Forecasting stock returns with large dimensional factor models

We study equity premium out-of-sample predictability by extracting the information contained in a high number of macroeconomic predictors via large dimensional factor models. We compare the well-known factor model with a static representation of the common components with the Generalized Dynamic Factor Model, which accounts for time series dependence in the common components. Using statistical and economic evaluation criteria, we empirically show that the Generalized Dynamic Factor Model helps predicting the equity premium. Exploiting the link between business cycle and return predictability, we find accurate predictions also by combining rolling and recursive forecasts in real-time.     

Improved estimation of the covariance matrix of stock returns with an application to portfolio selection

This paper proposes to estimate the covariance matrix of stock returns by an optimally weighted average of two existing estimators: the sample covariance matrix and single-index covariance matrix. This method is generally known as shrinkage, and it is standard in decision theory and in empirical Bayesian statistics. Our shrinkage estimator can be seen as a way to account for extra-market covariance without having to specify an arbitrary multifactor structure. For NYSE and AMEX stock returns from 1972 to 1995, it can be used to select portfolios with significantly lower out-of-sample variance than a set of existing estimators, including multifactor models.     

A 10-year retrospective on the determinants of Russian stock returns

We study factors influencing returns at the Russian stock market from 1995 to 2004, putting emphasis on how these evolved over time. We find that the relationship is highly unstable and this instability is not confined to financial crises alone. Most computed statistics exhibit constant ups and downs, but there has been recently a sharp rise in explainability of stock returns. Domestic factors have been playing a gradually diminishing role, while the importance of international factors has been increasing. In recent years, the effect of oil prices and foreign exchange rates has diminished, the impact of US stock prices and international and domestic interest rates has increased, while the influence of monetary aggregates such as gold reserves and credit balances has fallen to practically zero.     

Trading volume and stock returns: A meta-analysis

I examine 468 estimates on the relationship between trading volume and stock returns reported in 44 studies. I study publication bias together with Bayesian and frequentist model averaging to explain the heterogeneity in the estimates. The results yield three key conclusions. First, publication bias distorts the findings of the primary studies. Second, the predictability of stock returns varies with different markets and stock types. Third, different data characteristics, structural variations and methodologies used drive the heterogeneity in the results of the primary articles. In particular, one should be cautious when using monthly data or VAR models.     

Presidential administrations and the day-of-the-week effect in stock returns

This study examines the presence of a day-of-the-week effect over different presidential administrations. The results indicate that the day-of-the-week effect prevails during the Democratic and Republican administrations. However, the pattern of the day-of-the-week effect differs between the two presidential administrations. Specifically, the negative returns on Monday are more pronounced during the Republican than during the Democratic administrations. Therefore, explanations for the day-of-the-week effect should take into account the changing pattern of the day-of-the-week effect across presidential administrations.     

The role of current and non-current accruals in the relation between stock returns and earnings

This study examines the role of accruals in the relation between stock returns and earnings for intervals of one to four years. We argue that the roles of current and non-current accruals differ because the former turn over more frequently while the latter include long term timing differences and permanent differences. Accordingly, the roles of both categories of accruals are examined over intervals within and beyond the cycle of current accruals. The results suggest that accruals strengthen the association between stock returns and earnings and that they are more important for shorter intervals. Further, non-current accruals play a dominant role in the relation between stock returns and earnings while the effect of current accruals is negligible for all intervals examined.     

Identifying common dynamic features in stock returns

This paper proposes volatility and spectral based methods for the cluster analysis of stock returns. Using the information about both the estimated parameters in the threshold GARCH (or TGARCH) equation and the periodogram of the squared returns, we compute a distance matrix for the stock returns. Clusters are formed by looking to the hierarchical structure tree (or dendrogram) and the computed principal coordinates. We employ these techniques to investigate the similarities and dissimilarities between the ‘blue-chip’ stocks used to compute the Dow Jones Industrial Average (DJIA) index.     

