استفاده از اطلاعات غیر اصلی برای تجزیه و تحلیل سیاست های پولی در اقتصادهای در حال گذار
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|26725||2009||12 صفحه PDF||سفارش دهید||محاسبه نشده|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Journal of International Money and Finance, Volume 28, Issue 5, September 2009, Pages 868–879
Empirical macroeconomics is plagued by small sample size and large idiosyncratic variation. This problem is especially severe in the case of the transition economies. We utilize a mixed-estimation method incorporating prior information from OECD country data to estimate the parameters of a reduced-form transition economy model. An exactly identified structural VAR model is constructed to analyze monetary policy in the transition economies. The OECD information increases the precision of the impulse response functions in the transition economies. The method provides a systematic way to analyze monetary policy in the transition economies where data availability is limited.
The nature of the monetary transmission mechanism in market economies is difficult to ascertain.1 It is even more difficult to identify these mechanisms in transition economies.2 During the planned-economy era and the early-transition period, a market-type economy (MTE) monetary transmission mechanism did not exist in the formerly centrally planned, now transition, economies because of the underdevelopment of financial institutions and markets. Nor could such a mechanism be measured, since the data generation and collection processes also did not exist. By the middle of the 1990s, institutions and financial markets had developed sufficiently for policymakers to begin employing traditional MTE monetary policy tools, resulting in consistent and purposeful monetary policy.3 However, data availability still limits policy analysts' ability to do quantitative analysis. When time series are short, economists rely on common theories and experience gleaned from economic history in other countries. Even in the Organization for Economic Co-operation and Development (OECD) countries, macroeconomists use common theories to do analysis across time and national borders. Typically, the structures of the models are similar, but the particular empirical estimates vary. The variation may be due to systematic differences in institutions or policy. It may also reflect the presence of idiosyncratic shocks that can dominate econometric estimates in data sets. Forecasters typically find that in-sample model-selection procedures are of little use in choosing models that will do well in out-of-sample forecasting experiments. Stock and Watson (2003) attribute this result to large idiosyncratic shocks. In a similar vein, Devereux (2003) argues that the structures of transition economies (TEs) are similar to the MTEs, but that the shock processes are different. This implies that the typical MTE macroeconometric model and even MTE data may be employed to improve the precision of estimated TE models. Using a small structural vector autoregression (SVAR) Kim (2002) compares the reaction function and effects of monetary policy across West European countries. He found similarities in the effects of policy across Denmark, France, Germany and Italy. We suggest that using data from OECD countries can improve the confidence we have in empirical policy analysis for some transition economies. In particular we suggest using information from the MTEs in those economies where basic reforms have been enacted, but in which there is little history from which to estimate econometric relationships. By basic reforms we mean only that (1) disciplined monetary, fiscal and regulatory institutions have operated effectively for some time; (2) prices and output are determined by market forces in most sectors; and (3) a standardized system of national income and product accounts have been implemented and corresponding economic data are reported. If the economy is thought to be operating like a market economy, but there is only a short history of macroeconomic time-series data, our approach can reduce the sampling error associated with small data sets. The Czech Republic, Hungary and Poland fit these criteria – the above reforms have taken place, but there is insufficient data to estimate models with high precision. Our approach should prove useful in forecasting macroeconomic aggregates and analyzing potential effects of monetary policies.4 In this paper we evaluate the impact of monetary policy in these three transition economies using a structural vector autoregressive (SVAR) model that has been widely used to analyze policy in developed market economies. The monetary policy shocks are identified in a Wold recursive ordering as in Eichenbaum and Evans (1995) and Christiano et al., 1999 and Christiano et al., 2005. This framework, presented in Section 2, requires more data than is typically available in the TEs. Therefore, we explore the use of OECD country data as extraneous information to improve the precision of our estimates for the three TEs. Section 3 explains the panel SVAR model and the mixed-estimation procedures that are used to analyze the effects of a contractionary monetary policy shock. As a preliminary check on the robustness of our method, we compare the impulse responses to a contractionary monetary policy shock estimated using the 15-country OECD MTE panel with those from a model estimated using a three-country TE panel. The dynamic patterns evident in the two models are quite similar. In Section 4 we examine the impulse response functions and variance decompositions from the three-TE country models estimated individually and with the extraneous information taken from the OECD MTE panel. The individual country model with extraneous information from the MTE model provides more precise estimates regarding the monetary policy shock than the models with only that single country data. Although the results are conditioned by the MTE data and look similar, differences remain. Section 5 offers a summary of the analysis and a discussion of policy implications
نتیجه گیری انگلیسی
This paper illustrates how one might incorporate cross-country information in a time-series analysis of the monetary transmission mechanism. Our method is quite general and could be applied to a variety of identification schemes, economies and empirical models. We use the panel data to estimate the common elements in the mechanisms that propagate shocks. Our main point is to demonstrate that the OCED histories could be used to improve the analysis of TE policies. Our results for the MTE common model are consistent with the conventional wisdom about a monetary policy shock. Our method is more likely to be useful if the emerging market economy is operating much like a typical OECD country. Some evidence in favor of this assumption is that the general patterns in the impulse responses for a common model estimated using a panel of data for only the transition economies look much like the patterns found in the common model estimated using panel data for OECD countries. Our interpretation is that the underlying economic dynamics are quite similar, but difficult to recognize for individual economies because of the small sample size and large idiosyncratic shocks. Augmenting the structural VARs with information from the OECD countries provides more confidence in the empirical model. This is a systematic method for quantitative analysis in countries where the economic structure has changed in important ways, whether through the emergence of markets or a fundamental reform in economic policy, yet data is limited.