دانلود رایگان مقاله زمان بندی بازار در طول چرخه تجاری

Abstract

This paper analyzes the economic value of linking return predictability to the business cycle. Recent studies show that stock returns are predictable in recessions while bond returns are predictable in expansions. I examine whether these findings can be exploited in real-time trading by letting the coefficients of popular return regressions switch across states of the economy. The switching models I propose are easy to implement and provide meaningful economic gains relative to their constant coefficient versions. However, choosing a good recession signal is important as inaccurate business cycle turning points corrupt the switching extensions.

7. Conclusion

I measure the economic value of allowing return predictability to switch across the business cycle. I do this for a mean-variance utility investor choosing between stocks, bonds, and a risk-free asset. First, I let stocks only be predictable in recessions and bonds only be predictable in expansions. Second, I let the regression intercept and slope coefficients change freely across recessions and expansions. Both strategies combine recession dummies with standard return forecasting methods such as bivariate regressions and diffusion indices. For the most accurate recession signals, these dummy based strategies considerably improve risk-adjusted returns when compared to constant coefficient forecasts. However, the gains depend strongly on the choice of recession indicator, and I show how choosing the wrong turning points can have large economic consequences.