Document Sample
WORKSHOP ON KEY ECONOMIC INDICATORS Uses of Economic Indicators Powered By Docstoc
					ORGANISATION FOR                                          ECONOMIC AND SOCIAL
ECONOMIC CO-OPERATION                                     COMMISSION FOR ASIA
AND DEVELOPMENT                                           AND THE PACIFIC

Centre for Co-operation with
Non-Member Economies

                               WORKSHOP ON

                   KEY ECONOMIC INDICATORS

                                   22-25 May 2000

                       Uses of Economic Indicators
                       Measuring Economic Trends

                                    Ronny Nilsson

                               Division for Non-Members
                                 Statistics Directorate

1        Introduction

           There are many ways to use and analyse economic indicators – from simple to very
complicated statistical techniques. However, some basic methods are common to all techniques.
They are all concerned with tracking the current phase of the business cycle in the light of the
latest release of an economic indicator.

          Economic indicators – monthly and quarterly economic statistics are an important
element in the total information framework for a market economy. A well-established system of
quantitative and qualitative information is of key importance in allowing a market economy to
work efficiently. The statistical data in such a system measure both economic phenomena and
economic behaviour. Economic agents (households and enterprises) operating in the market as
well as governments need this information for taking decisions on actions in the market and on
economic policies.

          Business tendency surveys collect “qualitative” information from business managers on
their assessment of the current economic situation and on their intentions and expectations for the
immediate future. Such surveys are conducted in most market oriented countries and they have
proved a cost-effective means of generating timely information on short-term economic

          The next section of this paper briefly describes how key economic indicators are used to
track the economy and provides background material on using economic indicators, in particular,
how to interpret changes in them and how to evaluate their accuracy. The third section outlines
the kinds of economic indicators that are normally used for monitoring macro-economic
performance in the medium-term in OECD countries and describes how individual short-term
economic indicators are used for current analysis of economic developments. A final section
discusses issues related to the presentation of short-term indicators, in particular, short-term

2        Analysing Trends in Economic Indicators

           The economy continually operates in recurring phases of expansion and contraction in
activity, which are referred to as business cycles or growth cycles, if activity is measured as
deviation from the trend and not the level of activity. Economic activity is measured in real terms
i.e. in constant prices.

          The analysis of economic indicators is mainly concerned with changes in the indicators
over time i.e. the cyclical fluctuations. For this purpose, the indicators must be transformed to
show change in terms of percent (growth rates) or as index numbers, making relative comparisons
over longer time periods easier. In this respect it is important to understand how alternative
methods of calculating growth rates lead to different perceptions of economic performance. In
addition, indicators measured in current prices must be corrected for price changes. Finally, to
analyse changes from one period to another within a year it is necessary to work with data
adjusted for seasonal and other calendar related effects, and if the data show strong irregular
movements, it may be necessary to smooth the indicator.

          Short-term statistics is in general based on sample information and this in combination
with the need for timely information may hamper the quality of the data. This may mean that the
indicator is based on incomplete data and will be revised at a later time.

          Seasonal adjustment, smoothing (short-term trends) and data revisions are key issues
when analysing economic indicators for short-term fluctuations and are outlined in more detail in
the following sections.

         Seasonal adjustment

          Seasonal adjustment of economic indicators is important because the adjusted series
may be more easily interpreted and used to measure changes between consecutive periods.
Seasonal adjustment identifies and removes be means of analytical techniques the regular pattern
within a year to highlight the underlying trend and short-term movements in the series. A
seasonally adjusted series consist of the trend plus the irregular component, and if the irregular
component is strong, may not represent a smooth easily interpreted series. To further highlight the
trend-cycle, the irregular component may be removed and the underlying trend-cycle estimated
by smoothing the seasonally adjusted series using a moving average or other trend estimation

          It is important to emphasise that seasonal adjustment and trend-cycle estimations
represent an analytical massaging of the original data. As such, the seasonally adjusted data and
the estimated trend-cycle component complement the original data but can never replace the
original. The non-seasonally adjusted data shows the actual changes that have taken place, while
the seasonally adjusted data and the trend-cycle estimate represent an analytical elaboration of the
data showing the underlying developments.

          In many situations it is necessary to pre-process or adjust the data after they have been
collected, but before they are used for monitoring and analysing short-term fluctuations in
economic time series. A number of the most common types of adjustments carried out prior to
seasonal adjustment are outlined in the following

         Calendar adjustment

          Many monthly time series contain variations, which result from the weekly/monthly
cycle in the daily data. Such variations include:

•   different lengths of months;
•   number of Saturdays and Sundays in a month;
•   official holidays and regional official holidays;
•   differences in the importance of certain working days.

           These may obscure important movements in the series and should be adjusted for. This
type of adjustment is often refereed to as trading day adjustment and in particular effects retail
sales series.

