Suppose that the equity premium is forecasted by dividend yields. Even if such a relationship does exist, there is so much noise in the equity premium that estimation, inference and forecasting cannot be carried out using the faint signal coming from the dividend yields. For analyzing equity/dividend data, it is useful to quantify the signal in a given sample. We define an index of signal strength or information accumulation, by renormalizing the signal to noise ratio. The novelty in our parameterization is that the index of information accumulation explicitly influences rates of convergence and can even lead to inconsistent estimation, inconsistent testing, unreliable R^20s and no out of sample forecasting power. Indeed, we prove that if the signal to noise ratio is close to zero, forecasts from the existing model will not do better than the simple unconditional mean. Thus, it is not surprising that dividend yield forecasts of the equity premium cannot outperform its mean. The analytic framework is general enough to capture most previous econometric findings related to the equity/dividend relationship.

We analyze several ways of conducting long-horizon regressions, taken from the empirical literature. Asymptotic arguments are used to show that, in all cases, the t-statistics do not converge to well-defined distributions, thus explaining the tendency of long-horizon regressions to find ‘significant” results, where previous short-term approaches have failed. Moreover, in some cases, the ordinary least squares estimator is not consistent, and the R^2 cannot be interpreted as a measure of the goodness of fit. Those results cast doubt on the conclusions reached by most previous long-horizon regression studies. We propose a rescaled t-statistic, whose asymptotic distribution is easy to simulate, and re-visit some of the evidence on the long-horizon predictability of returns and the long-horizon tests of the Fisher Effect.

If yields are assumed to have an exact unit-root, it has previously been shown that the rational expectations hypothesis of the term structure (REHTS) has been rejected by single-equation tests. However, small deviations from exact unit-root produce substantial changes in the small sample distributions of those tests and the normal approximation is no longer satisfactory. We assume that the yield of 1-period zero-coupon bond follows a local-to-unity process with parameter c (c=0 for exact unit root) and use asymptotics to derive alternative distributions, which are far better approximations to finite sample distributions. Those asymptotic distributions depend crucially on c, and that allows us to analyze the impact of small deviations from unit-root on the distribution of the tests. Interestingly, for small values of c, the results obtained in the data do not imply a rejection of the REHTS. The above results are useful only when c is known or consistently estimable. Thus, the REHTS is cast into a triangular representation, where the cointegrating vectors are a function of c. Consistent and asymptotically unbiased estimators of c are proposed. A Wald test for the restrictions imposed by the REHTS on the cointegrating relationship is derived. The relevance of the asymptotic results for samples of practical sizes is investigated with Monte Carlo simulations. The methods are applied to the yield data of McCulloch and Kwon (1993). Although the REHTS is statistically rejected, the results are encouraging and suggest interesting directions for further research.

Even if returns are truly forecasted by variables such as the dividend yield, the noise in such a predictive regression may overwhelm the signal of the conditioning variable and render estimation, inference and forecasting unreliable. Unfortunately, traditional asymptotic approximations are not suitable to investigate the small sample properties of forecasting regressions with excessive noise. To systematically analyze predictive regressions, it is useful to quantify a forecasting variable’s signal relative to the noisiness of returns in a given sample. We define an index of signal strength, or information accumulation, by renormalizing the signal-noise ratio. The novelty of our parameterization is that this index explicitly influences rates of convergence and can lead to inconsistent estimation and testing, unreliable R2s, and no out-of-sample forecasting power. Indeed, we prove that if the signal-noise ratio is close to zero, as is the case for many of the explanatory variables previously suggested in the finance literature, model based forecasts will do no better than the corresponding simple unconditional mean return. Our analytic framework is general enough to capture most of the previous findings surrounding predictive regressions using dividend yields and other persistent forecasting variables.

