of
taxis away from the pre-booked market and toward serving the rank and hail market. By this
mechanism, supply in all of these markets could be expected to increase.

A "two track" model, in which there are large numbers of both taxis and hire cars, can be observed
in a number of major cities worldwide. Two obvious examples are New York City and London. In
both cases, the "hire car" market (called car services in the former and mini-cabs in the latter city)
represent a very large proportion of total "taxi" supply. In both cases, this situation has developed in
the context of strong entry restrictions to the taxi market proper - albeit that these restrictions are
quantitative in the case of New York City, where no new taxi medallion has been issued since 1937,
and qualitative in the case of London42.

Licence fee reductions

Melbourne is clearly another city in which strong quantitative controls have existed in the taxi
market over a long period, yet the relative size of the hire car sector is substantially smaller than in
these two cities: the current number of hire car licences in Victoria as a whole is around 900,
compared with over 5,000 taxi licences43. One possible explanation for this difference is that hire car
licences have, since 2004, been sold by the government for a fee initially set at $66,000 and
subsequently reduced to $60,50044. To the extent that this fee is seen as an impediment to entry to
the industry, reductions might be expected to assist in securing greater competition for the taxi
industry from the hire car sector.

However, the historical record does not suggest that reductions in the fee would be likely to have a
substantial impact. In practice, the hire car fee has fallen significantly since 2004, both in absolute
and relative terms, yet entry to the industry has been limited.

Table 2: Changes in Melbourne taxi and hire car values - 2004 to 2011

   Higher vehicle standards constitute a further additional restriction on hire car operations, although this appears to be of
lesser significance to their competitive position in the market.
   Completion of "The Knowledge" typically requires 2 - 4 years' study and requires several, successive examinations to be
passed. This represents a significant regulatory barrier to entry. See:
   Taxi Industry Inquiry (2011). Setting the Scene. P 19.
   Prior to 2004, the licences were available at a nominal cost, subject to satisfaction of a "public interest" test, similar to
that historically adopted in relation to taxi licence applications.

Final draft: 18 October 2011
                         Taxi licence                  Hire car licence          Hire car licence as % of taxi licence
2004                     $340,000                      $66,000                   19.4%
2011                     $510,000                      $60,500                   11.8%
Change                   +$170,000 (50%)               -$5,500 (-8.3%)           -40%

The current fee of $60,500 is currently 8.3% lower in nominal terms than in 2004, while the real
value of a hire car licence is currently approximately 25% below its level of 7 years ago45. Moreover,
while the value of the hire car licence has fallen, the taxi licence value has grown by around 50%
since 200446. Thus, the value of a hire car licence has fallen from 19.4% of the value of a taxi licence
in 2004 to around 11.8% of the value of a taxi licence currently. Hire car licences are thus 40%
lower, relative to the value of a taxi licence, than seven years ago. Moreover, in recent times, hire
car licences have been available at a fee of $40,000 to operators willing to operate hybrid vehicles,
implying an even greater decline in licence costs47.

Despite these reductions in both absolute and relative licence fees, the rate of growth in the sector
has been relatively low: the current total of approximately 900 hire cars compares with a 2004 total
of 53848, indicating that entry has totalled around 360 in seven years, or around 7.6% per annum on

These observations suggest there must be some doubt as to the likely effectiveness of a strategy of
reducing the hire car licence fee as a means of encouraging greater competition for the taxi industry
from this sector.

Removal of other regulatory restrictions

The question of why the hire car sector has been less successful in expanding to compete with the
taxi sector than in other cities, such as London and New York City, that are also characterised by
severe supply limits in the taxi market, was considered in the ESC's 2004 review of the hire car
licence fee. The ESC noted that hire cars had historically focused on the provision of a premium
service at higher prices than taxis, particularly to the business and tourism sectors. This was
attributed in part to regulatory factors including higher vehicle standards and limits on the minimum
period of hire.

