We estimate latent factor models of liquidity, aggregated across various liquidity measures. Shocks to assets’ liquidity have a common component across measures which accounts for most of the explained variation in individual liquidity measures. We find that across-measure systematic liquidity is a priced factor while within-measure systematic liquidity does not exhibit additional pricing information. Controlling for across-measure systematic liquidity risk, there is some evidence that liquidity, as a characteristic of assets, is priced in the cross-section. Our results are robust to the inclusion of other equity characteristics and risk factors, such as market capitalization, book-to-market, and momentum.

We use an asymptotic principal components technique to estimate the pervasive factors influencing asset returns and to test the restrictions imposed by static and intertemporal equilibrium versions of the arbitrage pricing theory (APT) on a multivariate regression model. The empirical techniques allow for fairly arbitrary time variation in risk premiums. We find that the APT provides a better description of the expected returns on assets than the capital asset pricing model (CAPM). However, some statistically reliable mispricing of assets by the APT remains.

This paper develops a theory and econometric method of portfolio performance measurement using a competitive equilibrium version of the Arbitrage Pricing Theory. We show that the Jensen coefficient and the appraisal ratio of Treynor and Black are theoretically compatible with the Arbitrage Pricing Theory. We construct estimators for the two performance measures using a new principal components technique, and describe their asymptotic distributions. The estimators are computationally feasible using a large number of securities. We also suggest a new approach to testing for the correct number of factors.

We study market-making high-frequency trader (HFT) dynamics around large institutional trades in Canadian equities markets using order-level data with masked trader identification. Following a regulatory change that negatively affected HFT order activity, we find that bid-ask spreads increased and price impact decreased for institutional trades. The decrease in price impact is strongest for informed institutional traders. During institutional trade executions, HFTs submit more same-direction orders and increase their inventory mean reversion rates. Our evidence indicates that high-frequency trading is associated with lower transaction costs for small, uninformed trades and higher transaction costs for large, informed trades.Received May 24, 2016; editorial decision June 21, 2018 by Editor Andrew Karolyi.

With time-varying adverse selection in the market for new equity issues, firms will prefer to issue equity when the market is most informed about the quality of the firm. This implies that equity issues tend to follow credible information releases. In addition, if the asymmetry in information increases over time between information releases, the price drop at the announcement of an equity issue should increase in the time since the last information release. Using earnings releases as a proxy for informative events, we find evidence supporting these propositions.

This article develops an intertemporal, discrete-time, competitive equilibrium version of the arbitrage pricing theory (APT) and explores the econometric implications of this model under various restrictions on investor preferences and on the dynamic behavior of dividends. We describe conditions under which the econometric techniques typically used for estimating and testing the APT can be shown to be consistent with our economic model. We relate our intertemporal version of the APT to the static APT and to Merton’s intertemporal capital asset pricing model.

Two-pass cross-sectional regression (TPCSR) is frequently used in estimating factor risk premia. Recent papers argue that the common practice of grouping assets into portfolios to reduce the errors-in-variables (EIV) problem leads to loss of efficiency and masks potential deviations from asset pricing models. One solution that allows the use of individual assets while overcoming the EIV problem is iterated TPCSR (ITPCSR). ITPCSR converges to a fixed point regardless of the initial factors chosen. ITPCSR is intimately linked to the asymptotic principal components (APC) method of estimating factors since the ITPCSR estimates are the APC estimates, up to a rotation.

We test whether momentum strategies remain profitable after considering market frictions induced by trading. Intraday data are used to estimate alternative measures of proportional and non‐proportional (price impact) trading costs. The price impact models imply that abnormal returns to portfolio strategies decline with portfolio size. We calculate break‐even fund sizes that lead to zero abnormal returns. In addition to equal‐ and value‐weighted momentum strategies, we derive a liquidity‐weighted strategy designed to reduce the cost of trades. Equal‐weighted strategies perform the best before trading costs and the worst after trading costs. Liquidity‐weighted and hybrid liquidity/value‐weighted strategies have the largest break‐even fund sizes: $5 billion or more (relative to December 1999 market capitalization) may be invested in these momentum strategies before the apparent profit opportunities vanish.

An important issue in applications of multifactor models of asset returns is the appropriate number of factors. Most extant tests for the number of factors are valid only for strict factor models, in which diversifiable returns are uncorrelated across assets. In this paper we develop a test statistic to determine the number of factors in an approximate factor model of asset returns, which does not require that diversifiable components of returns be uncorrelated across assets. We find evidence for one to six pervasive factors in the cross-section of New York Stock Exchange and American Stock Exchange stock returns.

An important issue in applications of multifactor models of asset returns is the appropriate number of factors. Most extant tests for the number of factors are valid only for strict factor models, in which diversifiable returns are uncorrelated across assets. In this paper we develop a test statistic to determine the number of factors in an approximate factor model of asset returns, which does not require that diversifiable components of returns be uncorrelated across assets. We find evidence for one to six pervasive factors in the cross‐section of New York Stock Exchange and American Stock Exchange stock returns.