          A special case of trading day adjustment is working day adjustment. This type of
adjustment is performed on production data. The adjustment is made by finding the average
number of working days for each period in the series and to adjust the raw data so that it reflects
actual values comparable to a standard period.
          Outlier adjustment

         Variation due to moving holidays is not corrected for in the calendar adjustment and
must be treated separately. This type of variation is not removed by seasonal adjustment, which
only removes fluctuations, which regularly occur every period. The correction is performed by
use of both external and internal evidence, i.e. usually derived from the irregular factors
computed in the seasonal adjustment process.

          Variations due to special events, e.g. the effect of a severe strike, should be adjusted for
in order not to interfere with the estimation of the seasonal factors performed in the next step.

          Seasonal adjustment is easy to do with off the shelf programs such as X-11 and X12-
ARIMA developed by the United States Census Bureau, or X-11-ARIMA developed by Statistics
Canada. These programs include options for adjustment for varying numbers of working days,
holidays, irregular events, treatment of extreme values, tests for stable seasonality.

          OECD Member countries apply seasonal adjustment at different levels. Some countries
adjust all individual series, while others make the adjustments only at aggregated levels. Most
countries obtain seasonally adjusted totals as the sum of the adjusted components, whilst others
also adjust the totals independently.

          The traditional practice in many countries was to run seasonal adjustment at the end or
at the start of each calendar year and to use the projected seasonal factors in turn to seasonally
adjust each new monthly or quarterly observation. This practice was necessary because of
constraints on computer resources. With the advent of powerful PCs this constraint is less valid

          However, many statistical agencies in OECD countries still adopt this policy. The use of
projected seasonal factors does not take into account the latest information in the series, as would
be the case if seasonal adjustment was rerun every month or quarter. This practice is commonly
referred to as concurrent adjustment (current updating) and is used in some OECD countries. One
reason for not redoing seasonal adjustment every period is to avoid minor variations in the series,
which do not add to the analytical usefulness.

          The general policy in OECD countries is to publish both original and seasonally
adjusted data for each series. Smoothing of seasonally adjusted series is only recommended for
presentation in diagrams and for analytical purposes.

           Smoothing (short-term trends)

          Many economic indicators show high irregular variation even after seasonal adjustment
and some sort of smoothing is necessary in order to track cyclical fluctuations and to determine if
the indicator is actually expanding or contracting. This can be done by smoothing the seasonally
adjusted data and or calculation of changes (growth rates) over different time periods.

           The extent to which seasonally adjusted data are effected by irregular variation is
illustrated for monthly industrial production indices in selected OECD countries. In order to
detect the underlying trend-cycle in the data they need to be smoothed by a moving average. The
number of periods to be included in the moving average is determined by the MCD measure ,

 MCD (Months for Cyclical Dominance) is defined as the shortest span of months for which the I/C ratio is less than
unity. I and C are the average month-to-month changes without regard to sign of the irregular and trend cycle
component of the series, respectively. Although I remains approximately constant as the span of months increases, C

both the length of the moving average and the MCD measure are set out below for the different

No smoothing needed (MCD=1)                         United States
2 month moving average (MCD=2)                      Canada, Japan
3 month moving average (MCD=3)                      Mexico, Luxembourg, France, Germany, Ireland,
                                                    United Kingdom
4 month moving average (MCD=4)                      Denmark, Finland, Greece, Italy, Netherlands, Spain
5 month moving average (MCD=5)                      Austria, Belgium, Portugal, Sweden
6 month moving average (MCD=6)                      Norway

          The above information show that only the seasonally adjusted industrial production
index for the United States is smooth enough to give information on the underlying trend-cycle
development. For over half the number of countries the MCD measure is in the range 4-6, which
indicates that the series are very irregular and that the month to month developments in the
seasonally adjusted series are not very useful to track short-term trends in the data.

          For indicators that are not seasonally adjusted, an indirect technique of seasonal
adjustment is to compare trends of several consecutive months with the same months of the
previous year. However, this method cannot identify cyclical turning points on a current basis
because it focuses on year-to year monthly change and does not provide a seasonally adjusted
view of current movements in their own right.

         A common method to estimate short-term trends in economic indicators is to use the
trend-cycle calculated as part of the seasonal adjustment process. A short-term trend is considered
as one, which eliminates cycles of a frequency of less than 12 months. For relatively smooth
series (MCD = 2-3) a 13 terms moving average is normally sufficient and for more irregular
series (MCD = 4-6) a 23 term moving average is needed to estimate the short- term trend. This
method is used to estimate the level of the trend but is not as reliable for estimating changes in the

            Growth rates

           For estimating changes in the trend on a current basis period-to-period growth rates i.e.
month-to-month or quarter-to-quarter percentage changes would be the ones that give the most
up-to-date information. However, these growth rates give only reliable information for series with
no or little irregular variation (MCD=1).

         The irregular variation present in most economic indicators has led to an extensive use
of annual or annualised growth rates for analysing short-term trends. The most frequently used
growth rate related to annual trends is the 12-month (4-quarter) rate of change.

