有关泰勒规则Taylor rule论文的文献综述(英文)
2. Literature Review
There are numerous types of monetary policy rules have been discussed in economic literatures. In 1993, Taylor rule has been firstly put forward by Stanford University professor John.B.Taylor in USA, which aroused wide concern of both scholars and policians. According to Taylor rule, the central bank should make monetary policy to adjust nominal interest rate in short term based on the change of inflation rate and output gap for stabilizing real equilibrium interest rate. For example, when inflation rate exceeds target of inflation rate, the real interest rate would deviate from the real equlibrium interest rate, hence central bank should adjust norminal interest rate to keep them consistent. In recent years, Taylor rule has been widely experimented with in macroeconomic models, and examined or modificated in literatures. Taylor rule is simple and flexibal to guide American economy, and recieves extensive attention in academia. Besides, it has became a benchmark theory of monetary policy in Federal Reserve, Bank of England, European Central Bank and Bank of Canada soon. For these reason, there were extensive literatures about Taylor rule in recent decades, some proposed problems and others did a lot modification and extension, which formed a series of function that analyze the relationship between interest rate amd inflation rate and output gap (alleged Taylor-type rules). While the reaserches on this issue in China mainly focus on whether it is appropriate to adopt the Taylor rule in China’s monetary policy. The most Chinese scholars affirm that the Taylor rule was applicable for China’s monetary policy. However, some of them bring forward questions about its application, for example, Bian (2006) claimes that the Taylor tule was not stable and could not be adopted in Chinese monetary policy. Meanwhile, many scholars attempted to adopt different reaserch methods, such as co-integration analysis, Generalized Method of Moments (GMM) and time-varying parameter model. Specifically, this section will give a brief review some literatures in different respects.
2.1 Different methodologies
Multitudinous literatures have estimated Taylor rule by different research methods and variable selection. For example, Taylor (1993) measured inflation as the percent change in the price deflator for GDP over the previous four quarters in the original specification of Taylor rule. Subsequently, different price indexes have been
experimented. Kozicki (1999) said that if the rule recommendation was different when inflation was measured by the core consumer price index (CPI) or the chain price index for GDP, then the rule may become not that useful. In her research, there were four alternative inflation measures had been considerd into the estimation: CPI inflation, core CPI inflation, GPD price inflation and expected inflation. Kozicki estimated the Taylor rule to obtain rule recommendations, and the results shown that the recommendations were significant different in different inflation measures. Taylor (1993) mesured potential output by fitting a time trend to real output. Subsequently, other researchers chose different methods including regressing real output on segmented linear trends and quadratic trends, HP filters, and more structural approaches. The results stated that different measure methods work out different recommendations. Kozicki (1999) found that interest rate recommendation values obtained by alternative measure methods of potential output range from 0.9 percentage points to 2.4 percentage points. Zhao and Gao (2004) proposed a robust interest rate rule for China’s interest rate liberation by Levin, Wieland and Williams rule ( LWW rule, 2001), which was closer to China’s situation where the exchange rate influences on the long-term inflation target. Using the model of Ball (1999), they built a dynamic quarterly inflation rate model with the data period from 1993 to 2002, which possessed good statistical and econometric properties, and demonstrate forecasting precision. Finally, they said that the LWW rule may be more robust in emerging market economies and closer to China’s situation than the Taylor rule. Bian (2006) employed GMM and co-intergration test to finish the empirical analyses about the application of Taylor rule in China’s monetary policy. The reaserch results shown that both of the methods could prove that Taylor rule was able to discribe the trend of China inter-bank offered rates (CHIBOR) well. They obtained the reaction coeficienct of inflation gap was between 0.4 and 0.5, and the output gap was between 0.2014 and 0.4958. However, the results also indicated that although it could discribe the trend of CHIBOR, Taylor rule was not stable when applicated in China’s monetary policy and not suitable in long term using.
2.2 Interest rate smoothing
The central banks always prefer to change short-term interest rates in sequences of several small movements in the same direction and change its direction infrequently, which is so-called interest rate smoothing. Clardia, Gali and Gertler (1998) pointed
out that interest rate smoothing was able to help to solve two big problems: fluctuatoin in capital market and reduction of public trust to monetary policy that caused by dramatic adjustment in short term. In general, the Taylor rule is commonly modified with the introduction of interest rate smoothing by using a lagged interest rate term. Based on the Taylor rule reaction function, Sack and Wieland (1999) obtained that the reaction coefficient of lagged interest rate was 0.63, which shown that the Federal Reserve did not adjust the interest rate frenquently. Levin et al (1999) proposed that the optimal behavior of central bank could be explained in Taylor rule model with the introduction of interest rate smoothing. Sack and Wieland (1999) argued that interest rate smoothing may be the optimal behaviour when a central bank was aiming to stabilize the inflation and output. King (2000) also insisted to introduce the lagged interest rate into Taylor rule and the results shown that the interest rate changes smoothly. Orphanides (2001) compared the current-time data and lagged data, and concluded that the central bank should try to adjust the interest rate smoothly and avoid the adjustment in large range and opposite direction. However, Rudebusch (2002) and Soderlind et al (2003) proposed the doubts and claimed that the Taylor rule with smoothing interest rate may cost more to predict interest rate.
