International stock return predictability under model uncertainty

This paper examines return predictability when the investor is uncertain about the right state variables. A novel feature of the model averaging approach used in this paper is to account for finite-sample bias of the coefficients in the predictive regressions. Drawing on an extensive international dataset, we find that interest-rate related variables are usually among the most prominent predictive variables, whereas valuation ratios perform rather poorly. Yet, predictability of market excess returns weakens substantially, once model uncertainty is accounted for. We document notable differences in the degree of in-sample and out-of-sample predictability across different stock markets. Overall, these findings suggest that return predictability is neither a uniform, nor a universal feature across international capital markets.     

Stock return predictability and model uncertainty

We use Bayesian model averaging to analyze the sample evidence on return predictability in the presence of model uncertainty. The analysis reveals in-sample and out-of-sample predictability, and shows that the out-of-sample performance of the Bayesian approach is superior to that of model selection criteria. We find that term and market premia are robust predictors. Moreover, small-cap value stocks appear more predictable than large-cap growth stocks. We also investigate the implications of model uncertainty from investment management perspectives. We show that model uncertainty is more important than estimation risk, and investors who discard model uncertainty face large utility losses.     

Implementing statistical criteria to select return forecasting models: what do we learn?

Statisticl model selection criteria provide an informed choiceof the model with best external (i.e., out-of-sample) validity.Therefore they guard against overfitting ('data snooping').We implement several model selection criteria in order to verifyrecent evidence of predictability in excess stock returns andto determine which variables are valuable predictors. We confirmthe presence of in-sample predictability in an internationalstock market dataset, but discover that even the best predictionmodels have no out-of-sample forecasting power. The failureto detect out-of-sample predictability is not due to lack ofpower.     

Tactical Industry Allocation and Model Uncertainty

We use Bayesian model averaging to analyze industry return predictability in the presence of model uncertainty. The posterior analysis shows the importance of inflation and earnings yield in predicting industry returns. The out‐of‐sample performance of the Bayesian approach is, in general, superior to that of other statistical model selection criteria. However, the out‐of‐sample forecasting power of a naive i.i.d. forecast is similar to the Bayesian forecast. A variance decomposition into model risk, estimation risk, and forecast error shows that model risk is less important than estimation risk.     

Industry return predictability,timing and profitability

This paper aims to investigate the predictability of Australian industrial stock returns. Several identified economic variables are found to contain significant predictive power over industry portfolio returns in a Bayesian dynamic forecasting model. The Bayesian updating process was also applied in an investigation of out-of-sample prediction, timing ability and the profitability of an investment strategy of industry-rotation. When the predictor variables are employed in out-of-sample analysis, the predictive power is superior to the naïve prediction. The timing ability and profitability associated with predictability are also economically significant. When the industry momentum is examined, the results show that a group-rotation strategy can enhance the portfolio performance.     

Market timing and return prediction under model instability

Despite mounting empirical evidence to the contrary, the literature on predictability of stock returns almost uniformly assumes a time-invariant relationship between state variables and returns. In this paper, we propose a two-stage approach for forecasting of financial return series that are subject to breaks. The first stage adopts a reversed ordered Cusum (ROC) procedure to determine in real time when the most recent break has occurred. In the second stage, post-break data is used to estimate the parameters of the forecasting model. We compare this approach to existing alternatives for dealing with parameter instability such as the Bai–Perron method and the time-varying parameter (TVP) model. An out-of-sample forecasting experiment demonstrates considerable gains in market timing precision from adopting the proposed two-stage forecasting method.     

In-sample vs. out-of-sample tests of stock return predictability in the context of data mining

We undertake an extensive analysis of in-sample and out-of-sample tests of stock return predictability in an effort to better understand the nature of the empirical evidence on return predictability. We find that a number of financial variables appearing in the literature display both in-sample and out-of-sample predictive ability with respect to stock returns in annual data covering most of the twentieth century. In contrast to the extant literature, we demonstrate that there is little discrepancy between in-sample and out-of-sample test results once we employ out-of-sample tests with good power. While conventional wisdom holds that out-of-sample tests help guard against data mining, Inoue and Kilian [Inoue, A., Kilian, L., 2004. In-sample or out-of-sample tests of predictability: which one should we use? Econometric Reviews 23, 371–402.] recently argue that in-sample and out-of-sample tests are equally susceptible to data mining biases. Using a bootstrap procedure that explicitly accounts for data mining, we still find that certain financial variables display significant in-sample and out-of-sample predictive ability with respect to stock returns.     

