Long-horizon Predictability Research Articles

Equilibrium Predictability, Term Structure of Equity Premia, and Other Return Characteristics

This article presents a structural model of aggregate return characteristics based on a one-channel Bansal and Yaron (2004) economy under recursive preferences. The results rest on an endogenously determined price–dividend ratio that is not exponentially affine, which implies time-variation and predictability of equity premia. The predictability coefficient itself is stochastic. This provides theoretical foundations for recent works in predictability like Dangl and Halling (2011). In longer horizons, the predictability relationship is highly volatile making it difficult to make inferences about long-horizon predictability.

Nonparametric testing for long-horizon predictability with persistent covariates

We propose a testing procedure for long-horizon predictability via kernel-based nonparametric estimators of long-run covariances between multiperiod returns and persistent covariates. Asymptotic properties of the proposed tests are studied. As for implementation of the test, sieve bootstrap methods are employed to obtain reasonable approximation to the sample distribution of the test statistics. Monte Carlo simulations are conducted to verify the theoretical conjecture. Empirical analysis, using US monthly data from 1929 to 2011, are presented for testing stock return predictability of some forecasting financial variables. Long-term interest rates, unlike default spreads or price-earning ration, are found to show some forecasting power.

Comment

Comment

Short-horizon return predictability and oil prices

This paper shows that oil price changes, measured as short-term futures returns, are a strong predictor of excess stock returns at short horizons. Ours is a leading variable for the business cycle and exhibits low persistence which avoids the fictitious long-horizon predictability associated with other predictors used in the literature. We compare our variable with the most popular predictors in a sample period that includes the recent financial crisis. Our results suggest that oil price changes are the only variable with forecasting power for stock returns. This significant predictive ability is robust against the inclusion of other variables and out-of-sample tests. We also study the cross-section of expected stock returns in a conditional CAPM framework based on oil price shocks. Our model displays high statistical significance and a better fit than all the conditional and unconditional models considered, including the Fama–French three-factor model. From a practical perspective, ours is a high-frequency, observable variable that has the advantage of being readily available to market-timing investors.

A Path-Planning Method for Human-Tracking Agents Based on Long-Term Prediction

This paper deals with a multiagent path-planning problem where several robots track humans to obtain detailed information on human behaviors and characteristics. For this, agents' paths are planned on the basis of the similarity between the predicted positions of humans and the agents' field of view. The long-horizon path planned on the basis of an accurate long-horizon prediction improves the tracking performance. However, it requires heavy computation and is less useful if the prediction is inaccurate. Since the accuracy of the prediction depends on the situation, the prediction term is determined by the similarity between the current and previous predictions. The results of computer simulation showed that our path-planning method works well for trajectories of humans in a dynamic environment by changing the horizon length of the path planning.

Modelling Time-Variation in the Stock Return-Dividend Yield Predictive Equation

Using data for forty markets, this paper examines the nature and possible causes of time-variation within the stock return-dividend yield predictive regression. The results in this paper show that there is significant time-variation in the predictive equation for returns and that such variation is linked to economic and market factors. Furthermore, the strength and nature of those links are themselves time-varying. The inclusion of this time-variation in the predictive equation increases the predictive power compared to the standard constant parameter predictive model. Evidence is also reported for time-varying dividend growth predictability. Long-horizon predictability is also examined with evidence reported that the nature of the factors affecting time-varying predictability changes with horizon. The results here, while directly contributing to the returns predictability debate, in particular regarding its existence and source, may also inform the discussion that links time-varying expected returns (and risk premium) to economic factors.

Return Predictability under the Alternative

Long-horizon predictability is not a myth. We propose a new analytical standard error for predictive regressions that does not impose the null hypothesis that returns are unpredictable and exhibits substantial power gains relative to popular tests. Deriving the covariance matrix under the alternative hypothesis produces two new terms capturing the volatility of shocks to the regressor and their correlation with shocks to the prediction equation. Empirically, we show that failure to detect long-horizon predictability comes from lower power in tests derived under the null hypothesis. For many predictors, giving the alternative a chance allows short-run predictability to survive at long-horizons.

Exchange rates and individual good's price misalignment: Evidence of long-horizon predictability

Although purchasing-power-parity fundamentals, in general, have only weak predictability, currency misalignment may be indicated by price differentials for some individual goods, which could then have predictive power for subsequent re-evaluation of the nominal exchange rate. We collect good-level price data to construct deviations from the law of one price and examine the resulting price-misalignment model's predictive power for the nominal exchange rates between the U.S. dollar and two other currencies: the Japanese yen and the U.K. pound. We find that the slope coefficients and R-squares of in-sample forecasting regressions for almost all goods in our data increase with the forecast horizon for the bilateral exchange rates between the U.S. dollar and the Japanese yen and the U.S. dollar and the U.K. pound. The results of tests for out-of-sample superior predictive ability suggest that our price-misalignment model outperforms a random walk model either with or without drift for the U.S. dollar vis-à-vis the Japanese yen at the 5 percent level of significance over long horizons.

