آیا بانک های مرکزی به بازار سهام واکنش نشان می دهند ؟ مورد آلمان
کد مقاله | سال انتشار | تعداد صفحات مقاله انگلیسی |
---|---|---|
23150 | 2007 | 15 صفحه PDF |
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Journal of Banking & Finance, Volume 31, Issue 3, March 2007, Pages 719–733
چکیده انگلیسی
In this paper, we ask whether the Bundesbank, prior to the European Central Bank taking responsibility for monetary policy in 1999, reacted systematically to stock price movements. In contrast to the results for the US, our empirical findings show a generally weak relationship between German stock returns and short-term interest rates at the daily and the monthly frequency. The results are extremely robust to alternative model specifications. The evidence is inconsistent with the hypothesis of a systematic reaction of the Bundesbank to German stock prices. However, we do find that, as in the US, the Bundesbank may have reacted to the stock market crash of 1987 by loosening monetary policy.
مقدمه انگلیسی
Econometrics has made great strides in identifying the reactions of monetary authorities to economic developments from data that simultaneously reflect the behavior of agents in the economy. Rigobon, 2003 and Rigobon and Sack, 2003 are the latest to offer a solution to the identification problem. Their approach relies on an identification procedure that exploits the heteroskedasticity of shocks to stock returns to identify regimes when the central bank reacts to stock market developments. Relying on US data covering the 1985–1999 period, Rigobon and Sack conclude that rising stock prices drive short-term interest rates in the same direction, suggestive of a systematic reaction of the Federal Reserve to stock price movements. Bjørnland and Leitemo (2005) rely on the short-run identification assumption of Christiano et al. (1999), and combine this with a restriction that the long-run impact of monetary policy on stock returns is neutral. Their findings are compatible with ones reported in Rigobon and Sack. The impact of asset prices, especially stock prices, on the conduct of monetary policy has been debated for some time. For example, in 1987, the Federal Reserve was given credit for stemming the perceived negative macroeconomic effects of the stock market crash of that year (Blinder and Reis, 2005). Further impetus for this debate comes from the increased visibility and importance over the past several years given to the stock market’s role in the monetary transmission process (Chami et al., 1999 and Mishkin, 2001). Even in the European context, while stock markets are thought to pay a less prominent role than in the US, their importance is rising quickly suggesting that the gap between the two continents is narrowing (Goodhart and Hofmann, 2001 and European Central Bank, 2002). Regardless of the extent of market capitalization in Germany and elsewhere in Europe, all that is needed is for asset prices to deviate far enough from fundamental values to influence expectations of inflation and, hence, central bank credibility. Therefore, stock market performance can have an indirect influence on the stance of monetary policy. Moreover, it is highly likely that events in the US, including ones pertaining to developments in stock markets, would have repercussions on the conduct of Bundesbank policy. It has been popular to interpret monetary policy decisions by framing them around some monetary policy rule, namely a Taylor rule. Such rules have also been found to adequately describe the behavior of the Bundesbank (e.g., Clarida et al. (1998); but see Faust et al. (2001)). An unresolved issue is whether inflation and the output gap are the variables central banks respond to. The output gap is sufficiently imprecisely measured as to raise doubts about whether it plays a significant role in such policy rules. Others have asked whether, in addition to an inflation and an output gap, a term representing asset price movements should also be included (e.g., Bernanke and Gertler, 1999, Bernanke and Gertler, 2001, Cecchetti et al., 2000, Bullard and Schalling, 2002 and Gilchrist and Leahy, 2002). The rising volatility of asset prices, such as stock prices, seems to have been associated with a diminution of volatility in business cycle movements. However, the stock price-interest rate relationship involves endogenous variables. Moreover, conventional methods, such as instrumental variables approaches, are unable to satisfactorily correct the endogeneity problem because of the near non-existence of instruments highly correlated with asset prices but uncorrelated with interest rate movements.2 The aim of the paper then is twofold. First, we estimate a model of the stock price-interest rate link using German data covering a sample when the Bundesbank was solely responsible for the conduct of monetary policy in Germany, namely February 1, 1985 to December 30, 1998. As one of the world’s most successful central bank in terms of controlling inflation, it has long practised monetary policy in a highly pragmatic fashion that has involved monitoring a variety of economic indicators around a policy of monetary targeting (Deutsche Bundesbank, 1999).3 Based on a reading of the Bundesbank’s Monthly Reports, it appears that stock market developments did not attract the regular attention of policy makers in Germany, at least during the period considered.4 We provide a brief discussion suggesting varied theoretical motivations linking stock prices to monetary policy. However, a common thread is the concern that a central bank might express over the volatility of stock returns. Institutional differences between the Bundesbank and its US counterpart suggest that the former might be less inclined to react to stock market developments. Second, we modify Rigobon and Sack’s (2003) testing strategy in a number of ways. In particular, we use different methods to identify volatility regimes. We also consider the possibility that the Bundesbank may have reacted instead to deviations from average or trend returns driven by ‘fundamentals’. In addition, we also consider the case where the reaction is based on conditional, as opposed to unconditional, volatility in stock returns. Whereas Rigobon and Sack rely on a 30-day rolling variance of the residuals from a VAR, and recognize that this is an ad hoc procedure, we instead posit a Markov switching process to identify different volatility regimes. The volatility regimes separately identified via the Markov switching approach serve as the input for the identification of the parameters of interest using the Rigobon–Sack approach. Next, we implement a bootstrapping procedure to calculate the t-statistics of the parameter of interest that recognizes the fat tailed nature of asset prices instead of the normality assumption made by Rigobon and Sack (2003). Finally, our VARs capture the potential spillovers of stock market developments in the US. Hence, even if the German stock market, per se, has a limited impact on German monetary policy it is conceivable that stock market developments in the US, where stocks play an increasingly important role in the transmission process, may indirectly influence Bundesbank actions. The paper proceeds as follows. In Section 2, we discuss the relationship between stock price developments and monetary policy. The identification technique is outlined in Section 3. In Section 4, data and model specifications are presented. Section 5 contains the empirical results, and Section 6 concludes with a summary and suggestions for future research.
نتیجه گیری انگلیسی
The nature of the relationship between asset price movements and monetary policy is currently a hotly debated topic in macroeconomics. Relatively little empirical evidence is available that estimates the relationship between asset price movements and monetary policy measures. This paper provides new empirical findings on the role of stock price movements on interest rates using data from Germany over the 1985–1998 period. The relevant empirical evidence is difficult to obtain because of an identification problem that cannot be adequately solved with conventional methods such as instrumental variables. Rigobon and Sack, 2003 and Rigobon, 2003 propose a new identification technique based on the heteroskedasticity of shocks to stock returns. We implement an improved version of their approach. Our empirical results show that, for daily and monthly time series, interest rates did not respond to stock returns Germany over the 1985–1998 period. The findings are fairly robust with respect to a large number of various model specifications. These results stand in contrast with the US evidence provided by Rigobon and Sack. While it is possible that our estimation approach provides one reason for the discrepancy between the German and US evidence, the theoretical rationale linking central bank reactions to asset prices is not yet sufficiently well developed to provide definite guidance on the question. It is also possible that the institutional structure of the Bundesbank and the German stock market may have led to a different response to stock market movements than at the Federal Reserve. In particular, the 1987 stock market crash dwarfs by a considerable margin all other events that raises the volatility in US stock returns, whereas the German stock market experienced three such shocks over the sample considered. More than 90% of the US sample used by Rigobon and Sack (2003) is concentrated around the 1987 period, only 20% of our sample makes up this regime.