          The 12-month (4-quarter) rate of change is calculated as the percentage variation over
12 months for monthly series, over 4 quarters for quarterly series and over one year for annual
series. This type of growth rate has the same problems as noted above on the indirect technique of

should increase. Therefore, the I/C ratio, itself a measure of smoothness, should decline and eventually become less
than unity. In practice, there are some series for which the I/C ratio at first declines as the span of months increases, and
then starts to increase again without ever having dropped as low 1. Hence, there is a convention that the maximum
value of MCD should be six. For quarterly series there is an analogous measure, Quarters for Cyclical Dominance
(QCD) which has a maximum value conventionally defined as 2.

seasonal adjustment. This growth rate gives no information on current developments because it is
measuring changes over the same period of the previous year.

          The smoothed 12-month rate of change is a better measure because it is calculated by
dividing the figure for a given month m by the 12-month moving average centred on m-12. This
means that it is less influenced by abnormal developments for a single period in comparison to
the 12-month rate of change.

          The smoothed 6-month rate of change at annual rate is an even better measure in that it
both smoothes and compares with a more recent period, which makes it well adapted for
identification of correctly dated turning points in the cycle. This rate is calculated by dividing the
figure for a given month m by the 12-month moving average centred on m-6.5. This smoother
growth rate is annualised. This method is a way of smoothing out some volatility inherent in
percentage changes while attempting to preserve the cycle.

          Accuracy and Revisions

           There are two fairly simple ways to evaluate the accuracy of economic indicators. The
effect of data errors and relative accuracy of an indicator can be estimated by taking into account
the extent of revisions to the preliminary data and, in the case of indicators based on surveys, the
sampling reliability of the surveys. However, such estimates are not often available but when they
are they should be taken into account.

           It is very important that the series are not revised to a significant extent in later periods
if they are to be used for analysing the present economic situation and for forecasting. Business
survey series rarely are revised whilst in many countries preliminary data for conventional
statistics are released very quickly but later revised up to three times. For a few indicators - in
particular indices of production and new orders - about 30-40 per cent of the forecasting errors
are due to revisions of the first published data in some countries.

3         Economic Indicators Used for Analysis and Forecasting

           Table 1 lists the kind of economic statistics that are normally used for the review of
macro-economic performance in market economies, for forecasting economic developments in
the medium term (over the next 18 months) and statistics used for short-term analysis. The second
to third columns of the table indicate uses of short-term economic indicators for current analysis
of economic developments, which is the focus of this section. Economic indicators are used for
different purposes but some of the major uses are:

•   as leading indicators for short-term developments;

•   as proxies for other related variables which become available later;

•   as policy variables (inflation, interest rates etc.) or market variables indicative for business
    and consumer confidence or market expectations.

Table 1 Uses of Economic Indicators for Measuring Economic Trends

                                          Leading          Proxy for other         Policy/market
                                         indicator            variable               variable

National accounts
Final expenditure on GDP                                                        Demand side of GDP

Industrial production                                      Changes in GDP
Capacity utilisation                                       Potential output

Cyclical indicators
Business tendency surveys                Output                Output           Business confidence
Consumer surveys                       Consumption          Consumption         Consumer confidence
Composite leading indicators           GDP/output          Changes in GDP

Deliveries, sales, stocks, orders         Output

Construction; approvals/starts      Construction output    Changes in GFCF

Domestic trade
Wholesale/retail sales                                    Private consumption
Passenger car registration/sales    Private consumption   Private consumption   Consumer confidence

Labour force
Employment/unemployment                                                            Policy variable
Vacancies                              Employment
Labour productivity                                                               Competitiveness

Prices and wages
Consumer/producer prices                                                          Policy variable
Raw material prices/oil prices           Inflation                               Market expectation
Hourly wage rate                                             Labour costs
Unit labour costs                         Output                                  Competitiveness

Money and interest
Money supply                                                                      Policy variable
Interest rates                                                                    Policy variable
Share prices                                                                     Market expectation
Exchange rate                                                                     Competitiveness

Foreign trade
Imports/exports, net trade                Output                                   Policy variable

Balance of payments
Current/capital account                                                            Policy variable
Change in reserves                                                                 Policy variable

         National accounts

          National accounts data are central for the review of economic performance and to
forecast economic developments. The expenditure components of GDP are of main interest to
market economies where the economies are demand driven. Statistics on national accounts are,
however, only available on annual or quarterly basis in most countries and the data are released
with relatively long delays so that other economic statistics have to be used to give up-to-date
information for the forecasting process.


           Industrial output indices are available monthly and they provide indication on changes
in overall economic activity. Movements in the overall index are usually well correlated with
changes in GDP. This is illustrated in Chart 1 and 2, which show the developments in terms of
growth rates measured over 4 quarters of GDP and industrial production for Japan and Australia
over the period 1970-1999 and 1976-1999 respectively. The industrial sector is the most cyclical
sector of the economy and its cyclical turning points coincide normally with corresponding
turning points in GDP. The experiences in OECD countries are that the turning points of GDP
occur usually within the quarter of the corresponding industrial production turns.

