Comparing Asset Pricing Models

A Bayesian asset pricing test is derived that is easily computed in closed form from the standard F‐statistic. Given a set of candidate traded factors, we develop a related test procedure that permits the computation of model probabilities for the collection of all possible pricing models that are based on subsets of the given factors. We find that the recent models of Hou, Xue, and Zhang (2015a, 2015b) and Fama and French (2015, 2016) are dominated by a variety of models that include a momentum factor, along with value and profitability factors that are updated monthly.     

Costs of Equity Capital and Model Mispricing

Costs of equity for individual firms are estimated in a Bayesian framework using several factor-based pricing models. Substantial prior uncertainty about mispricing often produces an estimated cost of equity close to that obtained with mispricing precluded, even for a stock whose average return departs significantly from the pricing model's prediction. Uncertainty about which pricing model to use is less important, on average, than within-model parameter uncertainty. In the absence of mispricing uncertainty, uncertainty about factor premiums is generally the largest source of overall uncertainty about a firm's cost of equity, although uncertainty about betas is nearly as important.     

Do subjective expectations explain asset pricing puzzles?

The structural uncertainty model with Bayesian learning, advanced by Weitzman (AER 2007), provides a framework for gauging the effect of structural uncertainty on asset prices and risk premiums. This paper provides an operational version of this approach that incorporates realistic priors about consumption growth volatility, while guaranteeing finite asset pricing quantities. In contrast to the extant literature, the resulting asset pricing model with subjective expectations yields well-defined expected utility, finite moment generating function of the predictive distribution of consumption growth, and tractable expressions for equity premium and risk-free return. Our quantitative analysis reveals that explaining the historical equity premium and risk-free return, in the context of subjective expectations, requires implausible levels of structural uncertainty. Furthermore, these implausible prior beliefs result in consumption disaster probabilities that virtually coincide with those implied by more realistic priors. At the same time, the two sets of prior beliefs have diametrically opposite asset pricing implications.     

Weighing asset pricing factors: a least squares model averaging approach

Empirical evidence has demonstrated that certain factors in asset pricing models are more important than others for explaining specific portfolio returns. We propose a technique that evaluates the factors included in popular linear asset pricing models. Our method has the advantage of simultaneously ranking the relative importance of those pricing factors through comparing their model weights. As an empirical verification, we apply our method to portfolios formed following Fama and French [A five-factor asset pricing model. J. Financ. Econ., 2015, 116, 1–22] and demonstrate that models accommodated to our factor rankings do improve their explanatory power in both in-sample and out-of-sample analyses.     

The pricing of volatility risk in the US equity market

We analyze whether the pricing of volatility risk depends on the asset pricing framework applied in the tests, the specified volatility proxies, and the portfolio sorts used for spanning the asset universe. For this purpose, we compare the results using a macroeconomic and fundamental based asset pricing model using three proxies of volatility and uncertainty, using size/value sorted and industry sector portfolios. Our results reveal that the marginal pricing effect of the VIX volatility factor is strong and statistically significant throughout the models and specifications, while the effect of an EGARCH-based volatility factor is mixed, mostly smaller but with the correct sign. In most cases, the EGARCH factor does not impair the pricing effect of the VIX. The portfolio sorts have a substantial impact on the volatility premiums in both model frameworks. The size of the volatility risk premium is more uniform across the models if the industry sector portfolio sort is used. Finally, the size/value portfolio sort generates larger volatility risk premiums for both models.     

Correlation expansions for CDO pricing

This paper develops numerical approximations for pricing collateralized debt obligations (CDOs) and other portfolio credit derivatives in the multifactor Normal Copula model. A key aspect of pricing portfolio credit derivatives is capturing dependence between the defaults of the elements of the portfolio. But, compared with an independent-obligor model, pricing in a model with correlated defaults is more challenging. Our approach strikes a balance by reducing the problem of pricing in a model with correlated defaults to calculations involving only independent defaults. We develop approximations based on power series expansions in a parameter that scales the underlying correlations. These expansions express a CDO tranche price in a multifactor model as a series of prices in independent-obligor models, which are easy to compute. The approach builds on a classical approximation for multivariate Gaussian probabilities; we introduce an alternative representation that greatly reduces the number of terms required to evaluate the coefficients in the expansion. We also apply this method to the underlying problem of computing joint probabilities of multivariate normal random variables for which the correlation matrix has a factor structure.     