Wage Rigidity: A Solution to Several Asset Pricing Puzzles

This paper studies asset prices in a general equilibrium model with staggered wage contracts. We show that infrequent renegotiation of labor wages is crucial for production-based models to explain the high equity return volatility (excess volatility puzzle) observed in the data. Our preferred model combining wage rigidity with long run productivity risk can produce several hard to explain features of both financial and business cycle data: i) high unconditional equity premium and equity volatility; ii) countercyclical variation in the equity premium, in its volatility, and in Sharpe Ratio; iii) long-horizon predictability of returns with high R² values; iv) a higher expected returns for value stocks over growth stocks; and v) smooth wages and volatile aggregate profits. The intuition is that in standard models, highly pro-cyclical and volatile wages act as a hedge for the firm, reducing profits in good times and increasing them in bad times; this causes profit and returns to be too smooth. Infrequent renegotiation smoothes wages and smooth wages act like operating leverage, making profits more risky. Bad times and value firms are especially risky because committed wage payments are high relative to output. Consistent with our model, we show that in the data wage growth can forecast long horizon returns; furthermore we find the same predictability at the industry level, with more rigid industries having stronger predictability; lastly we show that growth stocks have lower degree of wage rigidity than value stocks in the data.

Asset Prices and the Heterogeneous Impact of News

This paper analyzes a general equilibrium exchange economy with a continuum of agents who have catching up with the Joneses preferences and differ with respect to the curvature of their utility functions and - as a peculiarity - the impact of news about uncertainty. The dynamic redistribution of wealth among the agents generates endogenously the state-dependent preferences of the representative agent and the variation in the pricing kernel volatility, i.e. the price of risk. The model exhibits many of the empirically observed properties of the unconditional and conditional moments of stock returns. Indeed, the new source of heterogeneity has a strong impact on asset prices whereas heterogeneity in risk aversion has just a marginal effect once it is estimated from data. In particular the heterogeneous impact of news helps in explaining the equity premium and risk-free rate puzzles, the high and countercyclical Sharpe ratio, the long-horizon predictability of excess returns and the GARCH and leverage effects in return volatility.

Equilibrium Predictability and Other Return Characteristics

This paper presents a structural model of aggregate return characteristics based on a one-channel Bansal and Yaron (2004) economy under recursive preferences. The predictability result rests on an endogenously determined price-dividend ratio that is not exponentially affine which implies time variation and predictability in equity premia. This result is new within the context of a one-channel Bansal and Yaron (2004) economy. Furthermore, the predictability coefficient is stochastic which provides theoretical foundations for recent works in predictability like Dangl and Halling (2011). In longer horizon, the predictability relationship is highly volatile making it difficult to make inference about long-horizon predictability.

Long-Horizon Stock Predictability: Evidence and Applications

Long-Horizon Stock Predictability: Evidence and Applications

The dividend–price ratio does predict dividend growth: International evidence

Abstract Unpredictable dividend growth by the dividend–price ratio is considered a ‘stylized fact’ in post war US data. Using long-term annual data from the US and three European countries, we revisit this stylized fact, and we also report results on return predictability. We make two main contributions. First, we document that for the US, results for long-horizon predictability are crucially dependent on whether returns and dividend growth are measured in nominal or real terms, and this difference is due to long-term inflation being strongly negatively predictable by the dividend–price ratio. The impact of inflation is to reinforce real return predictability and to reduce – or change direction of – real dividend growth predictability. This provides an explanation for the strong predictability of long-horizon real returns in the ‘right’ direction, and the strong predictability of long-horizon real dividend growth in the ‘wrong’ direction, that we see in US post war data. Second, we find that predictability patterns in three European stock markets are in many ways different from what characterize the US stock market. In particular, in Sweden and Denmark dividend growth is strongly predictable by the dividend–price ratio in the ‘right’ direction while returns are not predictable. The results for the UK are mixed. Our results are robust to a number of changes in the modeling framework. We discuss the results for dividend growth predictability in terms of the ‘dividend smoothing hypothesis’.

Mean Reversion in Stock Prices in an Emerging Market: Sri Lankan Evidence

This paper examines long-horizon predictability of stock prices in Sri Lanka using 1-4 year real returns on indices for market and industry portfolio during the 1985-97 period. In a model of prices, which contains a random walk component and a slowly decaying transitory price component, which are uncorrelated, the slope of regression of returns of certain period on lagged return of equal length measures the predictable variation of returns. The results show that, consistent with the predictions of the model and findings of previous research, long-horizon returns in Sri Lanka are highly negatively correlated implying mean reversion. Predictable variation in 1-3 year returns is very large in the order of about 50 percent. The results also reliably reject the random walk behaviour of prices.

Structural breaks in financial ratios: evidence for nine international markets

Financial ratios have recently reached unprecedented levels and despite price falls remain at ahistoric levels. This is at odds with the theoretical present value model. Whilst, several researchers have attempted to reconcile theory and data using fractional integration and nonlinear techniques, an alternate view is that such ratios do not revert to a single point but exhibit a time-variation in their level. This article, tests for and supports, the belief that financial ratios exhibit structural breaks, or level shifts, through time. As such, the present value model does not hold exactly, but needs to accommodate such shifts. This may also explain apparent contradictory findings in long-horizon return predictability regressions performed over different sample periods.