Predicting the bear stock market: Macroeconomic variables as leading indicators

This paper investigates whether macroeconomic variables can predict recessions in the stock market, i.e., bear markets. Series such as interest rate spreads, inflation rates, money stocks, aggregate output, unemployment rates, federal funds rates, federal government debt, and nominal exchange rates are evaluated. After using parametric and nonparametric approaches to identify recession periods in the stock market, we consider both in-sample and out-of-sample tests of the variables’ predictive ability. Empirical evidence from monthly data on the Standard & Poor’s S&P 500 price index suggests that among the macroeconomic variables we have evaluated, yield curve spreads and inflation rates are the most useful predictors of recessions in the US stock market, according to both in-sample and out-of-sample forecasting performance. Moreover, comparing the bear market prediction to the stock return predictability has shown that it is easier to predict bear markets using macroeconomic variables.     

Stock market return predictability: A combination forecast perspective

Based on traditional macroeconomic variables, this paper mainly investigates the predictability of these variables for stock market return. The empirical results show the mean combination forecast model can achieve superior out-of-sample performance than the other forecasting models for forecasting the stock market returns. In addition, the performances of the mean combination forecast model are also robust during different forecasting windows, different market conditions, and multi-step-ahead forecasts. Importantly, the mean combination forecast consistently generates higher CER gains than other models considering different investors' risk aversion coefficients and trading costs. This paper tries to provide more evidence of combination forecast model to predict stock market returns.     

Out-of-Sample Predictability in International Equity Markets: A Model Selection Approach

For 13 major international stock markets, there is much evidence of out-of-sample predictability for growth stocks especially when evaluated with economic criteria, and to a noticeably lesser extent for general stock markets and value stocks. Our results shed light on the recent debate about stock return predictability, using different assets (growth-style indexes), forecasting variables (past returns), forecasting models (nonlinear models), and alternative forecasting evaluation criteria (economic criteria). Our analysis suggests that (growth) stock returns might be predictable.     

Pockets of Predictability

For many benchmark predictor variables, short-horizon return predictability in the U.S. stock market is local in time as short periods with significant predictability (“pockets”) are interspersed with long periods with no return predictability. We document this result empirically using a flexible time-varying parameter model that estimates predictive coefficients as a nonparametric function of time and explore possible explanations of this finding, including time-varying risk premia for which we find limited support. Conversely, pockets of return predictability are consistent with a sticky expectations model in which investors slowly update their beliefs about a persistent component in the cash flow process.     

Learning about Predictability: The Effects of Parameter Uncertainty on Dynamic Asset Allocation

This paper examines the effects of uncertainty about the stock return predictability on optimal dynamic portfolio choice in a continuous time setting for a long-horizon investor. Uncertainty about the predictive relation affects the optimal portfolio choice through dynamic learning, and leads to a state-dependent relation between the optimal portfolio choice and the investment horizon. There is substantial market timing in the optimal hedge demands, which is caused by stochastic covariance between stock return and dynamic learning. The opportunity cost of ignoring predictability or learning is found to be quite substantial.     

Out-of-Sample Equity Premium Predictability in South Africa: Evidence from a Large Number of Predictors

This article uses a predictive regression framework to examine the out-of-sample predictability of South Africa’s equity premium, using a host of financial and macroeconomic variables. We employ various methods of forecast combination, bootstrap aggregation (bagging), diffusion index (principal component), and Bayesian regressions to allow for a simultaneous role of the variables under consideration, besides individual predictive regressions. We assess both the statistical and economic significance of the individual predictive regressions, combination methods, bagging, principal components, and Bayesian regressions. Our results show that forecast combination methods and principal component regressions improve the predictability of the equity premium relative to the benchmark autoregressive model of order one (AR[1]). However, the Bayesian predictive regressions are found to be the standout performers with the models outperforming the individual regressions, forecast combination methods, bagging and principal component regressions, both in terms of statistical (forecasting) and economic (utility) gains.     

