The relevance of reliable statistical systems for monetary policy making in the euro area

Speech by Professor Otmar Issing Member of the Executive Board of the European Central Bank CEPR/ECB Workshop on issues in the measurement of price indices, Frankfurt am Main, 16 November 2001

The construction of the European Union has been made a reality by the combined effort of many different bodies. Eurostat and the National Statistical Institutes of the European Union took part in this combined effort. The most common role played by statistics is that of providing assistance to policy makers. But, in the European Union, statistics have often played a key additional role. Statistics have been instrumental in many political agreements that lead to the construction of the European Union. The Treaty on European Union makes references to statistics in numerous occasions. For example:

  • on rules on the provision of cohesion funds,

  • on the financing of the European Union by setting limits to the tax base of Member States that meet a certain criteria,

  • on the rules of subscription of the capital of the ECB and the allocation of monetary income,

  • on the convergence criteria which guided decisions on the passage to the third stage of economic and monetary union.

On top of the efforts devoted to improving the reliability of statistics, a major challenge in the path towards the creation of a monetary union was the issue of harmonizing statistical series across member states. This is an ongoing project, but major achievements have been made. The issues that will be discussed in this Workshop, and which I have the pleasure to open, are in my view at the core of those challenges that Statistical Institutes currently face.

Over the last years we have witnessed rapid technological developments. As a result, the degree of sophistication of a growing percentage of consumer goods and services is now much larger than has been the case in the past. This raises problems for statisticians because there are recognized difficulties in measuring price changes in high technology goods and services whose characteristics are changing rapidly. Price indices are typically computed by matching prices of identical items over different periods of time. This means that the list of products being compared, cannot be identical to the list of all goods traded in the economy. In periods of rapid technology developments, it is plausible that this sample may not be a good representation of all products. Hedonic techniques have been introduced to deal with the problems of matching among goods due to changes in their characteristics. Hedonic techniques make use of regression methods to link difference in the price of products to differences in their characteristics. But there are also difficulties associated with the use of hedonic techniques. Firstly, these techniques cannot be implemented for pricing an entirely new good. Secondly, the implementation of these techniques requires us to select a collection of its characteristics that are valued by the users. This may not be a simple task. It is also problematic that certain assumptions need to be made with respect to the functional form of the hedonic regression.

The efficient functioning of modern economies depends to an increasing extent on reliable and harmonized statistical systems. Economic policy in general, and a single monetary policy in particular, could not operate properly without them.

To give some examples: Numerous analytical studies have been devoted to compare patterns of changes in productivity in different industrial sectors across countries. These studies provide an indication of the successes or failures of different a) regulatory frameworks, b) industrial organizations, or c) human capital compositions. As regards regulatory frameworks, some countries have played a leading role in the privatization of certain industries and some others have followed. It is certain that the followers looked carefully to the sign of the productivity gap between their industries and those of the leading country. For most of the seventies and part of the eighties the Americans looked at the industrial organization of Japanese companies to try to understand the productivity gap between both countries. Similarly with human capital compositions, there are large disparities among European countries in the levels of 'employment based' training and how much of this training is funded by the public sector. Concerns on productivity differentials that may be thought likely to result from disparities in the composition of human capital, may translate into educational policy measures. The problem is that disparities in levels of productivity are distorted by disparities in the techniques used to measure prices in different countries. This can change significantly the conclusions. Recent work conducted at the OECD has pointed to large disparities between computer equipment deflators among certain countries. These disparities cannot be explained simply by domestic factors but most likely are due to the different methods used to adjust for quality. This poses the risk that the rates of growth of investment in the information and communication industries might reflect these methodological disparities rather than any real developments.

One of the main effects that the mismeasurement of price indices may have on fiscal policy is that it may distort the fulfillment of a planned public budget. Many public expenditure items are linked to price indices. For instance, social transfers. The impact of a certain bias in a price index may be large when measured in the resulting amount of extra cash to add or subtract from total public expenditure. More importantly, the political intentions of a certain government with respect to its social policy would either be unfulfilled, if the price level is underestimated or vice versa. The latter may translate into a lack of funds and force a government to run a deficit to finance the extra expense. Or force a reduction in public expenditure from other items to keep in line with deficit ratio recommendations. Whether the bias leads to an increase in public expenditure or a reduction, it means that the final composition of the budget will differ from that initially intended. This planned budget is almost surely that in line with the political agenda that counted with the support of the majority of the citizens when they agreed to support that government.

