Three essays on the relationship between asset prices and macroeconomic fundamentals.

摘要

This thesis consists of three essays, which examine the relationship between asset prices and macroeconomic fundamentals.;In the first essay, we assess the ability of the standard consumption-based power utility asset pricing model to identify possible low frequency movements in real dividend growth and equity returns and thus potentially shed light on the question of what the main driver of aggregate stock market fluctuations is. In addition to the restrictions implied by the traditional power utility asset pricing model, we impose a long-run restriction on real consumption growth and real dividend growth in the spirit of Bansal and Yaron (2004). We provide empirical evidence that the standard consumption-based power utility asset pricing model, when taken literally, with its assumptions of constant relative risk aversion and constant conditional variances of the macroeconomic fundamentals, is unable to simultaneously account for the observed fluctuations in the aggregate U.S. stock market, in the real risk-free interest rate, and in macroeconomic fundamentals. As a result, it does not allow the data to give a definitive answer to the question whether stock price movements are primarily driven by changes in expectations of future dividend growth or changes in expectations of future equity returns.;In the second essay, we introduce additional macroeconomic-fundamentals-based factors into the original state-space model from the first essay. We achieve that by using two different channels. The first one involves relaxing the assumption that the variances of the macroeconomic fundamentals are constant through time. Time-varying variances dramatically affect the implications of the consumption-based asset pricing model as they introduce new dynamics for the expected excess returns series, thus improving the ability of the model to capture asset price movements, while keeping it grounded in macroeconomic fundamentals. The second channel allows for habit persistence in the utility function, similar to the one proposed by Campbell and Cochrane (1999). This assumption introduces time-varying risk aversion into the model, a feature which generates time-varying expected excess returns by introducing new state variables which are functions of macroeconomic fundamentals. The former modification of the benchmark model is one in which excess returns vary over time as a result of a time-varying quantity of risk, while the latter modification is one in which excess returns are time-varying because of a time-varying price of risk. Both specifications explain movements in asset prices and macroeconomic fundamentals reasonably well. A comparison of the out-of-sample forecasting performances of the two models indicates that the TVRA model does a slightly better job at explaining movements in the aggregate stock market, while the TVV model is significantly superior at forecasting movements in the real risk-free rate.;In the third essay, I build a dynamic stochastic general equilibrium (DSGE) model, estimate it using Bayesian MCMC methods, and use the results in order to assess how asset prices and macroeconomic aggregates respond to news about changes in the long-run components of macroeconomic fundamentals. My model is set in a real business cycle environment that is augmented with four real rigidities: external habit formation in consumption, internal habit formation in leisure, investment adjustment costs, and variable capacity utilization. Fluctuations in the endogenous variables of my model are driven by the three exogenous processes which govern the evolution of labor-augmenting technological change, investment-specific productivity, and total factor productivity. Each of the exogenous processes is assumed to have two components: a long-run component, which is stationary, but very persistent, and a short-run contemporaneous i.i.d. innovation.;I demonstrate that shocks to the long-run components of the three exogenous processes are the main contributors to fluctuations in both, stock prices and macroeconomic aggregates. I further show that fluctuations in asset prices, on one side, and in macroeconomic aggregates, on the other, are driven by two fundamentally different types of news. Namely, my results indicate that the main source of movements in the aggregate U.S. stock market is news about changes in the long-run component of investment-specific productivity. More specifically, that type of news explains about 84% of the variance of the ratio of total market valuation (TMV) to real GDP. (Abstract shortened by UMI.)