Stock Return Predictability: A Bayesian Model Selection Perspective

Attempts to characterize stock return predictability have resultedin little consensus on the important conditioning variables,giving rise to model uncertainty and data snooping fears. Weintroduce a new methodology that explicitly incorporates modeluncertainty by comparing all possible models simultaneouslyand in which the priors are calibrated to reflect economicallymeaningful information. Our approach minimizes data snoopinggiven the information set and the priors. We compare the priorviews of a skeptic and a confident investor. The data implyposterior probabilities that are in general more supportiveof stock return predictability than the priors for both typesof investors.     

Predicting stock returns: A regime-switching combination approach and economic links

This paper introduces a regime-switching combination approach to predict excess stock returns. The approach explicitly incorporates model uncertainty, regime uncertainty, and parameter uncertainty. The empirical findings reveal that the regime-switching combination forecasts of excess returns deliver consistent out-of-sample forecasting gains relative to the historical average and the Rapach et al. (2010) combination forecasts. The findings also reveal that two regimes are related to the business cycle. Based on the business cycle explanation of regimes, excess returns are found to be more predictable during economic contractions than during expansions. Finally, return forecasts are related to the real economy, thus providing insights on the economic sources of return predictability.     

Payout yield,risk, and mispricing: A Bayesian analysis

We develop a simple parametric model in which hypotheses about predictability, mispricing, and the risk-return tradeoff can be evaluated simultaneously, while allowing for time variation in both risk and expected return. Most of the return predictability based on aggregate payout yield is unrelated to market risk. We consider a range of Bayesian prior beliefs about the risk-return tradeoff and the extent to which predictability is driven by mispricing. The impact of these beliefs on an investor's certainty-equivalent return when choosing between a market index and riskless T-bills is economically significant, in both ex ante and out-of-sample analyses.     

Forecasting the price of gold using dynamic model averaging

We develop several models to examine possible predictors of the return of gold, which embrace six global factors (business cycle, nominal, interest rate, commodity, exchange rate and stock price) extracted from a recursive principal component analysis (PCA) and two uncertainty and stress indices (the Kansas City Fed's financial stress index and the U.S. economic policy uncertainty index). Specifically, by comparing alternative predictive models, we show that the dynamic model averaging (DMA) and dynamic model selection (DMS) models outperform linear models (such as the random walk) as well as the Bayesian model averaging (BMA) model. The DMS is the best predictive model overall across all forecast horizons. Generally, all the predictors show strong predictive power at one time or another though at varying magnitudes, while the exchange rate factor and the Kansas City Fed's financial stress index appear to be strong at almost all horizons and sub-periods. However, the forecasting prowess of the exchange rate is supreme.     

Do oil futures prices predict stock returns?

This paper explores stock return predictability by exploiting the cross-section of oil futures prices. Motivated by the principal component analysis, we find the curvature factor of the oil futures curve predicts monthly stock returns: a 1% per month increase in the curvature factor predicts 0.4% per month decrease in stock market index return. This predictive pattern is prevailing in non-oil industry portfolios, but is absent for oil-related portfolios. The in- and out-of-sample predictive power of the curvature factor for non-oil stocks is robust and outperforms many other predictors, including oil spot prices. The predictive power of the curvature factor comes from its ability to forecast supply-side oil shocks, which only affect non-oil stocks and are hedged by oil-related stocks.     

Stock return predictability and the adaptive markets hypothesis: Evidence from century-long U.S. data

This paper provides strong evidence of time-varying return predictability of the Dow Jones Industrial Average index from 1900 to 2009. Return predictability is found to be driven by changing market conditions, consistent with the implication of the adaptive markets hypothesis. During market crashes, no statistically significant return predictability is observed, but return predictability is associated with a high degree of uncertainty. In times of economic or political crises, stock returns have been highly predictable with a moderate degree of uncertainty in predictability. We find that return predictability has been smaller during economic bubbles than in normal times. We also find evidence that return predictability is associated with stock market volatility and economic fundamentals.