However, the ESC noted that a number of these restrictions (including the minimum hire time) had
been removed or made less stringent over time. Moreover, the ESC noted evidence that country
hire cars in particular had increasingly moved to embrace a price-sensitive "mainstream" market,
rather than simply serving a premium market, stating that they:

   CPI all groups Melboune - June 2011/June 2004 = 175.6/143.9. The current value of the hire car licence is therefore
equal to around $49,580 in 2004 dollars, or 25% below its then level of $66,000.
   See ESC (2004) Review of Hire Car Licence Fees: Final Report, p 17. The metropolitan hire car licence fee fell from
$66,000 to $60,500 in response to the ESC Report. See also: ESC (2005), p 23 (2004 value of a taxi licence).
   Moreover, the price of a hybrid vehicle would, in many cases, be lower than that of "traditional" hire cars, as would its
operating costs, suggesting further reductions in overall business costs.
   ESC (2004), p 17.

Final draft: 18 October 2011
            "...currently undertake substantial contract based work for the Department of Veterans
            Affairs (‘DVA’) at prices that are significantly below regulated taxi fares. ...This experience in
            country areas appears to indicate both that hire cars are economically able to compete
            directly with taxis and that some elements of the industry are moving toward serving this
            wider ‘non-premium’ market. This suggests the possibility of a significant further expansion
            in the market for hire car services over time."49

Thus, the ESC was confident that a reduction in hire car licence fees would lead the industry to
provide significantly greater competition to the taxi industry over time. However, the above data
indicates that this hope has yet to be realised. While regulatory restrictions on the sector are less
substantial than was previously the case, action to further reduce these restrictions might, a priori,
be expected to significant in encouraging greater service expansion. The key potential reforms in
this regard to be:

Restrictions on vehicle type and age

There is no clear policy reason for hire car vehicle requirements to differ from those applicable to
taxis. Removal of the requirement to operate a "premium" vehicle would be expected to have some
impact in reducing operating costs. That said, the size of the cost impact would be likely to be
relatively modest.

Prohibition on use of taxi ranks

The major regulatory difference between taxis and hire cars is that the latter are prohibited from
operating in both the rank and hail markets. While data are scarce, the ESC (2005) reported
estimates that these two markets could account for around 50% of the total taxi market.
Consideration could be given to allowing hire cars to use taxi ranks, thus opening a significant
additional proportion of the taxi industry to competition.


The historical data indicates that a gradual reduction in the real value of hire car licences over a
seven year period has been associated with only modest rates of entry, despite the fact that taxi
licence values have grown strongly over the same period. This inevitably raises some question as to
the effectiveness of licence price reductions as a strategy for enhancing competition for the taxi
industry from the hire car sector. Nonetheless, given that licences currently cost over $60,000, it
could be expected that a move to issue licences at administrative cost would potentially yield a
substantial supply response.

This could be increased further by relaxing other regulatory restrictions on the operation of hire
cars, thus enhancing further their potential profitability. A removal of the prohibition on the use of
taxi ranks by hire cars would be likely to be the most significant initiative in this regard, though
moving to a "level playing field" in terms of vehicle requirements would also have a potential impact.

     ESC (2004), p 23.

Final draft: 18 October 2011
Moreover, were a sufficiently strong supply response to be forthcoming, it could be expected that
network effects would operate (i.e. hire cars would form larger networks, thus improving their
potential response times and making them more competitive in the market for immediate dispatch
by phone). This would further enhance their competitive position and bring forth further entry.

Of course, while the value of the hire car licence is considerably smaller than that of the taxi licence,
it remains significant. This means that incumbent hire car licence-holders could be expected to
lobby strongly against any reduction in its nominal value. A basic problem with using this strategy to
improve taxi supply is that there is no policy or equity basis for acting primarily against the interests
of incumbent hire car licence-holders rather than taxi licence-holders.

Regardless of this issue, there must be doubt as to the size of the likely supply response to a fall in
hire car licence fees and, more specifically, the impact of an increase in supply on competition with
the taxi industry. Previous reports50 suggest that, while there is some competition between the two
sectors in the business traveller market, much of the activity in the hire car market remains focused
on tourism and special events, as has traditionally been the case, rather than on competition with
the taxi industry. However, the absence of sound data on industry operators - itself partly a
reflection of the atomised nature of the industry - means that this conclusion must be somewhat

The current relative prices of taxi and hire car licences must be considered in this context: The hire
car licence value of $60,500 is little more than one tenth of the size of the taxi licence value, despite
the fact that hire cars are able to compete in the 50 - 60% of the taxi market that is accounted for by
phone-booked services. Moreover, the proliferation of secondary networks and the improvements
in communications technologies in recent years would appear to provide opportunities for
businesses to combine both taxis and hire cars in order to meet consumer demand.

