Hiroyuki Oi, Shigenori Shiratsuka, and Toyoichiro Shirota
This paper comprehensively investigates long-run monetary neutrality in Japan, with due consideration to the order of integration of the money stock and real output, mainly using long-term time-series data retroactively available from the Meiji Period (1868–1912). The empirical results indicate little evidence against the long-run neutrality of money (especially defined as M2) with respect to real GNP. In addition, such findings are robust to a wide range of identifying assumptions. Keywords: Long-run monetary neutrality; Long-term time-series data; Structural changes; Unit root tests; Bivariate structural vector autoregression (VAR) JEL Classification: C22, C32, E40, E51
Hiroyuki Oi: Institute for Monetary and Economic Studies, Bank of Japan (E-mail: hiroyuki.ooi @boj.or.jp) Shigenori Shiratsuka: Institute for Monetary and Economic Studies, Bank of Japan (E-mail: firstname.lastname@example.org) Toyoichiro Shirota: Institute for Monetary and Economic Studies, Bank of Japan (currently Ohio State University) (E-mail: email@example.com) The authors would like to thank Professors Yuzo Honda, Yukinobu Kitamura, Ryuzo Miyao, Masao Ogaki, and Etsuro Shioji, and the staff of the Monetary Affairs Department, Research and Statistics Department, and the Institute for Monetary and Economic Studies (IMES) of the Bank of Japan for their valuable comments. Regardless, the opinions expressed in this paper are those of the authors, and do not represent the official views of the Bank of Japan or IMES. MONETARY AND ECONOMIC STUDIES /OCTOBER 2004 DO NOT REPRINT OR REPRODUCE WITHOUT PERMISSION.
This paper comprehensively investigates long-run monetary neutrality in Japan, with due consideration to the order of integration of the money stock and real output, mainly using long-term time-series data retroactively available from the Meiji Period (1868–1912). The issue of whether or not changes in the money stock influence real variables, such as real output and the unemployment rate, has long been a major topic in monetary economics. If changes in the money stock and changes in real variables are independent, then money is deemed to be neutral, and if not then money is non-neutral. Specifically, long-run monetary neutrality is said to exist if a permanent increase or decrease in the nominal money stock does not have any long-term effect on the level of output. The independence between real and nominal variables in the long run is a widely accepted notion in economic theory, the so-called classical dichotomy. In theoretical analyses of long-term consequences in the economy, including economic growth, long-run monetary neutrality is generally assumed.1 In empirical analyses of business cycles, long-run monetary neutrality is often employed as an identifying restriction in the structural vector autoregression (VAR) model, proposed by Blanchard and Quah (1989). Thus, it is deemed important to reexamine the empirical validity of long-run monetary neutrality, based on actually observed data. Empirical examinations to confirm long-run monetary neutrality have been conducted for many years. During the 1960s, many adopted the method of regressing output by the money stock.2 In response to these efforts, Lucas (1972, 1973) and Sargent (1971, 1976) pointed out the problems with investigating the long-run neutrality of nominal variables using reduced-form models. Lucas employs a simple rational expectations macroeconomic model to demonstrate that analyses using reduced-form models lead to the erroneous conclusion that long-run monetary neutrality does not exist, even in cases where it does.3 The critique offered by Lucas and Sargent is closely related to the stationarity of data. To verify long-run monetary neutrality, changes in the money stock are required to contain a permanent component that is independent from any changes in...