How many factors are important in U.K. stock returns?

I use the sequential approach of Harvey and Liu ([2018]. Lucky factors (Working Paper). Duke University) to build linear factor models in U.K. stock returns among a set of 13 candidate factors using individual stocks and three groups of test portfolios between July 1983 and December 2017. My study finds that the Market factor is the dominant factor in reducing mispricing in individual stocks and test portfolios regardless of the pricing error metric used. The Market factor has a bigger impact when using a value weighting pricing error metric. Whether a second factor is used or not depends upon which metric is used for mispricing and the time period examined. My study finds support for a two-factor model for the whole sample period of the Market factor and the Conservative Minus Aggressive (CMA) factor of Fama and French ([2015]. “A five-factor asset pricing model.” Journal of Financial Economics 116: 1–22) when giving greater weight to the mispricing of larger companies.     

Liquidity and asset pricing: Evidence from the Hong Kong stock market

This study investigates the role of liquidity in pricing stock returns in the Hong Kong stock market. Our results show that liquidity is an important factor for pricing returns in Hong Kong after taking well-documented asset pricing factors into consideration. The results are robust to adding portfolio residuals and higher moment factor in the factor models. The results are also robust to seasonality, and conditional-market tests. We also compare alternative factor models and find that the liquidity four-factor model (market excess return, size, book-to-market ratio, and liquidity) is the best model to explain stock returns in the Hong Kong stock market, while the momentum factor is not found to be priced.     

The Pricing of Event Risks with Parameter Uncertainty

Financial instruments whose payoffs are linked to exogenous events, such as the occurrence of a natural catastrophe or an unusual weather pattern depend crucially on actuarial models for determining event (e.g., default) probabilities. In many instances, investors appear to receive premiums far in excess of these modeled actuarial probabilities, even for event risks that are uncorrelated with returns on other financial assets. Some have attributed these larger spreads to uncertainty in the probabilities generated by the models. We provide a simple model of such parameter uncertainty and demonstrate how it affects rational investors' demand for event risk exposures. We show that while parameter uncertainty does indeed affect bond spreads, it does not tend to increase spreads by much. Indeed, the spread increases due to parameter uncertainty in our numerical examples are on the order of only 1–2 basis points. Moreover, in many instances, including those that have the most sensible correlation settings, parameter uncertainty tends to decrease the size of bond spreads. We therefore argue that parameter uncertainty does not appear to be a satisfactory explanation for high event-risk returns.     

Sources of the stock price fluctuations in Chinese equity market

This paper proposes a latent factor approach based on a state–space framework in order to identify which factor, if any, dominates price fluctuations in the Chinese stock markets. We also illustrate the connection of such stock price decomposition with several general equilibrium asset pricing models and show that the decomposition results can potentially offer useful insights with regard to the empirical relevance of asset pricing models. We use quarterly data of the Chinese A-Share equity market over the period 1995Q3–2011Q1 and find that the estimates of the state–space model suggest that the expected return is the primary driving force behind price fluctuations in the Chinese stock market. We show that the time-varying expected returns appear to be counter-cyclical and this result seems to be consistent with the habit formation model of Campbell and Cochrane [1999. By force of habit: A consumption-based explanation of aggregate stock market behavior. Journal of Political Economy 107, no. 2: 205–51.]. However, we also note that there is a great deal of uncertainty with respect to this variance decomposition due to the resulting small signal-to-noise ratio in the estimated state–space model.     