Given the above, it appears unlikely that the hire car licence price is itself a major impediment to this
sector providing stronger competition to the taxi industry. Nor are there clearly identifiable
regulatory impediments: while vehicle standards are higher in the hire car industry, and might
reasonably be relaxed, the contribution of these requirements to overall cost structures, and hence
to price competitiveness, seems likely to be relatively small.

A strategy of improving the ability of the hire car sector to compete with the taxi industry would
need to embrace both a reduction in the licence fee and, perhaps more importantly, a consideration
of the nature and extent of current regulatory impediments to such competition. However, a degree
of pessimism regarding the potential effectiveness of this course seems justified, based on the
apparent failure of this sector to take advantage of the opportunities for greater competition with
the taxi sector that have arisen to date.

     See ESC (2004), NCC (2002).

Final draft: 18 October 2011
7.      Conclusion
A wide range of transitional strategies have been proposed as means of rendering politically and
economically more palatable the move from a taxi market that is highly distorted by regulated entry
restrictions to an open entry market - or at least one that has a less distorted supply/demand
balance. Relatively few of these strategies have been adopted in practice.

This means that the relative merits of these strategies must, to a large extent, be assessed on
theoretical grounds. More importantly, however, it means that there is no known example of a taxi
market characterised by severe entry restrictions, like Melbourne's, successfully moving to an open
entry outcome via gradualist change. While at least six OECD countries have made the move to
open entry taxi markets in the last two decades approximately, and several additional countries have
open entry conditions in at least some cities, these changes were, in all cases, achieved through
immediate reforms and, in all but one case (that of the Northern Territory) without the payment of

Particularly given the relatively small number of observations involved, this cannot be taken to imply
that gradualist reform is impossible. However, the absence of even one successful instance of
gradualist reform necessarily casts a major question-mark over the practical feasibility of all of these
models. Moreover, the absence of any case in which compensation has been paid (save NT) and the
clear jurisprudence indicating that there is no legal requirement for governments to pay such
compensation suggest that the inclusion of such elements in a reform strategy is unlikely to enhance

Final draft: 18 October 2011
Appendix 1: Jurisprudence in relation to compensation
While the legal position in different jurisdictions will not be identical, available precedent in this area strongly
suggests that no legal obligation to pay compensation would be found to exist.

The most extensive jurisprudence on this issue arises from the Irish reforms. Legal challenges to the Irish
reforms by licence owners were unsuccessful and ended in clear judicial statements to the effect that there
was no legal obligation on governments to pay compensation. Barrett (2003) notes that:

         In Irish law there are three judgements that taxi licences do not confer property rights and that the
         terms under which they are held may therefore be altered without compensation.

The first of these is the decision in Hempenstall v Minister for the Environment (1994) which states:

         ‘property rights arising in licences created by law … are subject to the conditions created by law and
         an implied condition that the law may change those conditions. Changes brought about by law may
         enhance the value of those property rights …, or they may diminish them… But an amendment of the
         law which by changing the conditions under which a licence is held, reduces the commercial value of
         the licence cannot be regarded as an attack on the property right in the licence – it is the consequence
         of the implied condition which is an inherent part of the property right in the licence.’ (cited in Barrett
         (2003), pp 35-6)

That is, the nature of the “property right” conferred by a regulatory licence is that it is inherently conditional
on any future changes to the law that may alter its nature or value. Consequently, such changes do not
amount to an abrogation of a property right of the kind that would give rise to a legal right to compensation.
This position was subsequently confirmed in two further cases .