          Production indices by economic category (investment goods, consumer goods,
intermediate goods, raw materials etc.) give in addition information on the cyclical timing
relationships (lead/lag) between the different categories which give advanced indication on
changes in total industrial production. This is shown in Chart 3 and 4, which show the
development in terms of growth rates measured over 12 months of industrial production of
intermediate goods and investment goods for the United States. The series on intermediate goods
is leading the series on total production at cyclical turning points while the series on investment
goods is lagging the total industrial production series.

          Output of specific commodities (cement, steel, petroleum etc.) give signals on changes
in key sectors and are indicative for changes in total output. Cement production is for example a
good indicator for construction activity.

           Capacity utilisation is a potential output measure and is used to date cyclical turning
points in economic activity. Statistics on capacity utilisation are in most countries collected
through business tendency surveys and normally measured as an appreciation of current capacity
utilisation in per cent of technically and economically usable or competitive capacity.

     Chart 1
                                                                         J a pa n : G DP a nd Ind us t ria l p ro du c tio n
                                                                          P e r c e nt c h a ng e s o v e r fo ur q ua rt e rs
12                                                                                                                                                                              20

                                              I n d u stri al p ro d u ct ion , rs


 2                                                     G D P , ls


                                                                                                                                                                                -1 0

-4                                                                                                                                                                              -1 5

-6                                                                                                                                                                              -2 0
1 9 70 Q 1 1 9 7 2 Q 1 19 7 4 Q 1 1 9 76 Q 1 1 9 7 8 Q 1 1 9 8 0 Q 1 1 98 2 Q 1 1 9 8 4 Q 1 1 9 8 6 Q 1 1 98 8 Q 1 1 9 9 0 Q 1 1 9 9 2 Q 1 1 9 94 Q 1 1 9 9 6 Q 1 1 9 9 8 Q 1

     Chart 2
                                                                            Aust ra lia: G DP a nd ind us tria l pro du c tio n
                                                                              Pe r c e n t c ha n ge o v er f o ur q ua rte rs
  10                                                                                                                                                                            15

                                  I ndu st rial prod uct io n





                                                                                             G DP


     -6                                                                                                                                                                         -10
      197 6Q 1   1 978 Q 1      19 80 Q1      1 98 2Q 1         1 9 84 Q 1       19 8 6Q 1     198 8 Q 1   1 990 Q 1   19 92Q 1     1 99 4Q 1     19 96 Q 1      1 99 8Q 1

  Chart 3

                              Un ited States: Ind ustrial pro du ctio n total and intermediat go o ds
                                                 C han ges o ver twelve mon th s
20                                                                                                                  15

                                                                       I nterm edi at goods


 0                                                                                                                  0

                                                   Total product ion



-20                                                                                                                 -15
 1971M1 1973M1 1975M 1 1977M1 1979M1 1981M 1 1983M1 1985M1 1987M 1 1989M 1 1991M1 1993M1 1995M 1 1997M1 1999M1

  Chart 4

                                United States: Industrial production total and investmen t good s
                                                  Changes over twelve mon ths
      20                                                                                                             15


                                                                                     Total production




                                                                                              Inves tm ent goods


  -15                                                                                                                -15
   1971M1 1973M 1 1975M 1 1977M1 1979M1 1981M1 1983M 1 1985M 1 1987M1 1989M1 1991M1 1993M1 1995M 1 1997M 1 1999M1

         Cyclical indicators

          Cyclical indicators are of special interest for forecasting in that they give specific
information on changes in direction (turning points) of overall economic activity. Business and
consumer tendency surveys belong to this category of indicators. In addition to their cyclical
characteristics as leading indicators of economic activity, they also give unique information on
intentions and expectations of both entrepreneurs and households on a very timely basis.

           The cyclical characteristics of some key business tendency survey series and their
ability to predict turning points in industrial production (proxy for overall economic activity) in
selected OECD countries are set out in Table 2. These survey series are normally good leading
indicators for industrial production in most OECD countries and perform particularly well in the
countries set out in Table 2. They are smooth (low MCD) and show stable median leads in the
range of 4-6 months and of 4-8 months measured with the cross-correlation lead and with
correlation coefficients in the range 0.70-0.75.

Table 2 Cyclical characteristics of key business tendency survey series in selected OECD

                             Average of indicators, 1960-1985
        Median lead(-) or lag(+) and correlation against total industrial production
     Indicators       Number of     MCD        Median         Mean       Lead/ Correlation
                      indicators/              lead/lag     deviation      lag    coefficient
                       countries                 at all      around
                                                turning      median
Production                 15         3.3         -5.6          4.1       -6.0       .70
Stocks                     13         2.6         -3.8          3.6       -4.3       -.74
Order books                11         2.5         -4.1          3.1       -8.0       .75
New orders                 11         4.0         -6.5          4.2       -6.6       .71

          Business and consumer tendency survey series could also be used as proxies for related
quantitative variables The specific areas of quantitative (statistical) data with which business
tendency surveys series in industry and consumer surveys could be compared are set out in Tables
3 and 4.