The present scenario of rapid changing technology combined with low inflation makes the issue of measurement biases in price indices of the utmost relevance for monetary policy. The ECB took the issue of potential mismeasurement of prices into consideration when its monetary policy strategy was adopted.

The Treaty establishing the European Community in its article 105 stated that the primary objective of the ESCB is to maintain price stability. The high levels of inflation experienced during the seventies served to emphasize the relevance of achieving price stability for sustainable economic growth. There is general agreement now on the absence of a long run trade off between prices and activity. A long run trade off would imply some kind of money illusion or myopic behaviour by economic agents. High levels of inflation are usually associated with lower rates of growth. An environment of price instability is detrimental for an optimal allocation of resources.

Both anticipated and unanticipated inflation bring distortionary effects. The distortionary effects of anticipated inflation result from two sources. Firstly, inflation makes holding money less attractive and as a result economic agents search for alternative stores of value. This searching process will be costly. Secondly, anticipated inflation does not avoid the so called 'menu costs' of changing prices. Unanticipated inflation adds some further distortions to these. Firstly, it will bring wealth distributions from nominal creditors to nominal debtors with consequent real effects on the economy. Secondly, price instability makes the outcome of certain investment opportunities more uncertain. Also, economic agents must divert resources from productive investments to devote them to: firstly, a more detailed planning of the likely returns of an investment project under different price scenarios; and secondly, to insure themselves against the risks of price uncertainty. The interaction between the tax system and inflation in an environment of price instability will result in changes in the effective tax rate of different activities. This will affect the profitability of some activities in favour of others and would induce a flow of funds from those activities to the now more profitable ones. The mandate on the European System of Central Banks (ESCB) of maintaining price stability is thus key to achieve sustained growth and full employment.

On the 13th of October 1998, the Governing Council of the ECB reached an agreement on the main features of the stability-oriented monetary policy strategy to be pursued by the ECB. The strategy consists of three main elements. The first involves a quantitative definition of price stability. The remaining two elements refer to the so-called two pillars of the monetary policy strategy, namely a reference value for the growth of a broad monetary aggregate (first pillar), and a broadly based assessment of the outlook for future price developments (second pillar). By adopting a quantitative definition for price stability, the ESCB makes itself more transparent and also accountable to the public. A quantitative definition of price stability also provides agents with a benchmark for forming their expectations on future price developments. The Governing Council of the ECB adopted the following definition: 'price stability shall be defined as a year-on-year increase in the Harmonised Index of Consumer Prices (HICP) for the euro area of below 2%'. Furthermore, the Governing Council emphasised that price stability is to be maintained over the medium term. There are four main features of this definition of price stability. Two of those features are closely linked to the issue of measurement biases in price indices.

First, adoption of the HICP as the price index to be used. The HICP measures the change in expenditure required by a household to sustain in a certain period the consumption pattern of a base period. The underlying concept behind the HICP is therefore the final monetary consumption expenditure. The index measures expenditure resulting in monetary transactions. Focusing on a consumer price index means focusing on consumer goods and services that are at the very end of the production process. This means that changes in the consumer price index will be associated with changes in prices in the whole economy rather than changes in one particular sector. The adoption of the HICP responds to the fact that national Consumer Price Indexes (CPI) are not sufficiently comparable with one another. On establishing the convergence criteria for monetary union, the Treaty required inflation to be measured by means of the consumer price index on a comparable basis and taking into account differences in national definitions. The HICP meets these requirements. It is harmonised across EU countries with respect to the goods and services covered, and also with respect to methodology, such as the treatment of quality adjustment.

Second, the definition refers to the euro area as a whole. The monetary policy decisions taken by the ECB is orientated towards achieving price stability in the area as a whole and is not focused on specific national situations. Disparities in prices over the long term across countries are likely to be small as a result of the completion of the internal market. Nonetheless, divergences can still occur as a result of disparities in production structures, financial markets and institutional structures. If those disparities are large and it is concluded that action is needed, that action should come from national economic policies. The single monetary policy is in no position to address such cross-country divergences.

