اثر سیاست های پولی در چین: شواهدی از یک حالت FAVAR

We use a broad set of Chinese economic indicators and a dynamic factor model framework to estimate Chinese economic activity and inflation as latent variables. We incorporate these latent variables into a factor-augmented vector autoregression (FAVAR) to estimate the effects of Chinese monetary policy on the Chinese economy. A FAVAR approach is particularly well-suited to this analysis due to concerns about Chinese data quality, a lack of a long history for many series, and the rapid institutional and structural changes that China has undergone. We find that increases in bank reserve requirements reduce economic activity and inflation, consistent with previous studies. In contrast to much of the literature, however, we find that central-bank determined changes in Chinese interest rates also have substantial impacts on economic activity and inflation, while other measures of changes in credit conditions, such as shocks to M2 or lending levels, do not once other policy variables are taken into account. Overall, our results indicate that the monetary policy transmission channels in China have moved closer to those of Western market economies.

China's economy has experienced remarkable structural and institutional change in recent decades. This change may affect the efficacy of counter-cyclical monetary and fiscal policy for Chinese economic activity and inflation. In this paper, we examine whether or not that is the case in the context of a factor-augmented vector autoregression, or FAVAR, along the lines of Bernanke and Boivin (2003) and Bernanke et al. (2005). Most previous studies have found that market-based monetary policies, such as interest rates and reserve requirements, are unimportant in China relative to more direct, blunt credit policies such as “window guidance” for commercial bank lending levels. In contrast to this literature, we find, using recent data, that interest rates and reserve requirements are more important than direct quantity measures of lending (which, on their own, are insignificant). These results suggest that ongoing institutional changes in China may have led the monetary policy transmission mechanism in that country to have become more similar to that in the U.S. and other Western market economies (see, e.g., Bernanke and Blinder, 1992 and Christiano et al., 1999). Studying the monetary transmission mechanism in China raises two interesting challenges that motivate our approach of using a FAVAR model on relatively recent data. First, the well-known skepticism about the quality of Chinese data—which even Vice Premier Li Keqiang famously admitted were unreliable3—makes a FAVAR very appropriate. That is, we take a broad and expansive approach and use a large number of series associated with Chinese economic activity and inflation to estimate the true underlying, latent values of these series. Second, the rapid pace of institutional and structural change in China motivates our focus on the recent period, which includes both the Great Recession and the ensuing recovery. To the extent that China's economy, as well as its policy and banking institutions, have become more market-oriented, it is reasonable to think that the monetary transmission mechanism might have evolved as well. In terms of Chinese data, the quality of the reported output figures has long been under scrutiny (e.g., Holz, 2003, Holz, 2008 and Nakamura et al., 2014). One approach, which we follow in our FAVAR model, is to focus on a variety of measures of economic activity. For example, Vice Premier Li claimed that he looked at several indicators such as electricity production, rail cargo shipments, and loan disbursements to gauge Chinese economic activity. In a recent study, Fernald and Spiegel (2013) validate the information content of a range of indicators of Chinese economic activity relative to an external, independent statistical measure of that activity—namely, exports to China and Hong Kong, as reported by their major trading partners (the United States, European Union, and Japan). This measure should be highly correlated with true economic activity (either through domestic absorption or through re-processing for export), but is not subject to manipulation or bias by Chinese officials. Fernald and Spiegel report that a number of the alternative indicators they examine are more informative than is GDP as measures of economic activity. Moreover, they find that these alternative indicators typically do better in combination—i.e., taking the first principal component of a set of indicators. They find that a small set of indicators are particularly informative (electricity usage, new floor space added, China's reported exports, and raw materials used), but the more general point is the informational content of the economic indicators other than GDP. A FAVAR approach is particularly well suited to examine monetary policy effectiveness when output and inflation are imperfectly observed, latent variables. Under this approach, one considers a large number of economic indicators to estimate the unobserved latent variables that drive the systematic components of the economy. This approach also minimizes ad hoc decisions