Bayesian Solutions for the Factor Zoo: We Just Ran Two Quadrillion Models

We propose a novel framework for analyzing linear asset pricing models: simple, robust, and applicable to high-dimensional problems. For a (potentially misspecified) stand-alone model, it provides reliable price of risk estimates for both tradable and nontradable factors, and detects those weakly identified. For competing factors and (possibly nonnested) models, the method automatically selects the best specification—if a dominant one exists—or provides a Bayesian model averaging–stochastic discount factor (BMA-SDF), if there is no clear winner. We analyze 2.25 quadrillion models generated by a large set of factors and find that the BMA-SDF outperforms existing models in- and out-of-sample.     

Quantile regression analysis of hedge fund strategies

Extending previous work on hedge fund pricing, this paper introduces the idea of modelling the conditional quantiles of hedge fund returns using a set of risk factors. Quantile regression analysis provides a way of understanding how the relationship between hedge fund returns and risk factors changes across the distribution of conditional returns. We propose a Bayesian approach to model comparison which provides posterior probabilities for different risk factor models that can be used for model averaging. The most relevant risk factors are identified for different quantiles and compared with those obtained for the conditional expectation model. We find differences in factor effects across quantiles of returns, which suggest that the standard conditional mean regression method may not be adequate for uncovering the risk-return characteristics of hedge funds. We explore potential economic impacts of our approach by analysing hedge fund single strategy return series and by constructing style portfolios.     

Implications of state-dependent pricing for dynamic macroeconomic models

State-dependent pricing (SDP) models treat the timing of price changes as a profit-maximizing choice, symmetrically with other decisions of firms. Using quantitative general equilibrium models which incorporate a “generalized (S,s) approach,” we investigate the implications of SDP for topics in two major areas of macroeconomic research, the early 1990s SDP literature and more recent work on persistence mechanisms. First, we show that state-dependent pricing leads to unusual macroeconomic dynamics, which occur because of the timing of price adjustments chosen by firms as in the earlier literature. In particular, we display an example in which output responses peak at about a year, while inflation responses peak at about 2 years after the shock. Second, we examine whether the persistence-enhancing effects of two New Keynesian model features, namely specific factor markets and variable elasticity demand curves, depend importantly on whether pricing is state dependent. In an SDP setting, we provide examples in which specific factor markets perversely work to lower persistence, while variable elasticity demand raises it.     

State uncertainty in stock markets: How big is the impact on the cost of equity?

We propose a novel Bayesian framework to incorporate uncertainty about the state of the market. Among others, one advantage of the framework is the ability to model a large collection of time-varying parameters simultaneously. When we apply the framework to estimate the cost of equity we find economically significant effects of state uncertainty. A state-independent pricing model overestimates the cost of equity by about 4% per annum for a utility firm and by as much as 3% for industries. We also observe that the expected return, volatility, risk loading, and pricing error all display state-dependent dynamics that coincide with the business cycle. More interestingly, the forecasted market and Fama–French factor risk premiums can predict the future real GDP growth rate even though the model does not use any macroeconomic variables, which suggests that the proposed Bayesian framework captures the state-dependent dynamics well.     

Non-linear Gaussian sovereign CDS pricing models

Prior literature indicates that quadratic models and the Black–Karasinski model are very promising for CDS pricing. This paper extends these models and the Black [J. Finance 1995, 50, 1371–1376] model for pricing sovereign CDS’s. For all 10 sovereigns in the sample quadratic models best fit CDS spreads in-sample, and a four factor quadratic model can account for the joint effects on CDS spreads of default risk, default loss risk and liquidity risk with no restriction to factors correlation. Liquidity risk appears to affect sovereign CDS spreads. However, quadratic models tend to over-fit some CDS maturities at the expense of other maturities, while the BK model is particularly immune from this tendency. The Black model seems preferable because its out-of-sample performance in the time series dimension is the best.     