Table 3 Business and consumer tendency surveys and related quantitative statistical series

 Business Tendency Surveys in               Quantitative statistics
 Selected series                            Related quantitative series
 Industry confidence                        Real GDP
 Production assessment                      Industrial production index
 Production expectations                    Industrial production index
 Order book assessment                      Volume of new orders
 Finished goods stocks assessment           Volume of stocks of finished goods
 Selling price expectations                 Producer prices
 Employment expectations                    Employment in industry or manufacturing

    Capacity utilisation                         Industrial production index
              The relationships between a survey series on order books level and a quantitative
   statistical series on volume of new orders measured as changes over 12 months are illustrated in
   Chart 5. The correlation between the two series is high which means that the qualitative order
   books series is a good proxy for the quantitative new orders series. An advantage with the order
   books series is that it is relative smoother compared to the new orders series, which makes it more
   suitable for spotting cyclical turning points.

                                                                German y: Ord er bo oks le vel and new o rders
                                                                   Balance an d chan ge o ver 12 m on ths

 35                                                                                                                                                                         40

 25                                                                                                                                                                         20
                                                                                                   O rd e r b o o ks

 15                                                                                                                                                                         0

  5                                                                                                                                                                         -2 0

 -5                                                                                                                                                                         -4 0
                                                                                                       N e w o rd e rs

-1 5                                                                                                                                                                        -6 0

-2 5                                                                                                                                                                        -8 0
 1 9 7 0 M1 1 9 7 2 M 1 19 7 4 M 1 1 9 76 M 1 1 9 78 M 1 1 9 8 0M 1 1 9 8 2 M1 1 9 8 4 M 1 1 9 8 6 M 1 1 9 8 8 M1 1 99 0 M 1 1 9 92 M 1 1 9 94 M 1 1 9 9 6M 1 1 9 9 8 M 1

   Table 4 Business and consumer tendency surveys and related quantitative statistical series
   Consumer survey                      Quantitative statistics
   Selected series                      Related quantitative series
   Consumer confidence                  Private consumption (total)
   Financial situation of households    Income (disposable income)
   General economic situation           GDP (growth)
   Price development                    Consumer price index
   Unemployment                         Unemployment
   Major purchases                      Private consumption (furniture, el. Appliances, etc )
   Savings                              Household savings
   Car purchases                        Household car purchases
   House purchases/building             Household residential investment
   Home improvement                     Private consumption (central heating, sanitary ,etc)

               The relationships between consumer survey series and related quantitative statistical
   series is illustrated for a number of selected series. The following subject areas are analysed: GDP
   (growth) prices (CPI), unemployment, and private consumption (total). The comparisons are
   restricted to EU countries with available quantitative series comparable to the survey series. The

correlation results between survey variables and corresponding quantitative series are set out in
Table 5.

           The results of the comparisons may be summarised as follows. Consumer survey results
give a first quick indication of the development in some areas of the economy. This is noticed in
the areas of private consumption where the survey results are available long before corresponding
quantitative statistics. The usefulness of survey results is shown to be very important in countries
with no quarterly national accounts. The correlation between the survey results and corresponding
quantitative series is rather good for most variables and very good for prices and unemployment.
These results indicate that consumer survey results could be a very useful tool in assessing and
forecasting economic developments.

Table 5 Correlation between selected consumer tendency survey series and related
         quantitative statistical variables (changes over same period of previous year),
                   Consumer confidence/        Economic situation/    Price developments/
                   private consumption                GDP                     CPI
                   Lead/lag correlation Lead/lag correlation Lead/lag correlation
Europe                 -1          0.90          -2         0.90        -2         0.82
Belgium                 0          0.85           0         0.78         0         0.63
Denmark                 0          0.66           0         0.42        -1         0.87
Germany                 0          0.95          -2         0.92         0         0.90
Greece                  0          0.55          -1         0.71        -1         0.42
Spain                   0          0.89           0         0.76        -1         0.78
France                 -2          0.69           0         0.79        -1         0.72
Ireland                 0          0.42           0         0.74        -1         0.75
Italy                   0          0.90           0         0.72        -2         0.64
Netherlands            -2          0.52           0         0.42        -2         0.93
Portugal                0          0.79           0         0.57        -1         0.58
Finland                -2          0.84           0         0.86         0         0.58
United Kingdom         -2          0.83           0         0.84        -1         0.82

Composite cyclical indicators, in particular, composite leading indicators and confidence
indicators are of key interest for forecasting in that they give more reliable information on
changes in economic activity. These indicators are constructed from a set of individual series with
leading characteristics compared to overall economic activity. Different types of economic
rational explain the capacity of potential cyclical indicators to lead economic activity. The
indicators may be classified into the following categories: (1) early stage indicators, (2) rapidly
responsive indicators, (3) expectation-sensitive indicators and (4) prime movers;

          The first category contains indicators which measure an early stage of production, e.g.
new orders, order books, construction approvals, etc. The second contains indicators, which
respond rapidly to changes in economic activity such as average hours worked, profits and stocks.
The third cover indicators, which measure expectations or are sensitive to expectations and
includes stock prices, raw material prices and expectations based on business survey data
concerning production or the general economic situation. The fourth contains measures relevant
to monetary and fiscal policies and foreign economic developments such as money supply, terms
of trade and indicators for foreign countries.

