If returns are not predictable, dividend growth must be predictable, to generate the observed variation in divided yields. I find that the absence of dividend growth predictability gives stronger evidence than does the presence of return predictability. Long-horizon return forecasts give the same strong evidence. These tests exploit the negative correlation of return forecasts and dividend-yield autocorrelation across samples, together with sensible upper bounds on dividend-yield autocorrelation, to deliver more powerful statistics. I reconcile my findings with the literature that finds poor power in long-horizon return forecasts, and with the literature that notes the poor out-of-sample R² of return-forecasting regressions.

We investigate the empirical implications of using various measures of payout yield rather than dividend yield for asset pricing models. We find statistically and economically significant predictability in the time series when payout (dividends plus repurchases) and net payout (dividends plus repurchases minus issuances) yields are used instead of the dividend yield. Similarly, we find that payout (net payout) yields contains information about the cross section of expected stock returns exceeding that of dividend yields, and that the high minus low payout yield portfolio is a priced factor. WHILE THE IRRELEVANCE THEOREM of Miller and Modigliani (1961) implies that there is no reason to suspect that dividends play a role in determining equity price levels or equity returns, the theorem is silent on the usefulness of dividends in explaining these variables. It is then, perhaps, not surprising that there is a considerable literature exploiting the properties of dividends and dividend yields to better understand the fundamentals of asset pricing both in the time series and in the cross section. Motivation for the former comes from variations of the Gordon growth model in which dividend yields can be written as the return minus the dividend’s growth rate (see, e.g., Fama and French (1988)), from consumption-based asset pricing models in which the firm’s dividends covary with aggregate consumption (e.g., Lucas (1978) and Shiller (1981)), and so forth. Additional motivation comes from cross-sectional heterogeneity in tax, agency, and asymmetric information considerations (e.g.,

Goyal and Welch (2006) argue that the historical average excess stock return forecasts future excess stock returns better than regressions of excess returns on predictor variables. In this paper we show that many predictive regressions beat the historical average return, once weak restrictions are imposed on the signs of coefficients and return forecasts. The out-of-sample explanatory power is small, but nonetheless is economically meaningful for mean-variance investors. Even better results can be obtained by imposing the restrictions of steady-state valuation models, thereby removing the need to estimate the average from a short sample of volatile stock returns. Towards the end of the last century, academic finance economists came to take seriously the view that aggregate stock returns are predictable. During the 1980’s a number of papers studied valuation ratios, such as the dividend-price ratio, earningsprice ratio, or smoothed earnings-price ratio. Value-oriented investors in the tradition of Graham and Dodd (1934) had always asserted that high valuation ratios are an indication of an undervalued stock market and should predict high subsequent returns, but these ideas did not carry much weight in the academic literature until authors such as Rozeff (1984), Fama and French (1988), and Campbell and Shiller (1988a,b) found that valuation ratios are positively correlated with subsequent returns and that the implied predictability of returns is substantial at longer horizons. Around the same time, several papers pointed out that yields on short- and long-term Treasury and corporate bonds are correlated with subsequent stock returns [Fama and Schwert

We propose a theoretical framework for assessing whether a forecast model estimated over one period can provide good forecasts over a subsequent period. We formalize this idea by defining a forecast breakdown as a situation in which the out-of-sample performance of the model, judged by some loss function, is significantly worse than its in-sample performance. Our framework, which is valid under general conditions, can be used not only to detect past forecast breakdowns but also to predict future ones. We show that main causes of forecast breakdowns are instabilities in the data generating process and relate the properties of our forecast breakdown test to those of structural break tests. The empirical application finds evidence of a forecast breakdown in the Phillips ’ curve forecasts of U.S. inflation, and links it to inflation volatility and to changes in the monetary policy reaction function of the Fed.

I survey work on the intersection between macroeconomics and finance. The challenge is to find the right measure of “bad times,” rises in the marginal value of wealth, so that we can understand high average returns or low prices as compensation for assets’ tendency to pay off poorly in “bad times.” I survey the literature, covering the time-series and cross-sectional facts, the equity premium, consumption-based models, general equilibrium models, and labor income/idiosyncratic risk approaches.

We study an investor’s optimal consumption and portfolio choice problem when he confronts with two possibly misspecified submodels of stock returns: one with IID returns and the other with predictability. We adopt a generalized recursive ambiguity model to accommodate the investor’s aversion to model uncertainty. The investor deals with specification doubts by slanting his beliefs about submodels of returns pessimistically, causing his investment strategy to be more conservative than the Bayesian strategy. This effect is large for high and low values of the predictive variable. Unlike in the Bayesian framework, the hedging demand against model uncertainty may cause the investor’s stock allocations to first decrease sharply and then increase with his prior probability of the IID model, even when the expected stock return under the IID model is lower than under the predictability model. Adopting suboptimal investment strategies by ignoring model uncertainty can lead to sizable welfare costs.

The prevailing view in finance is that the evidence for long-horizon stock return predictability is significantly stronger than that for short horizons. We show that for persistent regressors, a characteristic of most of the predictive variables used in the literature, the estimators are almost perfectly correlated across horizons under the null hypothesis of no predictability. For example, for the persistence levels of dividend yields, the analytical correlation is 99% between the 1- and 2-year horizon estimators and 94 % between the 1- and 5-year horizons, due to the combined effects of overlapping returns and the persistence of the predictive variable. Common sampling error across equations leads to ordinary least squares coefficient estimates and R 2 s that are roughly proportional to the horizon under the null hypothesis. This is the precise pattern found in the data. The asymptotic theory is corroborated, and the analysis extended by extensive simulation evidence. We perform joint tests across horizons for a variety of explanatory variables, and provide an alternative view of the existing evidence. I.

This paper suggests a new state space approach to analysis of stock return predictability. Acknowledging that expected returns and expected dividends are unobservable, the Kalman filter technique is used to extract them from the observed history of realized dividends and returns. This approach explicitly accounts for the variation in expected dividend growth and allows to make estimates more robust to structural breaks in the means of dividend growth and returns. The constructed predictor outperforms the dividend-price ratio both in and out of sam-ple, providing statistically and economically significant forecasts. The finite sample likelihood ratio test reliably rejects the hypothesis of constant expected returns.

We provide an approach to forecasting the long-run (unconditional) distribution of equity returns making optimal use of historical data in the presence of structural breaks. Our focus is on learning about breaks in real time and assessing their impact on out-of-sample density forecasts. Forecasts use a probabilityweighted average of submodels, each of which is estimated over a different history of data. The empirical results strongly reject ignoring structural change or using a fixed-length moving window. The shape of the long-run distribution is affected by breaks, which has implications for risk management and long-run investment decisions. KEY WORDS:

We develop asset pricing models ’ implications for portfolio efficiency when there is conditioning information in the form of a set of lagged instruments. A model of expected returns identifies a portfolio that should be minimum variance efficient with respect to the conditioning information. Our framework refines previous tests of portfolio efficiency by using a given set of conditioning information optimally. The optimal use of the lagged variables is economically important. With standard portfolio designs and lagged instruments, by using the instruments optimally we reject several efficiency hypotheses that are not otherwise rejected. The Sharpe ratios of a sample of hedge fund indexes appear consistent with the optimal use of conditioning information.