Business Cycles and the Stock Market in Israel: An analysis using Markov Switching Models
This paper uses Markov Switching Models (MSM) to assess the differences in timing between turning points in the stock market and turning points in a number of macroeconomic series representing the level of activity in the Israeli economy. These models enable us to estimate the expectations and the variances of series from the stock market and of macroeconomic series in a state of growth on one hand and in a state of slowdown on the other. Moreover, running a uniform Markov Model enables us to map the chronology of the turning points (from growth to slowdown and vice versa) in retrospect for each series, and then to determine the timing differentials between series from the stock market and the other macroeconomic series. The benefit in using stock indices rather than other indices to forecast business cycles derives from the fact that on the assumption of an efficient capital market, all of the information regarding future corporate cash flows is included in the current share prices, the information is available in real time, and it is not subject to revisions. At the same time, deviations from the forecast are caused by expectation errors, particularly from unexpected shocks, and as is their manner, the stock markets tend toward overshooting. The findings show that during the sample period (January 1990 to December 2010), the turning points (from growth to slowdown and vice versa) on the industrial shares index preceded the turning points on the Industrial Production Index by an average of 7 months, with a standard deviation of 3 months. From the standpoint of the forecasting quality, there was one overidentification and one other borderline overidentification. In contrast, there were no underidentifications. The rest of the share indices and macroeconomic time series that were examined show less clear results. The results obtained are significant from the standpoint of economic decision makers and investors, since: (1) they enable us to map the chronology of business cycles by means of a non-linear model; (2) in addition to the forecasts that currently serve the central bank, they provide a further signal of a slowdown in economic activity based on identifying the turning points in the stock market and the timing differentials between them and those on the Industrial Production Index; and (3) after identifying the slowdown or growth in economic activity, we can estimate how long it will last according to the chronology of the stock market.