A more recent case from the United States reached the same conclusion, while subtly differing in terms of the
reasons advanced. Following the removal of entry restrictions in the city of Minneapolis, judgement was made
in August 2007 in the case of Minneapolis Taxi Owners Coalition vs City of Minneapolis . The plaintiffs’
argument for compensation for the loss of value of their taxi licences was again dismissed. However, in
contrast to the Irish cases noted above, the court distinguished between a property right and an “economic
interest” and concluded that

         “…the license holders do not have a property interest in the value of the taxicab vehicle license on the
         secondary market because the issuance of the license does not entitle them to that value, nor does it
         provide for its legal protection. The license allows its holders to drive a taxi in the City. It does not
         guarantee that the City would indefinitely limit the number of taxi licenses issued. Even under the
         former ordinance, the City was required to conduct a hearing at least every 24 months to determine
         whether “public convenience and necessity” warranted additional licenses. (Compl. Ex. E at fmr. sec.
         341.270(a)). Furthermore, the City, as a law-making entity, is free to amend ordinances as it sees fit.
         Like the trucking authority holders in the Rogers Truck Line case, the taxicab vehicle license holders do
         not have a constitutionally protected freedom from competition.”

   Humphry and others vs Minister for the Environment (2000) and Gorman, Kearns, and National Taxi Drivers Union v the
Minister of State and the Attorney General (2001), a judicial review of the decision in Humphry. See www.bailii.org
   United States District Court of Minnesota, Civil No. 07-1789 (JMR/FLN).

Final draft: 18 October 2011
Thus, while the legal context in which compensation decisions are made varies, there appears to be strong
case law, derived from several industries and applied to the taxi industry case, establishing that compensation
is not payable in respect of losses arising from legislative changes which affect licences or other “property
rights” which are established solely via government regulation in the first instance.

Final draft: 18 October 2011
Appendix 2: Summary of reform strategies pursued in
             major jurisdictions

At least six OECD countries have removed quantitative entry restrictions from their taxi regulatory
structures since the late 1980s, while open entry has also been implemented on a city by city basis in
some countries. In each case, the removal of entry restrictions has occurred via a single reform
initiative, rather than through a process of staged reform. In no case was compensation paid to
incumbent licence holders. In cases in which licences were tradable prior to the reforms, valuations
were generally modest. The exception is that of Ireland, where pre-reform values in the late 1990s
reached £90,000 in Dublin (A$300,000 in current dollar terms).

The following provides information on the reform process followed in relation to five of the six
countries that now have open entry taxi markets nationally. Data for the Czech republic was not


The path to reform in Ireland began with a 1992 Interdepartmental Committee report, which
recommended gradual liberalisation of the industry. This met with a major industry response,
including a blockade of the hometown of the responsible Minister. Dublin City Council approved a
recommendation from its taxi and hire car committee to release 200 new licences in 1995, but
rescinded this recommendation following further lobbying and a blockade. A similar
recommendation from the committee in 1997 also brought forth a blockade.

In 1999, the government proposed to issue 3,100 new licences in Dublin, approximately doubling the
existing number of taxis. It was proposed that 2,600 of these licences would be issued to existing
licence-holders, with 500 being more widely available. The proposed reservation of most of the new
licences to incumbents was strongly opposed by private hackney (i.e. hire car) drivers, who sought
access to the industry. A small group challenged the Minister's proposals in the High Court. The
challenge argued that the legislative discretion conferred on the Minister in relation to licence issue
did not allow him a general right to restrict taxi numbers53. This action was successful, with the
Court finding that the Minister had acted ultra vires in restricting licence numbers for reasons
unrelated to ensuring that licence-holders could meet required quality standards. In Humphrey v
Minister for Environment and Local Government (2000), the court stated:

        … the Minister, in restricting the numbers for reasons unrelated to qualitative standards of
        the vehicles and of drivers has fettered the exercise of the discretion conferred upon him by

  Barry, F. (2009). Politics and Economic Policy-Making in Ireland. Trinity College Dublin. See:

Final draft: 18 October 2011
            Section 82 of the 1961 Act. A quantitative restriction not alone affects the rights of citizens to
            work in an industry for which they may be qualified but it also manifestly affects the right of
            the public to the services of taxis and, indeed, restricts the development of the taxi industry

            Regulations which restrict the number of public hire vehicles contradict the very concept of
            public service. It is, of course, open to the relevant authority to insist on quality as the base or
            threshold requirement in relation to a vehicle license as well as a drivers license. The 1961
            Act does not contemplate the restriction of numbers in order to enforce standards.
            Moreover, there would appear not to be any criteria in the Act, nor in the regulations, by
            which a determination should be made on the number of new licenses to be granted.