about which data to include in a VAR and which not. Indeed, even in the U.S. context with relatively reliable data, Bernanke and Boivin (2003) and Bernanke et al. (2005) note that series such as output and inflation are not directly observable—there are a variety of measures of each. They argue that the FAVAR approach leads to better empirical estimates. Applied to China, the dynamic factor model approach distills a diverse set of economic indicators into underlying factors representing Chinese economic activity and inflation. The factor-model logic suggests that such activity and inflation factors are plausibly more accurate measures of these variables than any individual data series, and therefore may better represent the information sets relevant for policymakers or used by economic agents to make decisions. Turning to the monetary transmission mechanism, market-based monetary policy instruments (such as changes in interest rates) were generally considered inadequate to control China's economy in the 1990s. For example, Qin et al. (2005) argue that this inadequacy reflected the slower pace of reforms in the banking and financial sectors relative to the rest of the economy. As a result, studies of this period suggest that interest-rate policies pursued by the Peoples' Bank of China (PBOC) had little, if any, impact on the real side of the Chinese economy (e.g., Geiger, 2006 and Laurens and Maino, 2007). Instead, policymakers seeking to control cyclical fluctuations relied on relatively direct credit policies, such as telling banks when to lend and not to lend. Studies of monetary policies pursued by the PBOC during this period tend to find that the monetary authority pursued a simple money growth rule, as in the case of Burdekin and Siklos' (2008) findings for their 1990–2003 sample period. As the 1990s came to a close, financial liberalization in China appeared to increase the impact of monetary policy—and interest-rate policies in particular—on the real side of the Chinese economy (e.g., Dickinson and Liu, 2007). Zhang (2009) demonstrates in a DSGE model of the modern Chinese economy that an interest rate targeting rule employed by the PBOC would likely be more effective than a money supply targeting rule in stabilizing China's economy. Chen et al. (2011) argue that the effectiveness of non-standard forms of monetary policy, such as “window guidance” for commercial bank lending levels, are likely to diminish as financial markets become less distorted. Similarly, Fukumoto et al. (2010) argue that Chinese window guidance has been successful in the past, but that its success will diminish as the Chinese financial sector develops, in favor of more standard instruments, such as policy interest rates. Still, financial liberalization is incomplete in China, with remaining ceilings on bank deposit rates and floors on lending rates (e.g. Feyzioğlu et al., 2009). Given this institutional environment, it is an open question how standard China's monetary transmission mechanism currently is, where the standard monetary transmission channel would be exemplified by the findings in Bernanke and Blinder (1992) and Christiano et al. (1999).4 To investigate this issue, we incorporate our dynamic factor model estimates of Chinese output and inflation into a standard monetary policy VAR, identified via a recursive ordering as in the studies above. We begin with a simple three-equation FAVAR, comprising an economic activity factor, an inflation factor, and a (single) monetary policy instrument. In contrast to the previous literature, we find that increases in PBOC benchmark interest rates have standard impulse responses—that is, economic activity slows significantly in response to the shock, and inflation falls. Increases in reserve requirements also slow economic activity significantly. In contrast, innovations to M2 or lending have modest—and statistically insignificant—effects on output and inflation.5 These findings are representative of those from larger FAVAR systems as well: Innovations to interest rates and reserve requirements have economically and statistically significant effects, while our large standard error estimates for innovations to lending, M2, and government spending are consistent with no impact. Of course, lending could still be part of the monetary transmission mechanism, as it is in standard market-oriented economies such as the U.S. That is, increases in interest rates and (to a lesser extent) reserve requirements do reduce the pace of growth of lending and monetary aggregates. The lack of an independent effect of innovations in these aggregates to these variables is consistent with typical findings for the U.S., where much of the variation in quantity aggregates reflects idiosyncratic shocks to money demand rather than systematic policy (e.g., Bernanke and Blinder, 1992). Our FAVAR allows us to analyze the role of countercyclical policy during and since the recent global financial crisis. China's growth experience during this period was exceptional: while economic activity in both advanced and emerging economies fell dramatically, China's real GDP growth remained robust, averaging 7.4 percent during the period covering the U.S. recession. (This performance