Abstract: Abstract The long-run risks literature highlights the importance of the predictable long-run component in consumption growth to explaining the fluctuations in asset prices, but might overlook other potential driving forces that are not consumption-related. So as to understand the asset market phenomena from a wider perspective, this study proposes a consumption-based asset pricing model with recursive preferences, accommodating both long-run and valuation risks embedded in the low-frequency movements in consumption growth, dividend growth, and time-preference shocks. In the modeled economy, asset market fluctuates in response to long-run consumption growth, time-preference shocks and their respective volatilities; and the expected equity premium reflects the market compensation for households’ exposure to consumption growth uncertainty and valuation risks. Empirical evidence, based on the moment-matching methods and the particle smoothing algorithm, indicates that our model can: (1) resolve the equity premium puzzle, the risk-free rate puzzle and the weak correlation puzzle; and (2) account for the joint dynamics of the price-dividend ratio, the market returns and the risk-free rate. Moreover, we find that long-run consumption growth is the major contributor to asset market fluctuations, while time-preference shocks and valuation risks are significant determinants of the risk-free rate.
"Firm Size, Equity Premium and the Economic Policy Uncertainty" (Working in Progress)
Abstract: Recent macroeconomic studies argue that increased economic policy uncertainty reduces bank lending growth. If higher economic policy uncertainty triggers tighter credit market conditions, firms’ excess equity returns would increase to compensate shareholders for bearing larger risks associated with firms’ difficulties in raising funds through external finance. In addition, imperfect capital market theories predict that the effects of tightening credit market conditions on small and large firms are asymmetric. So as to assess these asset pricing implications, we employ the index of Economic Policy Uncertainty (EPU) and evaluate its effects on the expected equity returns on ten investment portfolios constructed based upon firm size. In this study, we find that the EPU index is an important predictor of the expected excess returns. First, regardless of the size of the firms, excess equity returns rise in response to higher economic policy uncertainty. Second, the excess returns on small firms are substantially more sensitive to changes in the EPU index than those on large firms in the recession state. In addition, the effects of the conditional volatility of the EPU index on excess returns are quantitatively similar across small and large firms. However, excess returns on small firms are more strongly affected by changes in the long-run component of the economic policy uncertainty than those on large firms.
"Structure of Exogenous Processes, Aggregate Fluctuations and Asset Price Movements: Bayesian Estimation of A DSGE Model" (Draft Coming Soon)
Abstract: This study proposes a DSGE model with multiple real rigidities, and identifies the major contributors to fluctuations in aggregate variables and asset prices. In particular, we take into account the possibility that the growth of total factor productivity, labor augmenting technology and investment-specific technology might consist of a permanent component and a transitory component. Experiments on various model specifications indicate that the identification of the major sources of aggregate fluctuations and asset price movements would depend heavily on researchers’ assumptions on the structure of the exogenous processes. Bayesian model comparison suggests that previous studies might have overlooked the persistent component embedded in the process of the investment-specific technology growth. In this study, we find that investment-specific technology shocks account for a remarkably large fraction of the short-run and the long-run fluctuations in output growth, investment growth, and total market values. Second, the long-run predicted error variance of consumption-output ratio and hours are overwhelmingly due to the permanent and the transitory components of the investment-specific technology growth. In addition, labor augmenting technology shocks, time-preference shocks and government spending shocks are only important contributors to variation in hours, consumption and government expenditures, respectively, in the very short-run.
"Teasing Out the Macroelasticities using the U.S. Budget Sequestration". joint with Carlos Zarazaga. (Draft Coming Soon)
Abstract: (Coming Soon)
"Optimizing Search Engine Marketing for StatsBank", joint with Zizhen Chen, Savitha Nathan, Peng Ren and Yixiang Zhang, Report of the Award Winning Solution in the Capital One® Modeling Competition 2014.
"Evaluation of Prevention Programs for Reducing Healthcare Costs in the State of New Hampshire", joint with Xusheng Chen, Shuling Liu, Ying Meng, Yibin Xu and Yixiang Zhang) Report of the AwardWinning Solution in the SAS® Data Mining Shootout Competition 2013.