Models for stock returns

Historically, the normal variance model has been used to describe stock return distributions. This model is based on taking the conditional stock return distribution to be normal with its variance itself being a random variable. The form of the actual stock return distribution will depend on the distribution for the variance. In practice, the distributions chosen for the variance appear to be very limited. In this note, we derive a comprehensive collection of formulas for the actual stock return distribution, covering some sixteen flexible families. The corresponding estimation procedures are derived by the method of moments and the method of maximum likelihood. We feel that this work could serve as a useful reference and lead to improved modelling with respect to stock market returns.     

A nesting framework for Markov-switching GARCH modelling with an application to the German stock market

In this paper, we establish a generalized two-regime Markov-switching GARCH model which enables us to specify complex (symmetric and asymmetric) GARCH equations that may differ considerably in their functional forms across the two Markov regimes. We show how previously proposed collapsing procedures for the Markov-switching GARCH model can be extended to estimate our general specification by means of classical maximum-likelihood methods. We estimate several variants of the generalized Markov-switching GARCH model using daily excess returns of the German stock market index DAX sampled during the last decade. Our empirical study has two major findings. First, our generalized model outperforms all nested specifications in terms of (a) statistical fit (when model selection is based on likelihood ratio tests) and (b) out-of-sample volatility forecasting performance. Second, we find significant Markov-switching structures in German stock market data, with substantially differing volatility equations across the regimes.     

Intraday volatility and periodicity in the Malaysian stock returns

Many empirical studies using high-frequency intraday data from a variety of markets indicate that PGARCH models give superior return volatility forecasts than those produced from standard GARCH models. This paper investigates into modelling approaches of four versions of PGARCH models of high-frequency data of Bursa Malaysia, in particular where the intraday volatility of double U-shaped pattern. It is examined through half-hourly dummy variables, quarterly-hourly dummy variables, Fourier Functional Form (FFF) based variables and spline-based variables. The non-periodic GARCH models, i.e., GARCH, EGARCH and TARCH are used for comparison of performance of best fit. The analysis show that among the four versions of PGARCH models, the half-dummy and the spline-based versions perform the best. EGARCH produced consistently superior results to other GARCH specifications.     

Alternatives to the normal model of stock returns: Gaussian mixture, generalised logF and generalised hyperbolic models

Simple parametric models of the marginal distribution of stock returns are an essential building block in many areas of applied finance. Even though it is well known that the normal distribution fails to represent most of the “stylised” facts characterising return distributions, it still dominates much of the applied work in finance. Using monthly S&P 500 stock index returns (1871–2005) as well as daily returns (2001–2005), we investigate the viability of three alternative parametric families to represent both the stylised and empirical facts: the generalised hyperbolic distribution, the generalised logF distribution, and finite mixtures of Gaussians. For monthly return data, all three alternatives give reasonable fits for all sub-periods. However, the generalised hyperbolic distribution fails to describe some features of the marginal distributions in some sub-periods. The daily return data are much more symmetric and expose another problem for all three distributions: the parameters describing the behaviour of the tails also influence the scale so that simpler alternatives or restricted parameterisations are called for.      

Changes in the risk structure of stock returns: Consumer Confidence and the dotcom bubble

Changes in the risk structure of stock returns may sometimes be very revealing. We examine economic variables that help explain principal components in UK stock returns, 01/1985 to 12/2001. The loading pattern on explanatory variables for the first component in a ‘bubble’ period is distinctive and consistent with a bubble/crash market. The second component shows a loading pattern on a Consumer Confidence variable in a pre-bubble period only. We observe apparently systematic changes in the structure of risk, and conjecture that Consumer Confidence captures a change in market sentiment that could be a signal for the evolution of stock prices.     

Predicting firm level stock returns: Implications for asset pricing and economic links

We examine the predictive ability of stock price ratios, stock return dispersion and distribution measures for firm level returns. Analysis typically focusses on market level returns, however, for the underlying asset pricing model to hold, firm-level predictability should be present. Additionally, we examine the economic content of predictability by considering whether the predictive coefficient has the theoretically correct sign and whether it is related to future output growth. While stock returns reflect investor expectations regarding future economic conditions, they are often too noisy to act as predictor. We use the time-varying predictive coefficient as it reflects investor confidence in the predictive relation. Results suggest that a subset of stock price ratios have predictive power for individual firm stock returns, exhibit the correct coefficient sign and has predictive power for output growth. Each of these ratios has a measure of fundamentals divided by the stock price and has a positive relation with stock returns and output growth. This implies that as investors expect future economic conditions to improve and earnings and dividends to rise, so expected stock returns will increase. This supports the cash flow channel as the avenue through which stock return predictability arises.     