In China, Xie and Luo (2002) finished the first Chinese paper to examine China’s monetary policy based on the Taylor rule, which introduced interest rate smoothing. In this paper, the historical analysis and the reaction function were used to conduct empirical analyses. Comparing the recommendation value from Taylor rule in China’s monetary policy with its real value, they concluded that the Taylor rule could discribe China’s monetary policy well, and the difference between actural value and rule value was caused by the lag of monetary policy. On the other hand, they estimated the reaction function of China’s monetary policy, and the results shown that the adjustment coefficient of the interest rates to the inflation rates was lower than 1 and the elasticity of the interest rate to the output gap was 2.84. Therefore, they concluded that China’s monetary policy had an overeaction to the output but a underreaction to the inflation rate. Also, they found the smoothness was 0.82 according to the estimation. Finally, they put forward suggestions that China’s monetary authority should employ the Taylor rule as the benchmark for measuring the stance of China’s monetary policy, because it could help enhance the transparency of China’s monetary policy, implement the interest liberalization reform and transform unstable monetary
policy rule to the stable monetary policy rule.
2.3 Forward-looking rules
Firstly, for the timming problem, the debate about data time mainly focus on whether to use current data or lagged data. Taylor (1999) employed current data to estimate that whether the Federal Reserve had set the Federal funds rate as recommended value from Taylor rule. While Levin et al. (1999) found that the empirical results did not show a substantial difference in the performance using lagged data instead of current data. Hamalainen (2004) explained that the costs were small because both inflation and output were persistent enough, thus the lags of inflation rate and output gap were good proxies for current values. Kozicki (1999) considered the lagged data of output gap and inflation data as a common approach to deal with lags of data. In his paper, he assumed that the Federal funds rate in a given quarter was set depending on the data in previous quarter. However, Orphanides (2001) demonstrated that the rule recommendations obtained with real-time data was significantly different from those obtained with lagged data in a Taylor-rule model. Furthermore, he suggusted that it was essential for monetary authority to make decisions according to the real-time available information.
Forward-looking rule specifies that central bank should focus on expected inflation gap and output gap when make monetary policy, rather than those of current and delay period. According to Batini (1999), there are three real benefits of forward-looking rules. Firstly, the monetary policy rule with forward-looking rule is able to embody explicitly the lags in monetary transmission. Secondly, forward-looking rule is far from output invariant. Thirdly, a forward-looking inflation rule embodies all relevant information for inflation predicting. When the nominal interest rate aggressively responds to ex inflation rates, Carles (2000) concluded that the monetary authority should follow a backward-looking rule for ensuring determinacy. Clarida, Gali and Getler (2000) were first to estimate the forward-looking Taylor rule for the postwar United States economy. Thire results supported the views that the anti-inflationary stance of the Fed has been stronger in the past twenty years. According to the forward-looking monetary policy model, they found that the target interest rate influences inflation rate and output gap based on the relationship between β and 1. The interest rate is stable when β>1, otherwise,
interest rate should be adjusted depends on economic fluctuation. Orphanides (2001) found that the Taylor rule with introduction of forward-looking behavior gave a better discription of the stable relationship between the federal funds rate and inflation gap, output gap. Considering the smoothing interest rate and forward-looking behavior, he used LOS and IV to estimate the Taylor rule in United States during 1987-1992. Based on the analyses of real time data and historical data, he drawn two conlcusions: firstly, the result from historical data was more accurate than the discription by the rules when use the real time data; Secondly, the Taylor rule with forward-looking behavior was more accurate when use the real time data. Huang et al (2001) also used OLS and IV to observe the data in New Zealand during 1989-1998, and concluded that forward-looking Taylor rule did better than the situation that only consider about the model of real time data, although there were only small difference between them. In China, based on forward-looking Taylor rule, Zhang and Zhang (2007) classified interest rate into three levels, which were market interest rate, regulated interest rate and spread between them, to estimate the monetary policy reaction function in China. They found that although the forward-looking Taylor rule could discribe the trend of three levels of interest rate, the three levels of interest rate were under-reaction to expected rate of inflation and expected output gap, which indicated the unstable monetary policy in China. With the co-intergration analysis, Lu and Zhong (2003) estimated China’s monetary policy based on Ta ylor rule. Results from esimation shows that Taylor rule could describe the trend of interbank offered rate well and it was able to be a benchmark in monetary policy making. Concerning about the time delay, they introduced a forward-looking rule into model to modificate Taylor rule. Chen, Yang and Tu (2006) introduced exchange rate factor into a forward-looking Taylor rule based on Lawrence model that under the open economy in China. The results shown that Taylor rule could be a benchmark for Chinese monetary policy. Comparing with the target of price stability, they said that monetary policy in China pays more attention to economic growth. Ye (2008) employed GMM method to estimate the reaction function of Taylor rule with the introduction of forward-looking rule again. He found that China’s inter-bank offered rates comformed to the essential features of the forward-looking Taylor rule and the differences between coefficients of inflation gap and output gap could not be ignored.