Third, setting a reference value of below 2%. Chairman Greenspan has proposed a much quoted definition for price stability; namely '[price stability corresponds with] price levels sufficiently stable so that expectations of change do not become major factors in key economic decisions'. This is not a precise, quantitative definition based on a specified statistical concept. Lack of precision reflects, here, among other factors, Mr. Greenspan's view that measurement problems in price indices are so relevant that a 'specific numerical inflation target would represent an unhelpful and false precision'. The ECB's Governing Council was of the view that the quality of the HICP made it feasible to set a precise definition of price stability as part of its monetary policy strategy. Implicitly this corresponds to a very strong recognition of the quality of the work carried out by European statisticians from National Statistical Offices and Eurostat. Most central banks agree that price stability means low and stable inflation. The ECB's quantitative definition is fully in line with this consensus.

As we know, there are three major potential sources of biases in the measurement of price indices: substitution bias, quality change bias and new product bias. The last two are particularly worrying during periods of rapid technological change. Large errors in measuring prices in an environment of low inflation may, for example, hide a situation of 'de facto' deflation. Whenever deflation is not anticipated it will bring distortions in the allocation of resources as a result of a transfer of wealth from debtors to holders of nominal debt. But deflation also causes specific problems for the implementation of monetary policy as a result of the zero lower bound on nominal interest rates. Although even in such a situation monetary policy still has means left to act. Setting a positive level of inflation as an objective serves as a cushion to prevent deflation. One of the reasons why the ECB did not set a "floor" for its definition is related to measurement problems in general and a positive bias in particular. As well as the size of any bias, its variance is also important. A near constant or low volatile bias is not of major concern for those responsible for monetary policy. It would be simple to adjust price estimates to account for it. A high volatility of the bias is a matter of much more serious concern, and makes achieving price stability more difficult. There is no clear indication on the size of this volatility. There is a need therefore to enhance our knowledge on how best to conduct monetary policy in the presence of a highly volatile measurement bias. Better still, there is a need to build robust price indexes with minimum and low volatile biases.

The fourth and final feature of our definition of price stability is on achieving this target over the medium term. This means that the monetary policy decisions made by the ECB will not respond to short run developments in prices associated with non monetary shocks. This type of shocks include changing seasonal patterns, oil price shocks, exchange rate changes, and changes in indirect taxation. Some economists have suggested calculating a low frequency trend of inflation, usually referred to as core inflation, to assess price developments that ignore transitory effects. Notwithstanding, the computation of a core inflation figure may only serve to confuse the public who understands better the standard definition of inflation. Reference to year on year changes in the definition of price stability reflects the concern with low frequency movements.

The ECB took the issue of potential mismeasurement of prices into consideration when our monetary policy strategy was adopted. This Workshop should serve to provide a forum for economists, statisticians and academics to discuss potential measurement issues in price statistics. We very much welcome new research that helps us to fully understand the implications that high technology goods and the complexity of the characteristics of services bring to the measurement of prices. This research may give an indication of the size of the problem, or better may serve to improve the reliability of price measures.

I would like to thank Vitor Gaspar for taking the initiative of organizing a Workshop around this topic. The responsibility for this workshop on the ECB side lies with as many as three Directorates: DG-Research, DG-Economics and DG-Statistics. This wide involvement on our side shows nothing but how relevant this topic is for us. I am thankful to all those of you from these Directorates that made this Workshop possible.

We are aware that Eurostat and the National Statistical Institutes are devoting large efforts to issues related to quality adjustment. However, the input of the academic community is also very much needed in building the required tools. As I have stressed in the past, a continuous exchange of ideas with other researchers and international organizations is instrumental in ensuring that our policies remain state-of-the-art. I am delighted and thankful to see the names of some very distinguished academics in the list of participants, and among those very special thanks go to those that also took part in the organization of this event. I would like also to thank the Center for Economic Policy Research for their involvement. I wish you all a very fruitful Workshop and look forward to learning from your discussions.

Speaking engagements

Media contacts