Given this outcome, the government determined to immediately open entry to the industry. It also
determined not to pay compensation to incumbent licence-holders. This decision was subject to
court challenge by the incumbent licence-holders, which failed. Barrett (2003)54 comments:

            In Irish law there are three judgements that taxi licences do not confer property rights and
            that the terms under which they are held may therefore be altered without compensation.

The third decision to whcih Barrett refers was an unsuccessful judicial review of the Humphry
decision cited above.

Following the failure of the legal action seeking compensation, further lobbying by former
incumbent licence-holders led to the government adopting a policy of making ex gratia payments on
a hardship basis. A Taxi Hardship Panel recommended payments totalling €12.6 million (£9.9
million). In the event, payments totalling €16,349,000 in hardship payments had been made by
2005, representing less than 5% of the total cost of paying full compensation. Payments were made
to 1,412 persons. There were around 3,900 licences on issue prior to reform, although the actual
number of licence-holders is unknown. Individual payments ranged between €3,000 and €15,000 .

An additional feature of the Irish reform process is notable: The Oscar Faber Report prepared for
the Irish government in 1998 argued that a fleet of 5900 taxis would be needed after 10 years in
order to service the Dublin market. In the event, two years after deregulation, 9000 taxis were
operating in Dublin (Barrett (2003)) and 10,865 in June 201155. This points to an apparent tendency
for formula based estimates of equilibrium taxi numbers of err strongly on the side of
underestimation of the true open market equilibrium number of taxis, perhaps in part because of a
failure to take into account the positive impact on demand of improved taxi service.

Ireland has maintained its open-entry arrangements, in combination with maximum fare regulation.
Further regulatory reforms were implemented after the initial reforms, mainly aiming at

  These decisions are Hempenstall v Minister for the Environment (1992), Humphry and others vs Minister for the
Environment (2000) and Gorman, Kearns, and National Taxi Drivers Union v the Minister of State and the Attorney General
(2001), a judicial review of the decision in Humphry. See www.bailii.org. See also: Barrett, S.D., (2003). Regulatory
Capture, Property Rights and Taxi Deregulation: A Case Study. Economic Affairs, December 2003, pp 35-40.


Final draft: 18 October 2011
strengthening quality regulation. These include the adoption of the Taxi Regulation Act 2003, the
establishment of a Commission for Taxi Regulation in 2004 and a national fare structure in 2006. A
2009 review commissioned for the regulatory agency56 found that trip numbers had risen 82 per
cent from 1997 - 2008 and that the annual value of reduced consumer waiting time was around
€780 million. These consumer benefits were found not to have come at the expense of quality
standards, which were rated as being good. Driver incomes were found to have declined slightly (by
about 5%) since 2005 and drivers were found to be earning well below the current average industrial
wage57. No longer-term driver income comparisons were available. A major shift in relative market
share toward taxis and away from hackneys (hire cars) since the opening of entry was observed and
it was concluded that the previous regulations had led to a major imbalance, or market distortion.
Continued problems in ensuring adequate services for people with disabilities were noted, and it
was concluded that there was "no one policy solution" to this problem. The current approaches of
providing financial support for purchase of WATs and requiring minimum service standards were
found to be appropriate.

The current 2010-2014 regulatory strategy document58 states that the existing regulatory model is
consistent with OECD and World Bank best practices and is expected to continue.

New Zealand

The opening of entry to the New Zealand taxi industry was originally intended to occur in concert
with the wide ranging reform of land transport regulation conducted in 1983. However, successful
lobbying by the industry resulted in the removal of the taxi-related provisions from the 1983
legislation. The legislation created a separate service category called a Passenger Service Vehicle,
with few restrictions on the grant of a PSV licence. In 1986 an application for a PSV licence was
received by an intending "limousine service" operator. According to Gaunt (1996)59:

         "The application was granted. The MOT strongly opposed the granting of the licence to Mr
         Tarr, claiming that a taxi service was being provided. The MOT appealed as far as the High
         Court, where it was ruled that the granting of the licence was legal".

The government of the day nonetheless continued to express dissatisfaction with the performance
of the taxi industry and sought change "from within the Industry". However, a new Minister
instituted a formal review of the industry, leading to proposals for reform being published in 1988
and eventually legislated in 1989.