                    The OECD calculates composite leading indicators for its member countries. The
         indicator for Japan is used to illustrate the construction and performance of such a composite
         indicator. The aggregate leading indicator is based on the following nine component series:
         (1) inventory to shipment ratio, (2) excess of imports over exports, (3) stocks in manufacturing,
         (4) assessment of finished goods stocks, (5) business situation prospects, (6) new loans for
         equipment, (7) ratio loans to deposits, (8) share prices and (9) new vacancies. The cyclical
         development of the composite indicator against industrial production is shown in Chart 6 and the
         historical performance indicates a high correlation of 0.87 at a lead of 6 months against industrial

                                               J ap a n : O E C D C o m o s it e Le a d in g Ind ic a t o r a n d In d u s t ria l p ro du c tio n
                                                                                    R at io to tre n d
 1. 2                                                                                                                                                                             120

1 .1 5                                                                                                                                                                            115
                                                                                             L e a d in g in d ica t o r

 1. 1                                                                                                                                                                             110

1 .0 5                                                                                                                                                                            105

    1                                                                                                                                                                             100

0 .9 5                                                                                                                                                                            95

 0. 9                                                                                                                                                                             90
                                         I n du s tr ia l p r od u c tio n

0 .8 5                                                                                                                                                                            85

 0. 8                                                                                                                                                                             80
  1 9 7 0 M 1 1 9 7 2 M 1 1 9 7 4 M 1 1 9 76 M 1 1 9 78 M 1 19 8 0 M 1 1 9 8 2 M 1 1 9 8 4 M 1 1 9 8 6 M 1 1 9 88 M 1 1 9 90 M 1 19 9 2 M 1 1 9 9 4 M 1 1 9 9 6 M 1 1 9 9 8 M 1

                   Indicators in the group - deliveries, sales, stocks and orders - refer to quantitative
         indicators. This type of indicators is, however, also available in qualitative form through business
         surveys and in this form they belong to the cyclical indicator category. All indicators in this group
         are very closely related to production activity and changes in production. Sales and new orders by
         economic category and stocks by stage of fabrication (finished goods-, raw material-, semi-
         finished stocks) give in addition information on the lead/lag relationships between the different
         categories and the production cycle and give signals on changes in production. Order books are of
         special interest as a leading indicator for production activity in that the time series data are very
         smooth which is very important for the detection of cyclical turning points.

                   Construction approvals or construction starts are leading indicators for construction
         activity and are used as proxies for gross fixed capital formation.

                   In the area of domestic trade, wholesale and or retail sales are used as proxies for
         private consumption and new passenger car registration give information about consumer

                                                                   J ap a n: P riv a te c o ns um pt ion an d re t a il s a le s
                                                                        P e r c e nt c h a ng e s o v e r fo ur q ua rt e rs
12                                                                                                                                                                            20


 8                                                            P riv at e c o n su m p t io n




                                          Re t ai l sa le s


-6                                                                                                                                                                            -15
19 7 0 Q 1 19 7 2 Q 1 1 9 7 4 Q 1 1 9 7 6 Q 1 1 9 7 8 Q 1 1 9 8 0 Q 1 1 9 8 2 Q 1 1 9 8 4Q 1 1 9 8 6Q 1 1 9 88 Q 1 1 99 0 Q 1 1 99 2 Q 1 19 9 4 Q 1 1 9 9 6 Q 1 1 9 9 8 Q 1

                Labour force statistics on employment and unemployment are monitored as key policy
      variables. Vacancies and other labour market series like hours worked are leading indicators of
      changes in activity. Other labour market series referring to marginal employment adjustments
      such as overtime and lay-off rate are normally also selected as leading indicators.

                Price indices like consumer and producer prices indices are monitored as key policy
      variables while raw material prices are monitored in that they give information on market
      expectations concerning future price changes and they are in this sense selected as leading

                 Wage statistic on hourly wage rate is monitored as a narrow, but timely, measure for
      labour costs; many countries calculate such data only for manufacturing industries. Unit labour
      costs give information on competitiveness and are normally calculated as ratios of compensation
      of employees to value added. Profits and costs are driving forces in a market economies and
      statistics on unit labour costs data (in inverted form) give indication on future developments of
      economic activity.

                Statistics on money and interest rates and other financial data are of particular interest to
      market economies in that they contain measures relevant to monetary and fiscal policies and are
      considered as key policy variables. Changes in money supply indicate whether monetary policy is
      accommodating, expansionary or restrictive stance of monetary policy. Interest rates help to
      determine the level of investment and consumption. Interest rates and exchange rates are
      presently the key indicators for monetary conditions and are used as leading indicators (in
      inverted form). Growing instability in the money-income relationship has meant that the
      significance of monetary aggregates for analysis and forecasting has diminished. Share prices
      reflect market expectations and is used as a proxy for business and consumer confidence.

