The changing sensitivity of realized portfolio betas to U.S. output growth: An analysis based on real-time data

Our results shed light on the sensitivity of the betas of portfolios formed on market capitalization (“size”) and book-to-market value (“value”) to output growth in the United States. We estimate a state-space model to analyze the sensitivity of portfolio betas to output growth. We measure output growth using real-time and revised data. Output growth has a significant effect on portfolio betas when size and value are high. Such portfolio betas exhibit countercyclical dynamics. They are more sensitive, in absolute terms, to output growth when the latter is measured using real-time data. Their sensitivity to output growth has grown over time. Portfolio betas with respect to output growth have become smaller over time, in contrast, when size is large but value is low.     

The business cycle and the equity risk premium in real time

Building on the stochastic discount factor model, we estimated a multivariate exponential GARCH-in-mean model to analyze the link between the business cycle and the equity risk premium in the United States. In order to measure the business cycle, we used revised and real-time monthly data on industrial production for the period from 1965 to 2008. The main result of our empirical analysis is that estimates of the equity risk premium based on real-time macroeconomic data may significantly differ from estimates of the equity risk premium based on revised macroeconomic data.     

Changes in the international comovement of stock returns and asymmetric macroeconomic shocks

We study whether asymmetric macroeconomic shocks help to explain changes in the international comovement of monthly stock returns in major industrialized countries over the period 1975–2004. Based on a time-varying parameter model, we trace out how the pattern of international comovement of stock returns changed over time. In order to identify asymmetric macroeconomic shocks, we estimate vector-autoregressive models. The results of estimating time-series regression models and panel-data models indicate that changes in the international comovement of stock returns are not systematically linked to macroeconomic shocks.     

Sources of time-varying exchange rate exposure

We report evidence of a time-varying link between returns on national stock market indexes and exchange rate returns (exchange rate exposure). We use this evidence to analyze the sources of changes over time in exchange rate exposure. Using monthly data for 14 industrialized countries for the period 1975–2006, we report evidence of a cointegration relation between exchange rate exposure and the industry composition of a country’s imports, and weaker evidence of a cointegration relation between exchange rate exposure and openness to trade.