Asset pricing models are most commonly built to address opuzzlespin either equity or bond markets. A question that seems worth asking is whether this habitual separation does not make one loose sight of much of the commonality in the behavior of asset pricing anomalies across markets. A commonality which by itself would not seem surprising, if one accepts a relation between macroeconomic aggregates and risk premia in asset markets to exist.
In the empirical asset pricing literature, the integration of bond and equity markets has long been an extensively studied topic. Keim and Stambaugh (1986), for example, construct bond and stock market variables which they find capable of jointly predicting risk premia of common stocks, as well as, corporate and government bonds. Cochrane and Piazzesi (2005) find a single return forecasting factor for one to five year maturity bonds, which turns out to be counter cyclical and to also forecast stock returns. Fama and French (1989) find a term premium with a clear business cycle pattern, showing that the returns on stocks and bonds move together. Moreover, they find that dividend yields, as well as, default spreads forecast returns on both stocks and bonds to be high, when business conditions are weak and low, when they are strong.
Elton, Gruber, Agrawal and Mann (2001) regress credit spreads against the Fama French three factor model and find positive loadings on all three factors. The exposure to explanatory factors for stocks can account for up to 85 percent of the default spread which is not explained by either loss on default or taxes. Cohen, Polk, and Vuolteenaho (2003) show that the value spread has a positive and statistically significant coefficient on the default yield spread. Equally investigating credit spreads, Collin Dufresne, Goldstein and Martin (2001) find an adjusted R of 60 percent when using Treasury yield changes and stock returns as explanatory variables.
The empirical asset pricing literature has thus made a strong case for the integration of equity and bond markets, and its significance for asset return predictability. There are common movements across the time series of asset markets that seem to be business cycle related. Moreover, the time series behavior of the cross section of equity and bond markets is correlated.
The present article presents a dynamic general equilibrium model that jointly addresses government bond, corporate bond, and equity markets. In light of the above mentioned empirical findings, comovements in returns across these markets, in the time series as well as the cross section, are driven by changes in long run output growth, the level of interest rates, and default risk along the business cycle.
As in the real business cycle literature, the model is built from specifying the dynamics of real output, which is assumed subject to technology shocks. Long run output growth follows a latent two state Markov switching process, leading to non stationarity in output: it is high, should the economy be in an expansion, and low during contractions. Moreover, GDP growth is assumed subject to i.i.d. shocks that create disparity in output growth at the firm level. This disparity is modelled to display a systematic bias in the cross section, such that fundamentals of (small) value firms react more strongly to downturns of the economy than those of (large) growth firms.
Download
PDF Ebook Macroeconomic Fluctuations, Comovements, and Return Predictability in Asset Markets
