Statistical Implications of Inflation Targeting

15 The Experience of the Czech National Bank

Carol Carson, Claudia Dziobek, and Charles Enoch
Published Date:
September 2002
  • ShareShare
Show Summary Details
Oldrich Dedek

When the Czech National Bank (CNB) announced its inflation-targeting regime at the end of 1997, the Czech Republic became the first transition economy to introduce this monetary policy framework, and it joined just a handful of advanced nations then practicing it. It should be noted, though, that this step was forced on the Czech Republic by circumstance; it was not the result of subtle theoretical considerations or careful weighing of the pros and cons of the various options. Following the currency crisis of mid-1997, the Czech Republic had lost the “nominal anchor of the transformation” in the form of a fixed exchange rate (already considerably softened, however, by a 15 percent fluctuation band), and a return to monetary targeting did not inspire any great confidence. In a way, then, the implementation of this new monetary regime helped the country out of a tight spot.

This explains why the introduction of inflation targeting in the Czech Republic was not fully in line with generally recognized requirements. First and foremost, as in all the other transition economies, there was no efficiently functioning financial and banking sector. The country’s knowledge of the transmission mechanism suffered—and continues to suffer—from numerous blank spots. And, in practice, the Czech Republic has not always been able to fulfill the requirements for timeliness and reliability. But these are all qualities that grow out of long maturation and gradual cultivation of the economic environment as a whole. In retrospect, one can say that the benefits to this general cultivation of introducing inflation targeting have been far greater than the harm caused by applying it in conditions of incomplete preparedness.

Although the introduction of inflation targeting involved a major change in regime, it certainly did not represent a fundamental turning point in demands on the scope and quantity of statistical data. Even under the previous monetary policy regimes—based on an exchange rate anchor and money supply management—it was unthinkable that decisions would be made without knowledge of the data needed for assessing inflation risks. A wide range of information—on the real economy and the balance of payments, on public finances and monetary developments, on the situation of financial markets, and on trends in the world economy—had been subject to appraisal and detailed discussion within the CNB Bank Board.

Nevertheless, several new factors emerged, both in relation to previous statistical practice and by comparison with the models applied in other countries. For example, an important factor specific to the Czech Republic was the defining of the inflation target in terms of “net” inflation, which had not previously been monitored in the statistics. In simple terms, net inflation monitors prices using an incomplete consumer basket. Items with regulated prices are excluded, as are the price effects of tax adjustments to items that remain part of the basket. The central bank’s aim was to separate the information on price movements from the uncertain and strongly politically oriented process of gradual elimination of major price distortions. Observing the principle that the institution targeting a particular measure of inflation should not be the one doing the measuring has enhanced the transparency of the new monetary policy regime. The Czech Statistical Office in January 1998 started to report not only the routinely monitored inflation indicators (consumer price index, producer price index, terms of trade, and so forth), but also monthly and annual net inflation data.

Although inflation targeting has not generated a need for any fundamental change in statistical measurement or for the monitoring of new data, it is clear that the new regime has placed much greater demands on the qualitative aspect of statistical measurement, particularly as to the accuracy, consistency, and systematic nature of the data. This is chiefly because inflation targeting, as a multi-criteria approach, is based on the use of all the information available for evaluating inflation risks with all their linkages. Order and transparency had to be introduced into the data monitored. The data had to be systematized into a new structure that would meet the requirements of this multi-criteria approach. A new structure for staff reports on monetary and economic developments was also required. These reports are written each month by the CNB’s Monetary Department, and they form the basis for the Bank board’s decision making on monetary policy measures.

The demand for higher-quality statistical data stems from the fact that, under inflation targeting, the inflation forecast becomes the focus of attention, the principal tool for the central bank’s ambition to suppress the adaptive behavior of economic agents in favor of forward-looking expectations. As a consequence, the predictive model, with its ability to identify turning points and changes in trend, takes on a fundamental significance. And that is unthinkable without a high-quality database containing sufficiently long and comparable time series for the selected indicators.