The new legislation removed both limits on licence numbers and maximum fare regulation, although
it did require that maximum fares be notified to the Minister by service providers. The number of

   Goodbody Economic Consultants (2009). Economic Review of the Small Public Service Vehicle Industry. Commission for
Taxi Regulation, Dublin. See: http://www.taxi-library.org/ireland-taxi-economic-review_2009.pdf
   The latter finding is unsurprising, given the essentially unskilled nature of the occupation.
   Gaunt, C. (1996). Taxicab Deregulation in New Zealand. Journal of Transport Economics and Policy, Vol 30, No. 1, pp

Final draft: 18 October 2011
taxis operating in major cities increased by almost 200% following deregulation in 198960. According
to Morrison (1997):

         The 1989 legislation, which removed the quantitative controls (deregulation), has been
         followed by a tripling of the number of companies in the metropolitan centres and a massive
         increase in the number of taxi cabs. A much wider range of taxi services now exploit different
         market segments and offer a wider geographic coverage. These changes have been
         accompanied by a decline in fares in real, if not nominal, terms.

Soon61 cites examples of service innovation in post-deregulation New Zealand including the
introduction of taxi vans and executive cabs, new taxi charge credit systems, more advertising on
cabs and tendering for public bus routes by some cab companies.

The immediate removal of quantitative restrictions on licence numbers was achieved in New
Zealand without payment of compensation to incumbents, although Soon notes that the context
was one of major reform being undertaken throughout the economy over a relatively short period of
time. Johnston62 also notes that pre-reform taxi licence values were substantially lower than those
current in Australian capitals at present, citing a 1989 figure of around NZ$20,000 (or around
NZ$34,000 in current dollar terms).

Open entry conditions have also been maintained to date. In common with the Irish experience,
fleet numbers appear to have broadly been maintained at levels near their post-deregulation highs.
For example, recent reports state that there are 1,237 taxis in Wellington, or slightly more than
200% more than the 400 comprising the pre-deregulation fleet. Notably, calls for the reintroduction
of limits on numbers continue to be heard, over 20 years after deregulation63.


The Swedish taxi industry was deregulated by the Transport Policy Act of 1989. Prior to this time
there was only one taxi company in each city and municipal councils determined the number of taxi
licences that would be issued with reference to their estimates of consumer demand.

The 1989 legislation deregulated both entry and fares. However, operators were required to inform
customers about fares prior to trips and cabs were required to be equipped with meters that
provided receipts. The taxi association also recommended that fares should be displayed on the
exterior of the cab. Previous requirements that cabs be affiliated with radio despatch networks
were removed, while government established its own despatch networks to provide competition to
the incumbents.

   According to NCC (2002, p 5-3), the number of taxis operating increased from 2,567 in 1989 to 6,907 in 1998. See:
National Competition Council (2002). Assessment Of Governments' Progress In Implementing The National Competition
Policy And Related Reforms. NCC, Melbourne (Australia). Available at www.ncc.gov.au
   Soon, J. (1999) Taxi!!: Reinvigorating Competition in the Taxi Market. Issue Analysis, No 7, May 1999, Centre for
Independent Studies.
   Johnston, A. (2000) Taxi Industry Reform: Should There Be Compensation? Agenda, Vol 7, No 2, pp 171-183.

Final draft: 18 October 2011
Quality controls were reinforced, with assessments made of competence, economic capacity and
character. Compulsory pre-entry training requirements were introduced as, in 1994, were
geographical knowledge tests64.

Taxi numbers had experienced little increase in the years immediately prior to deregulation. The
result was a rapid increase in numbers immediately following deregulation, with the size of the taxi
fleet quickly stabilising at the new higher level. Total taxi numbers in 1995/6 were reported to be 25
per cent higher than in 1989, while waiting times became significantly shorter65.

