Empirical Tests of the Real Business Cycle Model for the Indian Economy

In this paper, we have attempted to test some major implications of the real business cycles model of Long and Plosser [1931] using aggregate level annual data on the Indian economy for the period 1951-1985. Real business cycles imply that nominal magnitudes and real money balances cannot be exogenous driving mechanisms of the business cycle. These propositions have been tested in this paper. Section 1 discusses the low frequency properties of non-detrended data. In particular. those series in whose case the hypothesis of one or more unit ischolar_mains could not be rejected are identified on the basis of Phillips-Perron tests. Along with this, a system having secondary sector output and real balances and another having secondary sector output and wholesale prices is estimated via Johansen's 119881 cointegration, and tests of bidirectional causality following Miller and Russek [1990] have been taken. The short run business cycle is identified with data detrended using the Hodrick-Prescott filter with a weight of 400. Correlations of detrended gdp with various other detrended series are described. Section 2 takes tests of Granger causality and weak exogeneity for nominal and real money magnitudes and gdp over the cycle. Granger causality is confirmed from money to output. Though the results are somewhat ambiguous, they do not give clear cut support to the real business cycle model in the short run. In the long run, however, there is substantial support for the neutrality of money and prices.