A New Application of Exact Nonparametric Methods to Long-Horizon Predictability Tests

Empirical results from long-horizon regression tests have been influential in the finance literature. Yet, it has come to be understood that traditional long-horizon tests may be unreliable in finite samples when regressors are persistent and when the horizon is long relative to sample size. Recent research has provided valid alternative inference procedures in long-horizon regression in the case for which the regressor follows a near-unit root autoregressive process. However, in small samples, such processes may sometimes be difficult to distinguish with confidence from other persistent data generating processes, such as those displaying long-memory or structural breaks. In this paper, we demonstrate a simple means by which existing nonparametric sign and signed rank tests may be applied to provide exact inference in long-horizon predictive tests, without requiring any modeling assumptions on the regressor. Employing this robust approach, we find evidence of stock return predictability at moderate horizons using short-term interest rates, but little evidence of either short or long-run predictability using dividend-price ratios.

Lock-in of Extrapolative Expectations in an Asset Pricing Model

This paper examines an agent's choice of forecast method within a standard asset pricing model. To make a conditional forecast, a representative agent may choose one of the following: (1) a rational (or fundamentals-based) forecast that employs knowledge of the stochastic process governing dividends, (2) a constant forecast based on a simple long-run average of the forecast variable, or (3) a time-varying forecast that extrapolates from the last observation of the forecast variable. I show that a representative agent who is concerned about minimizing forecast errors may inadvertently become locked-in to an extrapolative forecast. In particular, the initial use of extrapolation alters the law of motion of the forecast variable so that the agent perceives no accuracy gain from switching to one of the alternative forecast methods. Under extrapolative expectations, the model can generate excess volatility of stock prices, time-varying volatility of returns, long-horizon predictability of returns, bubbles driven by optimism about the future, and sharp downward movements in stock prices that resemble market crashes. All of these features appear to be present in long-run U.S. stock market data.

Stock Return Predictability: Is it There?

We ask whether stock returns in France, Germany, Japan, the UK and the US are predictable by three instruments: the dividend yield, the earnings yield and the short rate. The predictability regression is suggested by a present value model with earnings growth, payout ratios and the short rate as state variables. We use this model imposing a constant risk premium to examine the finite sample evidence on predictability. Not only do we find the short rate to be a relevant state variable theoretically, it is also the only robust short-run predictor of equity returns. The evidence in Lamont (1998) on earnings and dividend yield predictability is not robust to our increased sample period, does not survive finite sample corrections and does not extend to other countries. We find no evidence of long-horizon predictability once we account for finite sample influence. Finally, cross-country predictability appears stronger than predictability using local instruments.

Non-linear long horizon returns predictability: evidence from six south-east Asian markets

Whilst the existence of long-horizon returns predictability has been a recurrent theme in empirical finance research, extant work has focused almost exclusively on the US, and more recent work has cast doubt over the validity of such potential predictability due to non-stationarity and serial correlation in the data. The present paper examines long-horizon returns predictability in six South-East Asian markets in the context of unit root tests conducted under corrections for serial correlation and heteroscedasticity. The analysis conducted also extends the investigation of long-horizon predictability to the non-linear setting, and examines whether any detected non-linear predictability is consistent with behavioural approaches to asset pricing which emphasise the role of noise traders. The results obtained suggest the following conclusions. First, long-horizon predictability is present in each of the six South-East Asian markets considered. Second, whilst forecast power increases with horizon, for the majority of series it is maximised for forecast horizons of between 12 and 48 months. Third, non-linear predictability is reported for all series, suggesting that positive and negative values of the (demeaned) dividend yield impart different levels of predictability for returns at different forecast horizons. However, there is no consistency in the pattern of non-linearity reported, and whilst the observed patterns are consistent with noise traders models for some of the markets considered, this is not true of all the markets considered. Nevertheless, the non-linear dynamics detected are suggestive of potential market inefficiencies in all six cases.

LOCK-IN OF EXTRAPOLATIVE EXPECTATIONS IN AN ASSET PRICING MODEL

This paper examines an agent's choice of forecast method within a standard asset pricing model. A representative agent may choose: (1) a fundamentals-based forecast that employs knowledge of the dividend process, (2) a constant forecast that is based on a simple long-run average, or (3) a time-varying forecast that extrapolates from the last observation. I show that an agent who is concerned about minimizing forecast errors may inadvertently become “locked-in” to an extrapolative forecast. In particular, the initial use of extrapolation alters the law of motion of the forecast variable so that the agent perceives no accuracy gain from switching to one of the alternative forecast methods. The model can generate excess volatility of stock prices, time-varying volatility of returns, long-horizon predictability of returns, bubbles driven by optimism about the future, and sharp downward movements in stock prices that resemble market crashes.