Abstract
This article presents evidence in favour of time-varying Markov regime-Switching (MS) properties in all shares stock returns in the USA. The model specifications include the US Institute for Supply Management's (ISM) manufacturing and Nonmanufacturing Business Activity Index (NMBAI) in the transition equations. We find that the developments in the ISM manufacturing index affect the regime-switching probabilities in both bull and bear stock market periods. The business activity in nonmanufacturing sectors, on the other hand, has a bearing only on bull market periods. We also test for the possibility of a common factor influencing both stock returns and business confidence in the manufacturing sector by estimating a time-varying MS model with the US industrial production in the transition equation. We find that the null hypothesis of a fixed transition probability MS model cannot be rejected when the US industrial production index is included in the transition equation of a time-varying MS model. We conclude that the information content in the ISM manufacturing confidence index, such as expectational shifts, has a separate influence on the stock market regimes over and above that of actual developments in industrial production.
Acknowledgements
Without implicating, the authors would like to thank Jochen Hartwig, Peter N. Ireland, and an anonymous referee for their valuable comments and suggestions on earlier versions of this article that led to many improvements.
Notes
1 The ISM-manufacturing is formerly called the Purchasing Managers’ Index, or the PMI.
2 His sample includes Australia, Austria, Belgium, Denmark, Finland, France, Germany, Ireland, Italy, Japan, the Netherlands, New Zealand, Norway, Spain, Sweden, Switzerland, the United Kingdom and the United States.
3 A two-state regime switching model with regime dependent autoregressive coefficients has not outperformed the two-state regime switching model with regime independent autoregressive coefficients.
4 Initially, we start with four lags of the confidence index and also evaluate them together and separately in this transition equation. Optimal lag length of the confidence index is decided according to the highest log-likelihood value and other model selection criteria.
5 We do not include a constant term in the transition equation as it leads to estimation problems in the Hessian matrix.
6 In order to justify whether low mean regimes are associated with high variance, we separate the sample according to the bear and bull market regimes and employ GARCH model to determine the degree of volatility of stock return series across the regimes. We conclude that GARCH parameter is 0.93 in the bull market regime, on the other hand, it is found to be more than 1 in the bear market regime.
7 In order to determine relative importance of PMI and NMBAI, we evaluate both confidence indices in the transition equations. However, the presence of high correlation between PMI and NMBAI generates insignificant estimates of coefficient.