Fares rose in real terms following deregulation, although significant regional variation was observed.
A concern was the ability of monopoly providers to raise prices in some smaller population centres.
Several years after deregulation, increasing innovation in pricing structures was observed. In larger
towns, central booking offices supplemented the standard tariff with fixed prices for some routes,
maximum prices within zones and discounts for certain categories of customers (e.g. women
travelling alone). However, some argued that a simpler tariff structure was needed in order to
enhance consumers' ability to estimate fares before trips as well as comparing different fare offers66.
It is notable that driver authorisations were not initially part of the regulatory system and were only
introduced in 1995. The impetus for this step seems to have evolved largely from concerns
regarding tax evasion, although drivers with criminal records for violent offences were excluded
from the industry.


Taxi fares in Singapore were deregulated in 1998, while entry was opened in 2003. According to
Chow & Leng67, the result has been a high level of innovation and diversification. They note, in
addition, that rigorous regulation of service quality and standards has been retained and that there
is a high level of surveillance of the industry to ensure compliance with these quality standards.
Regulatory requirements include a provision that operators must notify the regulator of changes in
fares a week before they are implemented.

The opening of entry appears to have yielded a substantial supply response: The taxi fleet grew in
size from 16,857 in 1996 to 24,446 taxis in 2007, while there are now eight taxi operating
companies. Despite this, there is recent evidence that the largest taxi operator has retained a
dominant market position and has been able to act as a "price leader" in practice. Thus, concerns

   Kang, CH., (1998). Taxi Deregulation: International Comparison. Institute of Transport Studies, University of Leeds.
Available at: http://www.taxi-library.org/kang0898.htm
   Ljusberg, JE (1998). Deregulation of the Taxi Market: The Swedish Experience. In Improving the Quality of Legislation in
Europe (TMC Asser Institut 1998, The Hague). Pp 145-150.
   Ljusberg (1998), op. cit., p 148.
   Chow, K. & Leng, LP (2008). The Evolution of the Taxi Industry in Singapore. Public Transport International, Vol 57, No 2

Final draft: 18 October 2011
regarding the consumer impact of fare deregulation persist, several years after entry to the market
was opened68.

In most cases, taxi drivers rent cabs for a fixed bailment fee, with these agreements being medium-
term in nature.


The regulation of the Dutch taxi industry was substantially reformed in 2000, although some changes
were subject to staged implementation. According to Bakker (2007)69:

            Considering previous negative experiences in other deregulating countries, the Dutch
            Government chose an implementation strategy consisting of deregulating as well as re-
            regulating elements. The philosophy was, firstly, to safeguard basic service quality and fair
            competition, and secondly, to abolish capacity and fare rules.

The Government initially replaced regionally based fare price regulation with a single national
maximum fare structure. It abandoned a regionally based licensing system, from 2002 replacing it
with a system that allowed any licensed taxi to operate anywhere within the country. Licensing is
based on operators, rather than individual vehicles and, since 2002, there has been no restriction on
the number of vehicles an operator can employ.

It was initially intended to deregulate fares, but this proposal was abandoned following early
experiences of significant fare increases. In the event, the system of nation-wide maximum fares has
been maintained.

Taxi numbers increased substantially in all areas of the country following deregulation. However,
the pattern of entry differed between markets: Where the rank and hail market dominated (i.e. in
major cities) new entrants tended to be small owner drivers. Conversely, where the pre-booked
market dominates, entry levels have largely been the result of expansions in the operations of
incumbent operators.

Taxi availability increased in most areas, notably at peak times. One dynamic operating has been
that operators from rural areas that faced low demand at nights and weekends were, due to the
abolition of the regional licensing system, now able to move to service urban centres at these times.

However, the improved supply of taxis has not brought forth the expected increases in patronage. It
has been speculated that this may be due to consumer price sensitivity and the failure of the
regulator to reduce fares70.

   Nugroho, A. (2011). Is Taxi Industry Should not Remained as Deregulated Market in Singapore?( sic). See:
     Bakker, P (2007). Deregulation of the Taxi Industry: Experiences in the Netherlands. In OECD/ECMT (2007), p 65.
     Bekken (2007), op cit, p 51.

Final draft: 18 October 2011
Appendix 3: Bibliography

Abelson, P (2010). Governance and Economics of the Taxi Industry with Special Reference to
Sydney. University of Sydney. Seminar Paper presented to the Economic Society of New South
Wales, 14 July 2010.

Barrett, S.D., (2003). Regulatory Capture, Property Rights and Taxi Deregulation: A Case Study.
Economic Affairs, December 2003, pp 35-40.