An application of a random level shifts model to the volatility of Peruvian stock and exchange rate returns

The literature has shown that the volatility of stock and forex rate market returns shows the characteristic of long memory. Another fact that is shown in the literature is that this feature may be spurious and volatility actually consists of a short memory process contaminated with random level shifts (RLS). In this paper, we follow recent econometric approaches estimating an RLS model to the logarithm of the absolute value of stock and forex returns. The model consists of the sum of a short-term memory component and a component of level shifts. The second component is specified as the cumulative sum of a process that is zero with probability ‘1-alpha’ and is a random variable with probability ‘alpha’. The results show that there are level shifts that are rare, but once they are taken into account, the characteristic or property of long memory disappears. Also, the presence of General Autoregressive Conditional Heteroscedasticity (GARCH) effects is eliminated when included or deducted level shifts. An exercise of out-of-sample forecasting shows that the RLS model has better performance than traditional models for modelling long memory such as the models ARFIMA (p,d,q).     

Are there Monday effects in stock returns: A stochastic dominance approach

We provide a test of the Monday effect in daily stock index returns. Unlike previous studies we define the Monday effect based on the stochastic dominance criterion. This is a stronger criterion than those based on comparing means used in previous work and has a well defined economic meaning. We apply our test to a number of stock indexes including US large caps and small caps as well as UK and Japanese indexes. We find strong evidence of a Monday effect in many cases under this stronger criterion. The effect has reversed or weakened in the Dow Jones and S&P 500 indexes post 1987, but is still strong in more broadly based indexes like the NASDAQ, the Russell 2000 and the CRSP.     

A simple two-component model for the distribution of intraday returns

We model the conditional distribution of high-frequency financial returns by means of a two-component quantile regression model. Using three years of 30 minute returns, we show that the conditional distribution depends on past returns and on the time of the day. Two practical applications illustrate the usefulness of the model. First, we provide quantile-based measures of conditional volatility, asymmetry and kurtosis that do not depend on the existence of moments. We find seasonal patterns and time dependencies beyond volatility. Second, we estimate and forecast intraday Value at Risk. The two-component model is able to provide good-risk assessments and to outperform GARCH-based Value at Risk evaluations.     

The behaviour of asset return and volatility spillovers in Turkey: A tale of two crises

This paper examines return and volatility spillovers between the Turkish stock market with international stock, exchange rate and commodity markets. Our aim is not only to examine spillover behaviour with a large emerging market but also to examine cross—asset spillovers and how they vary across two periods of financial market crisis; the dotcom crash and the liquidity-induced financial crisis. This is to be compared with existing work that typically focuses on industrialised countries or single asset markets only. Using the spillover index methodology we uncover an interesting distinction between these two periods of markets stress. Over the dotcom period spillovers are largely between the same asset class, notably two exchange rate series and two international stock markets series. However, in the period including the financial crisis, spillovers both increase and cross asset types and suggest a much greater degree of market interdependence. Understanding this changing nature in spillovers is key for investors, regulators and academics involved in theoretical model development.     

A multi-country test of the Fisher model for stock returns

This article provides a test of the Fisher model, linking expected stock returns and inflation, based on international data. Since the Fisher model is ‘universal’ and calls for a slope of 1 in any country, we improve the testing power by conducting a joint test over eight countries. The pooling of data for several countries seems to reduce the small-sample bias. We test the Fisher model, using an instrumental variable approach, for holding-period horizons ranging from 1–12 months. The Fisher model is not rejected at any horizon: however, the magnitude of the slope coefficient lends stronger support at long horizons. This study using multi-country panel data provides evidence corroborating the finding of Boudoukh and Richardson (1993) that the Fisher model holds at long horizons (5 years), using 180 years of US data.