We find that the average excess return in the stock market is higher under Democratic than Republican presidents– a difference of 9 percent per year for the value-weighted portfolio and 16 percent for the equal-weighted portfolio. The difference is economically and statistically significant, does not seem to be due to small sample biases, and is robust in different subsamples. There is a remarkable monotonicity in the difference of returns for size-decile portfolios, from 7 percent for large firms to about 22 percent for small firms. Presidential partisan cycles have a heterogeneous impact on industry returns: the tobacco, telecom, and chemical industries have performed better under Republican presidents, whereas the real estate, construction, and services industries have fared significantly better under Democrats. We test three plausible explanations for these findings. First, the relation might be due to political variables proxying for business-cycle factors. Second, the relation might be attributed to unexpected returns around elections, when information is revealed, rather than to expected returns varying with the political cycle. Lastly, differences in stock market riskiness across presidential regimes could account for the difference in average returns. We reject all three hypotheses. As it stands, the difference in excess returns during Republican and Democratic presidencies is a puzzle that cannot easily be explained. However, the cross-sectional evidence from size-sorted and industry portfolios suggests that the party in the presidency may affect the stock market through differences in fiscal and regulatory policies.

We test the hypothesis that the gradual diffusion of information across asset markets leads to cross-asset return predictability. Using thirty-four industry portfolios and the broad market index as our test assets, we establish several key results. First, a number of industries such as retail, services, commercial real estate, metal, and petroleum lead the stock market by up to two months. In contrast, the market, which is widely followed, only leads a few industries. Importantly, an industry’s ability to lead the market is correlated with its propensity to forecast various indicators of economic activity such as industrial production growth. Consistent with our gradual-information-diffusion hypothesis, these findings indicate that the market reacts with a delay to information in industry returns about its fundamentals.

We introduce Mixed Data Sampling (henceforth MIDAS) regression models. The regressions involve time series data sampled at different frequencies. Technically speaking MIDAS models specify conditional expectations as a distributed lag of regressors recorded at some higher sampling frequencies. We examine the asymptotic properties of MIDAS regression estimation and compare it with traditional distributed lag models. MIDAS regressions have wide applicability in macroeconomics and �nance.

We investigate whether the cap rate, that is, the rent-price ratio in commercial real estate incorporates information about future expected real estate returns and future growth in rents. Relying on transactions data spanning several years across fifty-three metropolitan areas in the U.S., we find that the cap rate captures fluctuations in expected returns for apartments, retail, as well as industrial properties. For offices, by contrast, the cap rate does not forecast returns even though adsditional evidence reveals that expected returns on offices are also time-varying. We link this varying success of the cap rate in forecastings commercial property returns to differences in the stochastic properties of their rental growth rates. The growth in office rents has a higher correlation with expected returns and is more volatile than for other property types. Taken together, these characteristics diminish the correlation between the cap rate and the future returns to offices.

This dissertation comprises three papers examining several questions in finance and accounting. A common thread is investigating the strategic interactions between public firms and stock market investors. Chapter 1 studies how investors with short-term horizons can impact firms' behaviors. Chapters 2 and 3 examine the impact of corporate disclosure and the market pricing of information.

In Chapter 1, I use the unique features of the margin trading system in China to identify the causal impact of transient investors on managerial myopia. Specifically, I employ a regression discontinuity design that exploits the ranking procedure that determines a stock's margin trading eligibility. I find that margin traders are extremely short-term oriented and cause a sharp increase in stock share turnover. Moreover, marginable firms cater to these transient investors by manipulating current earnings and reducing long-term investments. Consistent with managerial myopia, these firms experience a short-term price increase but a long-term decline in operating performance.

Chapter 2 is joint work with John Hughes, Jun Liu, and Dan Yang. We reexamine the relation between disclosure indices and cost of equity capital employing an empirical specification similar to that of \citet{botosan97} for a substantially larger sample over an extended time frame made possible by textual analysis. Our results provide no support for a hypothesis of a negative relation between disclosure indices and implied cost of equity capital. Rather, consistent with a bias of implied cost of equity capital as a proxy for expected return depicted by \citet{Hughes2009}, we find strong evidence of a positive relation.

Chapter 3 is joint work with Yibin Liu. We exploit an earnings-based delisting policy and examine its adverse effect on investor trust in earnings news. Besides providing prominent visual evidence of large-scale earnings management at the required earnings threshold, we find that firms close to this threshold are trusted less by investors, regardless of whether they have manipulated earnings. Moreover, we provide causal evidence by studying firms that approach this threshold due to a plausibly exogenous profitability shock. Our results suggest that earnings-based regulations with harsh punishment may lead to a decline in investor trust.