                              Un ited kin gd o m: In terest rate (3 mon th s) an d ind ustrial pro du ctio n
                                                             R atio to tren d

1.8                                                                                                                1. 15

                                                                                                                   1. 1

1.4                                                    I ndust rial produc tion

                                                                                                                   1. 05




                                                                                                                   0. 95

                                    Int eres t rate                                                                0. 9

0.4                                                                                                                0. 85
 1970M1 1972M 1 1974M 1 1976M1 1978M1 1980M1 1982M 1 1984M 1 1986M 1 1988M1 1990M1 1992M1 1994M 1 1996M 1 1998M1

                Foreign trade statistics. Total imports and exports are considered as important policy
      variables in open market economies. Detailed statistics on imports and exports by commodities
      and by partner countries give information on competitiveness. A global indicator of
      competitiveness is terms of trade, which reflect the relation between import and export prices.
      Export performance is in many market-oriented countries the driving force for economic activity
      and a positive export development is very much related to the competitiveness of commodities in
      the exporting country.

                The current account of the balance of payments is the account normally used in short-
      term analysis and is considered as a policy variable. The current account balance includes
      merchandise trade, factor services, non-factor services and transfers. The advantages with balance
      of payments data compared to foreign trade statistics is that trade in services is included and
      transactions in interest dividends are and current transfers are covered in addition to merchandise
      transactions; it therefore gives a complete picture a countries current transactions with the rest of
      the world. Change in reserves is an important indicator for monetary polices in that it shows the
      situation concerning short-term foreign assets. The capital account balance is not very often used
      for short-term analysis and it might be noted that, with the liberalisation of capital transactions in
      market oriented countries and the globalisation of financial markets, capital accounts data is
      increasingly unreliable.

               Among the economic statistics used for short-term analysis, the qualitative business and
      consumer surveys are special in that they give information on intentions and expectations of both
      entrepreneurs and households on a timely basis and are very well adapted to monitor the cyclical
      development in economic activity in market-oriented countries.

                                 N etherlan ds: T erm s of trad e and ind ustrial p ro ductio n
                                                        R atio to tren d

 1.1                                                                                                       1. 1

1.08                                                                                                       1. 08
                                                      T erms of t rade

1.06                                                                                                       1. 06

1.04                                                                                                       1. 04

1.02                                                                                                       1. 02

  1                                                                                                        1

0.98                                                                                                       0. 98

0.96                                                                                                       0. 96

0.94                                                                                                       0. 94

                                                              I ndustri al product ion
0.92                                                                                                       0. 92

 0.9                                                                                                       0. 9
  1970M 1 1972M 1 1974M 1 1976M 1 1978M 1 1980M1 1982M1 1984M1 1986M1 1988M1 1990M1 1992M1 1994M1 1996M1

4         Presentation of economic indicators

              Institutes that produce statistics normally put a great deal of effort into sampling,
collection and processing to produce the data. However, less effort is normally put into
presentation and publication of the results. In particular, little time is devoted to presenting short-
term indicators in a way that highlights short-term trend movements in the data.

          A study conducted by the Office for National Statistics in the United Kingdom in late
1996 showed that among National Statistical Institutes (NSI) in 21 countries about half of the NSI
produced trend estimates for a few key economic indicators. The study covered mainly OECD
countries so the number of NSI producing trend estimates would certainly be smaller on a larger
sample. A bit over half of the NSI that produced trend estimates published them as graphs and
numbers while about a quarter of the NIS published trend estimates as graphs only.

           There is no general accepted method for producing short-term trend estimates.
However, trend estimates produced by NSI are normally obtained from the trend-cycle calculated
as part of the seasonal adjustment process. The above study showed that most NIS used the X-11,
X-11-ARIMA or X-12-ARIMA program for seasonal adjustment.

          Both seasonally adjusted and trend-cycle data are produced on order to help the users in
their analysis. However, which one to consider for publication, especially for graphical
presentation, depends on several considerations. The Statistical Office of the European Union
(Eurostat) issued the results of a task force on seasonal adjustment in January 2000 which
includes the following recommendations to NIS. Both seasonally and trend-cycle data should be
calculated and published. However, concurrent estimation of trends leads to in general to
frequent and important revisions and a through evaluation of these revisions should be studied
before publication of trend estimates. Graphical presentation should preferably be based on trend-
cycle data. In special cases, and in the case of relative smooth data, seasonally adjusted data
should be used for presentation in graphs in addition to trend-cycle data.

          Users of economic indicators

          There are many different potential users of statistics in general and economic indicators
in particular. The following groups of users may be identified:

•   Economists, researchers and analysts in general
•   Senior executives in business generally
•   Politicians
•   Senior civil servants responsible for government policy
•   Senior personal in banks and financial institutions
•   The press and other media

          From the above list of potential users it is quite clear that the different categories do not
require the same type of information. In general, users can be divided into two categories: those
who intend to analyse the results in detail and those who simply want to know the main results.
To the first category i.e. the “analysts” will include economists, researchers and other analysts in
general. The second category may be labelled “executives” and will comprise the largest number
of potential users of the results.

