شفافیت بانک مرکزی در تئوری و عمل

We study the effects of Central Bank transparency on inflation and the output gap. Our intention is to illustrate, with the help of a small analytical model, how an imperfectly transparent Central Bank affects the two main macroeconomic variables, inflation and the output gap. The model tells us that transparency affects the variability of inflation and output but not their average levels. Then we examine the extent to which this conjecture is justified by the index of transparency constructed by Eijffinger and Geraats. Given the limitations of such indices, we only examine the correlations between the index of transparency and the macro variables in question. This analysis shows that the average magnitudes are not affected by transparency but their variability is. In the case of inflation, its variability benefits from the reduction of transparency and about 50% is explained by the variability in the transparency index. The effect on output volatility on the other hand is less clear, and in any case transparency seems to increase it rather than decrease it.

Most economists agree that greater transparency in monetary policy is desirable because it allows the private sector to make better – that is, welfare improving – decisions, as well as better informed decisions (Blinder, 1998). But not all agree. Some argue that incomplete transparency is optimal, as the effect on the Central Bank’s reputation and its consequent ability to control inflation has to be balanced against the private sector’s wish to see output, employment and prices stabilised (see for example Faust and Svensson, 2001 and Jensen, 2002). Others argue that certain restrictions on transparency are important for operational reasons. Once again the idea is to reinforce the Bank’s credibility (see Eijffinger and Hoeberichts, 2002) and to separate ‘the need to know’ from ‘the need to understand’ (Issing, 1999). In practice, many Central Banks have actually increased their transparency in recent years – using inflation forecasts, extensive explanations of the reasoning behind their decisions, and sometimes voting records on policy decisions or a discussion of the ‘bias’ in those decisions, to do so. Prominent examples are found in the Federal Reserve System in the US; but also in the Bank of England and the Central Banks of Canada, New Zealand and Sweden. Yet, on the basis of casual empiricism or circumstantial evidence, one cannot tell whether greater transparency has actually made any material difference to the policies chosen – or more exactly, to the impact and effectiveness of those policies. Nor is it possible to determine exactly what kinds of effects one should expect, in terms of average inflation, output performance and the stability of these indicators. Our purpose in this paper is therefore to determine what effects a lack of Central Bank transparency should be expected to have on monetary policies and what impact that would have on the economy. The main problem is that transparency has many different dimensions, and may mean different things to different people (Eijffinger and Geraats, 2002). Kuttner and Posen (2000) list a number of characteristics thought necessary for institutional transparency. These are: (1) a numerical goal for monetary policy, (2) an inflation report, explaining the expected effects of changes in monetary policy, (3) an inflation forecast (plus assumptions) explaining why those changes were necessary and, (4) a post-mortem evaluation of past policies and their achievements. These attributes cover both the information content as well as the way in which that information has been used. That distinction is important, but is seldom made.2 This distinction itself relates directly to the potential for conflict between the ability to control and the need for transparency. As a result, many commentators reach opposite conclusions about the need for transparency. Kuttner and Posen (1999) argue that it enhances the Central Bank’s ability to use discretionary policies, while Faust and Svensson (2002) argue the opposite (although for particular circumstances). In fact, both sets of authors agree that transparency will reduce the noise and the imprecision in the private sector’s decision making. They only differ as to whether greater transparency would increase the ability of the Central Bank and private sector to make consistent decisions, or whether it would reduce the Central Bank’s ability to control the private sector’s natural tendency to avoid monetary discipline. In terms of defining the concept, previous attempts identify the lack of transparency with the Central Bank having private information about the nature of the shocks (Cukierman, 1992 and Cukierman, 2001); or the public being unclear as to what the Central Bank’s objectives (Cukierman and Meltzer, 1986) or even preferences (Nolan and Schaling, 1998 and Muscatelli, 1998) might be. To allow for each of these possibilities, we present imperfect transparency in two forms: imperfect political transparency about priorities and objectives, and imperfect economic transparency about the conditioning information, shocks, target values, etc. In either case, a lack of transparency introduces a disturbance which distorts the private sector’ expectations for inflation. In this paper we examine these effects using a theoretical model of monetary policy, stripped down to its key components, and find that a lack of transparency does not alter the average inflation or output performance of an economy but it does affect the stability of the economy, increasing both inflation variability and output volatility. Then using a recently constructed index of Central Bank transparency due to Eijffinger and Geraats (2002), we test whether imperfect transparency has had the predicted effects in the main OECD countries. We find that the index appears to support the theoretical predictions of the model, except for output stability which appears to decrease with transparency instead of increase.3 The paper is organised as follows: Section 2 defines political and economic transparency and identifies the way they affect the moments of inflation and the output gap. Section 3 then employs the index of transparency by Eijffinger and Geraats (2002) to examine its empirical relevance. We look at two alternative groupings of this index and discuss the sensitivity of our results. Section 4 provides some comments and qualifications of our results and Section 5 concludes.

A simple model of transparency in Central Bank policy making shows that increases in the degree of transparency would not affect the average levels of inflation and output achieved. But it would decrease the degree of volatility in both inflation and output gap levels – and the more so, in output volatility if the Central Bank has conservative preferences. When we test these predictions against a numerical index of transparency available for nine OECD countries in the 1990s, we find that all the predictions are verified but for the reduction of output volatility. The latter appears to rise with increasing transparency. And although the size of the sample may account for this observation, this result remains robust across the different sensitivity checks performed. It is likely therefore, that the transparency index captures some other elements of Central Bank behaviour – for example, differing degrees of independence or uncertainty about the policy goals that the Central Bank ought to pursue. Looking at this point in greater detail would be the subject for future research.