is all the more remarkable given that China's heavy reliance on trade might have been expected to make it exceptionally vulnerable to the global downturn.) Surprisingly, we find that China's strong performance during the global financial crisis was not attributable to countercyclical monetary policy pursued by the Chinese government. That is, in the context of our FAVAR, when we set the interest-rate and reserve-ratio innovations to zero, China's economic performance during the crisis would actually have been better. The reason is timing: China was tightening monetary policy early in 2008. Given the estimated transmission lags, the contractionary effects of that tightening were hitting the economy in late 2008 and into 2009—exacerbating the effects of the global financial crisis on China. The paper most closely related to the present one is He et al. (2013), who also use a FAVAR model to estimate the effects of Chinese monetary policy on the Chinese economy.6 In contrast to the present paper, those authors treat Chinese output and inflation as observed variables (measured by industrial production and the consumer price index, respectively), and use latent factors to capture the effects of other variables on the Chinese economy, such as other Chinese economic indicators and measures of U.S. output and inflation. They find that industrial production responds modestly to a shock to either the benchmark lending rate or “market-based” PBOC policies, but responds strongly to shocks to total lending or M2. In contrast, we apply a dynamic factor model to Chinese output and inflation as well, given the widespread concerns about the quality of official Chinese measures of those variables. We also perform our analysis on a more recent sample then He et al. (2013), and are thus better able to study the performance of the Chinese economy during the recent global financial crisis and recovery. The remainder of our paper is organized into 6 sections. Section II introduces our FAVAR methodology. Section III describes our data. Section IV presents our main results. Section V conducts a number of robustness checks of those results. Section VI concludes. An Appendix provides some additional robustness analysis.

This paper uses a FAVAR methodology to assess the impact of Chinese monetary policy. A FAVAR approach appears to be particularly well suited for the Chinese economy, both due to its relatively short span of quality data available and to the likely fact that Chinese economic conditions and policy rules have changed dramatically over recent years. Our FAVAR approach accommodates data with different starting dates, and therefore allows for a broad set of indicator variables. Moreover, the FAVAR representation allows for the incorporation of many activity and price indicator variables while retaining a parsimonious VAR specification to assess policy impacts. Again, this is particularly promising for China, as the time series available for analysis are relatively short, and the data are likely to be noisy and therefore benefit from our broad set of indicator variables. Our results suggest that, contrary to earlier papers, China's monetary transmission mechanism is beginning to look more standard. In particular, we identify a substantive role for interest rate policies in the determination of both real economic activity and prices. The latter occurs with a lag. While these results contrast with earlier studies of the Chinese economy, they are directly consistent with a number of studies in the literature for the U.S., suggesting that China's idiosyncratic monetary transmission mechanisms may be a function of distortions to the Chinese economy, particularly in the Chinese commercial banking system. These studies predict that, as these distortions diminish, the standard monetary policy instruments are likely to gain importance in the monetary transmission mechanism and consequently, in Chinese monetary policy. An important caveat is that our monetary policy FAVAR analysis does not tell us whether the channels of the monetary transmission mechanism are the same in China as in Western economies. The VAR itself simply tells us about timing. That is, the VAR finds that changes in interest rates that cannot be predicted based on current as well as lagged activity and inflation are associated with later contractions in activity and prices. The mechanisms at work in Western economies are that changes in central bank policy rates lead to changes in other market interest rates as well as broader financial conditions. These changes, in turn, affect the economic decisions of various agents in the economy. Further research is needed to assess the degree to which these channels of monetary transmission are the same in China. Of course, China's economy is far from fully liberalized, and more non-standard Chinese monetary policies—such as window guidance to Chinese commercial banks—are likely to continue to play a role in its monetary policy going forward. Still, our results indicate that the liberalization of China's economy to date, particularly in its financial sector, has left that country's monetary transmission mechanism closer to those of Western economies than previously realized.