Testing conditional asset pricing models: An emerging market perspective

The CAPM as the benchmark asset pricing model generally performs poorly in both developed and emerging markets. We investigate whether allowing the model parameters to vary improves the performance of the CAPM and the Fama–French model. Conditional asset pricing models scaled by conditioning variables such as Trading Volume and Dividend Yield generally result in small pricing errors. However, a graphical analysis reveals that the predictions of conditional models are generally upward biased. We demonstrate that the bias in prediction may be the consequence of ignoring frequent large variation in asset returns caused by volatile institutional, political and macroeconomic conditions. This is characterised by excess kurtosis. An unconditional Fama–French model augmented with a cubic market factor performs the best among some competing models when local risk factors are employed. Moreover, the conditional models with global risk factors scaled by global conditioning variables perform better than the unconditional models with global risk factors.     

Nonlinear Pricing Kernels, Kurtosis Preference, and Evidence from the Cross Section of Equity Returns

This paper investigates nonlinear pricing kernels in which the risk factor is endogenously determined and preferences restrict the definition of the pricing kernel. These kernels potentially generate the empirical performance of nonlinear and multifactor models, while maintaining empirical power and avoiding ad hoc specifications of factors or functional form. Our test results indicate that preference-restricted nonlinear pricing kernels are both admissible for the cross section of returns and are able to significantly improve upon linear single- and multifactor kernels. Further, the nonlinearities in the pricing kernel drive out the importance of the factors in the linear multi-factor model.     

Consumption risk and the cross-section of government bond returns

In this paper we provide a consumption-based explanation of risk in nominal US Treasury bond portfolios. We use a consumption-CAPM with Epstein–Zin–Weil recursive preferences. Our model introduces two sources of risk: uncertainty about current consumption (reflected in contemporaneous consumption growth) and uncertainty about prospects of consumption in a long run (reflected in innovations to expectations about future consumption growth). We use a novel approach to estimate pricing factors in our model: we employ a factor-augmented VAR model with common factors, extracted from a large panel of macroeconomic and financial data, as state variables. We find that the important source of risk in US bonds is related to uncertainty in prospects in future consumption and it induces a positive and significant risk premium. We find as well that covariance risk related to innovations in expectations about future consumption growth is greater for long term bond portfolios than for short term bond portfolios, which is consistent with a duration measure of risk and justifies why long term bonds require greater premium than short term bonds. Our model explains well the cross-sectional variation in average excess returns of bonds with different maturities over the period 1975–2011 and compares favorably with competing models.     

Modeling the cross-section of stock returns using sensible models in a model pool

An increase in the number of asset pricing models intensifies model uncertainties in asset pricing. While a pure “model selection” (singling out a best model) can result in a loss of useful information, a full “model pooling” may increase the risk of including noisy information. We make a trade-off between the two methods and develop a new two-step trimming-then-pooling method to forecast the joint distributions of asset returns using a large pool of asset pricing models. Our method allows investors to focus on certain regions of the distributions. In the first step, we trim the uninformative models from a pool of candidates, and in the second step, we pool the forecasts of the surviving models. We find that our method significantly enhances portfolio performance and predicts downside risk precisely, and the improvements are mainly due to trimming. The pool of sensible models becomes larger when focusing on extreme events, responds rapidly to rising uncertainty, and reflects the magnitude of factor premiums. These findings provide new insights into asset pricing model evaluation.     

The Stochastic Seasonal Behaviour of Natural Gas Prices

Previous studies have explored the seasonal behaviour of commodity prices as a deterministic factor. This paper goes further by proposing a general (n+2m)‐factor model for the stochastic behaviour of commodity prices, which nests the deterministic seasonal model by Sorensen (2002) . We consider seasonality as a stochastic factor, with n non‐seasonal and m seasonal factors. The non‐seasonal factors are as defined in Schwartz (1997) , Schwartz and Smith (2000) and Cortazar and Schwartz (2003) . The seasonal factors are trigonometric components generated by stochastic processes. The model has been applied to the Henry Hub natural gas futures contracts listed by NYMEX. We find that models allowing for stochastic seasonality outperform standard models with deterministic seasonality. We obtain similar results with other energy commodities. Moreover, we find that stochastic seasonality implies that the volatility of futures returns follows a seasonal pattern. This result has important implications in terms of option pricing.