              The analysts are characterised by people who are capable and have time to analyse
the indicators. Such users may work for statistical institutes, economic research institutes and
organisations that have their own in-house research department. The main requirements of this
category in terms of short-term indicators are to:

•   have the original figures i.e. raw data in numbers;
•   have the seasonally adjusted data in numbers and graphical form;
•   have the trend-cycle data in numbers and graphical form;
•   receive historical data for a few years to reduce time looking through many back issues of the
    indicator reports;
•   not have the figures revised too frequently, and to have the full back run of data made
•   have any breaks in the series competently dealt with and noted, with at least one period on
    both old and new base;
•   have an appropriate amount of concise, easily absorbed, methodological information and if
    possible a telephone number or address to write to with queries;
•   receive the data quickly after publication.

           However, the analysts are a minority of potential users, but an important minority. The
executives are a much greater potential (and influential) user group, characterised by those who
do not want to look at the results as such, but want to know what the results mean, that is, they
need their information “predigested”. If the institute producing the indicators does not publish the
results in a way that enables these people to “use” them, the majority will never know what the
results mean.

          The results for senior executives should be presented in a way that encourages them to
read through the material and at the same time makes it easy for them to absorb the material. This
category of people does not have time to read everything and will select only, that, which seems
important. They will hardly look at statistical tables. On the other hand, the interpretation of some
economic indicators is not obvious and a simple explanation needs to be presented. However,
selected statistical tables should be included in the presentation, they add credibility to the
analysis, but they should play a relatively minor role or be put in an annex. The main
requirements for a report for executives may by summarised in the following points:

• the main part of the report should devoted to text;
• the text should focus on results which have some particularly relevant or important to say;
• the main results should presented in a simple way and give a good overview of the economic
• it should be brief, two or three pages of text with a few graphs containing seasonally adjusted
  and trend-cycle data (cyclical indicators are particularly suitable for graphical presentation);
• statistical tables should include seasonally adjusted and or trend-cycle data and be attached in
• the style of presentation of the text should not be bureaucratic, and text should be laid out in
  short sentences and paragraphs;
• the text should be accompanied by information showing where more detailed data could be
  accessed (in other publications) and a contact person (and telephone number) from where
  further information could be obtained;
• it will need to explain why the indicators can be important guides to economic activity and any
  relevant technical points.

              Reports for users

              Three main types of reports or publications may form the basis for a publication
strategy for presentation of short-term trends in economic indicators to different users. The four
main reports are:

• press release
• report for executives
• report for analysts

          Focus of information and content of these reports are summarised in Table 6 and
outlined in detail in the following paragraphs.

          The press release should be very short, one or two pages maximum. Only the main
features should be covered in the text. Seasonally adjusted data should be given more emphasis
than trend estimates in the text. Graphs should show both seasonally adjusted and trend data,
while tables should focus on seasonally adjusted data. The press notice will also include some
basic methodological information i.e. trend estimation method and reliability of estimates. A
telephone contact number should be indicated.

         In many ways everything that applies to the press release also applies to the report for
executives. However, this report could be a bit longer, two or three pages and could include one
or two graphs. The main message should be identified and summarised in headlines followed by
some text focusing on specific results for the current situation and main features of the future
development. A few tables with seasonally adjusted data could complement the text. The report
should include some methodological information, i.e. indicator characteristics and some
information about how to interpret the results.

           The report for analysts should contain a complete presentation of the indicator results.
This will include an overview of the main results and main features by sectors, regions etc. The
text should be complemented with graphs and tables of key results. This type of report should
contain detailed results with full data presentation in tables for all variables by sectors, regions
etc. for a reasonable period of past data. These tables could show original data, seasonally
adjusted data and trend estimates for a longer period of years so that the analysts can easily
compare current results with past trends.

          The methodological description should be extensive and give more technical and
detailed information. The following points should be covered and included in the report at least
once a year:

•   Seasonal adjustment method and program in use i.e. X-11-ARIMA, X-12-REGARIMA;
•   Decision rules for the choice of different options in the program i.e. forecasting model
    (ARIMA-model), additive or multiplicative seasonal behaviour, moving holidays;
•   Special constraints for time and activity aggregation i.e. annual time consistency on the sub-
    annual seasonally adjusted data, direct or indirect seasonal adjustment;
•   Other constraints in operation;
•   Outlier detection and correction method i.e. strikes, natural events;
•   Decision rules between different kind of data transformation;
•   Revision policy i.e. intervals for re-estimation of seasonal parameters and updating of
    seasonal factors;

•   Description of the working/trading day adjustment i.e. proportional adjustment, regression
•   Contact address

Table 6 Reports for different users

Reports/users    Focus of              Content                            Data presentation
                 information                                       Raw       Seasonally    Trend
                                                                   data       adjusted      data

Press release    Main features         Text, graph, table,           No         Tables       Graphs
                                       indicator characteristics                Graphs

Executives       Main features         Text, graphs, tables,         No         Tables       Graphs
                                       indicator characteristics                Graphs
                                       Interpretation of results

Analysts         Overview              Text, graphs, tables        Tables       Tables       Tables
                 Main features by      Detailed data                            Graphs       Graphs
                 components            Indicator characteristics
                                       Interpretation of results