Barry, F. (2009). Politics and Economic Policy-Making in Ireland. Trinity College Dublin. See:

Chow, K. & Leng, LP (2008). The Evolution of the Taxi Industry in Singapore. Public Transport
International, Vol 57, No 2

Deighton-Smith, R. (2000). Taxi Reform in Australia. National Competition Council Staff Discussion
Paper. NCC, Melbourne, October 2000. www.ncc.gov.au

Department of Treasury and Finance (2003). National Competition Policy: Report for the 2003
Assessment on Victoria’s Implementation of National Competition Policy. Available at

See ESC (2004) Review of Hire Car Licence Fees: Final Report. Essential Services Commission
(Government of Victoria), Melbourne (Australia), June 2005.

Essential Services Commission (2005) Taxi Fare Review: Final Report. Essential Services
Commission (Government of Victoria), Melbourne (Australia), June 2005.

Gaunt, C. (1996). Taxicab Deregulation in New Zealand. Journal of Transport Economics and Policy,
Vol 30, No. 1, pp 103-6.

Gaunt, C. & Black, T. (1994). The Unanticipated Effects of the Industry Commission's
Recommendations on the Regulation of the Taxicab Industry. Economic Analysis and Policy, Vol.
24, No. 2, pp 151-162.

Goodbody Economic Consultants (2001). Review of the Taxi and Hackney Market 2001 (Demand
and Supply). Goodbody Economic Consultants, Dublin, January 2001.

Humphrey v. Minister for Environment and Local Government [2000] IEHC 149; [2001] 1 ILRM 241
(13th October, 2000) http://www.bailii.org/ie/cases/IEHC/2000/149.html

Johnston, A. (2000) Taxi Industry Reform: Should There Be Compensation? Agenda, Vol 7, No 2, pp

Kang, CH., (1998). Taxi Deregulation: International Comparison. Institute of Transport Studies,
University of Leeds. Available at: http://www.taxi-library.org/kang0898.htm

Final draft: 18 October 2011
Ljusberg, JE (1998). Deregulation of the Taxi Market: The Swedish Experience. In Improving the
Quality of Legislation in Europe (TMC Asser Institut 1998, The Hague). Pp 145-150.

Nugroho, A. (2011). Is Taxi Industry Should not Remained as Deregulated Market in Singapore?(
sic). See: http://kumpulankaryasiswa.wordpress.com/2011/04/15/is-taxi-industry-should-not-

OECD/ECMT (2007) (De) Regulation of the Taxi Industry. Round Table No 133. Transport Research
Centre. OECD, Paris, 2007.

Schaller, B. (2005), A Regression Model of the Number of Taxicabs in US Cities. Journal of Public
Transportation, Vol 8, No 5, 2005.

Soon, J. (1999) Taxi!!: Reinvigorating Competition in the Taxi Market. Issue Analysis, No 7, May
1999, Centre for Independent Studies.

Taxi Industry Inquiry (2011). Setting the Scene

Trudel, M (1995). The Fundamentals of Taxi Regulation and the Quebec Experience. Paper
presented to the 7th Congress of the European Taxi Confederation, Donostia - San Sebastian, Spain,
February 1995

United States District Court of Minnesota, Civil No. 07-1789 (JMR/FLN).

Other readings

Ho, LS. (1993). An Optimal Regulatory Framework for the Taxicab Industry. Department of
Economics, Chinese University of Hong Kong. http://www.ln.edu.hk/econ/staff/taxi.pdf

Moore, AT., and Balaker, T. (2006) Do Economists Reach a Conclusion on Taxi Deregulation?
Economic Journal Watch, Vol 3, No. 1, January 2006, pp 109-132.

Morrison, PS. (1997), Restructuring Effects of Deregulation: The Case of the New Zealand Taxi
Industry. Environment and Planning A, Vol 29, No.5, pp 913-928.

OECD (2007). Taxi Services: Competition and Regulation. OECD Policy Roundtables. [Book based
on papers presented to the October 2007 roundtable of the OECD Competition Committee on
improving competition in taxi services. http://www.oecd.org/dataoecd/49/27/41472612.pdf

Final draft: 18 October 2011

To top