2.4 Taylor rule within open economy
The most scholars believed that exchange rate is a significant channel of monetary transmission under open economy, which can influence a country’s prices and interest rate by inport and export. There are economics discuss about some detail problems, for example, what the function of exchange rate in monetary policy is, whether exchange rate can be bring into monetary policy and how to do it. Obstfeld and Rogeff (1996), Svensson (2000) and Ball (1999) are representatives of economics who support the exchange rate should be bring into monetary policy. Obstfeld and Rogff (1996) built a Stackelberg compitition model with leader and follower countries under open economy to prove that it was hard for a follower country to keep exchange rate stable when it has to set interest rate following the leader country. Svensson (1999, 2000) claimed that it was beneficial to bring the exchange rate into monetary policy in inflation targeting regime. For example, exchange rate provides more transmission channels for monetary policy. As an asset price, the exchange rate is forward-looking and predictable, but some external disturbance may be transmitted to domestic market. If exchange rate was be concerned in monetary policy, the adjustment of exchange rate could help to avoid these external disturbance. Therefore, he proposed that exchange rate was ignored in the Taylor rule, and central bank should consider about it in inflation rate target under open economy. On the contrary, Laxton and Pesenti (2003) opposed to consider exchange rate in monetary policy. They built a global economitric model (GEM) to analyze open economy countries, and found that exchange rate plays a weak role in the model. Taylor (1999) also indicated that there was not great improvement in economy condition when bring exchange rate in monetary policy, and sometimes to do so would even makes it worse, according to the estimation of part of European countries. He provided two reason to explain this result: firstly, the effect of exchange rate has been transmitted to interest rate by other relative variables, such as inflation and output; Secondly, there is no need to change interest rate to offset the purchasing power parity deviations caused by exchange rate. However, Taylor (2000) found that exchange rate had great influence on monetary policy in emerging market countries, since these countries were underdeveloped and their foreign exchange markets were unsound. Therefore, he claimed that whether bring the exchange rate into monetary policy rule should depend on the specific economy condition.
In China, Wang and Zou (2006) estimated the application of Taylor rule in China’s monetary policy under the open economy, where exchange rate was influenced by United States, Japan and European countries. They tried to prove whether the Taylor rule was suitable for China’s monetary policy. According to their results, they concluded that the standard Taylor rule had strong steady, and the interest rate level in China was related to inflation rate, output gap and other factors, then they also proved that Chinese optimal level of interest rates was influenced by the development of foreign economies. Meanwhile, they indicated that foreign economy cannot be ignored in monetary policy making. Besides, Deng and Shi (2011) built a mixed Taylor rule including exchange rate to estimate China’s monetary policy, their results proved the importance of exchange rate as well. Wang and Wang (2011) varianted the standard Taylor rule to estimate the application of Taylor rule in China’s monetary policy under open economy, which used the data between the first quarter of 1995 and the fourth quarter of 2010 from the aspects of smoothing interest rates, forward-looking variables, exchange rates and assets prices. The test results of this model shown that there was a positive correlation of interest rate with inflation rate, output gap and asset prices, and the smoothing interest rates was effective in policy behavior explaination.
2.5 Taylor rule with asset prices
With the development of capital market and financial market, a highly integrated global market is formed. Therefore, more economics began to concern another question that whether should bring asset prices into monetary policy making. Bullard and Schaling (2002) sugested that Taylor rule should include asset prices. While Mishikin (2007) propsed that monetary policy need to concern about asset prices only when it influences output and inflation. Although this issue caused a heated debate, there is not a theory conclusion yet. On the other hand, the most Chinese scholars regard asset prices as an important variable in Taylor rule. Qian (1998) said that the fluctuation of asset prices could influence monetary policy since it reflected people’s long-term expected interest rate level. Yi and Wang (2002) concluded in their paper that monetary authority should pay attention to asset prices and stock prices during monetary policy making, because the target of Chinese monetary policy was domestic price stability. Peng and Liu (2004) analysed the Taylor rule focus on the relationship between asset price bubble and monetay policy. The reaserch selected China’s
financial data during the first quarter of 1994 to the fourth quarter of 2001, and considerd about the capital market factor based on Taylor rule estimation. They tried to find out whether China’s monetary policy can react to stock market bubble well. Their conclusion was that China’s central bank did not control the bubble by monetary policy successfully, and the the monetary policy performanced inconsistent and unstable at this stage. According to analyses of data during 1994-2006, Zhao and Gao (2009) concluded that asset prices was a significant endogenous variable in China’s monetary policy reaction function.