The credibility of the inflation forecast demands that much greater attention be paid to measuring inflation expectations. For this reason, direct measurement of such expectations among selected market participants has been introduced. This is essentially the only new element introduced into the statistical surveying at the CNB since the introduction of inflation targeting. Opinions on the development of prices are surveyed quarterly among nonfinancial corporations (160 respondents) and households (600 respondents), and monthly in the financial markets segment (6 domestic and 6 foreign firms). These surveys confirm the strongly adaptive nature of the expectations of all agents, including those from whom more sophisticated behavior might have been expected.

In parallel with the direct measurement of inflation expectations, indirect measurement methods are also being cultivated. The endeavor is to apply standard methods for extracting information from the behavior of the financial markets, in particular yield curve analysis. The stock market situation, as part of the more general problem of asset price trends, is also monitored. Despite undoubted analytical progress in this field, linked directly with the growing sophistication of the financial markets, many problems remain. Numerous segments of the financial market are still underdeveloped (options, for example). Price signals are marked by distortions (with the shallow stock market often involved in solving its own problems). And there are also clear cases of economic agents making short-term errors.

Inflation targeting has greatly increased the demands on the central bank’s analytical and predictive capacities. A more systematic understanding of the problems is leading to attempts to quantify known theoretical concepts such as potential output, the natural rate of unemployment, and the breakdown of the deficit into its structural and cyclical components. This is not standard statistical measurement, but rather a superstructure above the classic data system. Measurement of such data on the economy is much more the arena of modelers than of statisticians.

Can the CNB’s experience with inflation targeting be used to derive any implications for the IMF’s surveillance activities? Are there any fundamental discrepancies between the data required by the IMF and the data needed for inflation targeting? To answer these questions it is apposite to point out that the Czech Republic is not a debtor vis-à-vis the IMF, so it does not suffer from a feeling that it is being pestered by creditors constantly requesting information about this, that, or the other. The CNB voluntarily shares the logic of the new financial architecture, on the principle that greater openness and investor awareness are important components of the immune system’s protecting the economic organism against currency crisis contagion.

The CNB considers the IMF’s Special Data Dissemination Standard (SDDS) program—for which the CNB is the national coordinator for Czech institutions—to be a major contribution to the new financial architecture. Its usefulness is perhaps most tangible in the IMF’s initiatives to create international standards and codes. In the CNB’s experience, timeliness and comparability of data are greatly appreciated by foreign investors. The SDDS information requirement has not so far meant any duplication of statistics for the CNB. The data requested are those that the CNB uses routinely for assessing the monetary and overall economic situation and that form the basis for drawing up inflation forecasts. The CNB values the fact that the SDDS is an important driving force for data comparability at the international level. For the Czech Republic, as a candidate for membership in the European Union, it is important, for the purpose of harmonizing statistics, that the approaches selected be as close as possible to those required by European institutions.

Regarding the future of the SDDS project, it makes sense to respect and reinforce the tendency toward timely identification of economies’ external vulnerability. This orientation makes it possible to defend the information requirements for a more detailed foreign debt structure, international reserves composition, and other inputs needed for quantifying a full range of vulnerability indicators. If, however, one proceeds from the proposition that most of the time there is no threat of a currency crisis, the SDDS project may appear to lose motivation for further cultivation. Now, the question is whether to strengthen the SDDS project in the direction of a routine decision making on monetary policy, as shaped, for example, by the procedures of inflation targeting. The Czech experience does not indicate such a need if the implication is to go significantly beyond the current statistical reporting framework. Rather, it shows that the multi-criteria nature of inflation targeting from time to time pushes different aspects into the foreground, and that the domestic statistical service must adjust to these shifting current needs. And it is often difficult to promote these requirements, for the most part nonroutine, to the level of general statistical standards valid at the international level.

    Other Resources Citing This Publication