Travis L. Johnson

The Option to Stock Volume Ratio and Future Returns

We examine the information content of option and equity volumes when trade direction is unobserved. In a multimarket symmetric information model, we show that equity short-sale costs result in a negative relation between relative option volume and future �?rm value. In our empirical tests, �?rms in the lowest decile of the option to stock volume ratio (O/S) outperform the highest decile by 1.47% per month on a risk-adjusted basis. Our model and empirics both indicate that O/S is a stronger signal when short-sale costs are high or option leverage is low. O/S also predicts future �?rm-speci�?c earnings news, consistent with O/S reflecting private information.

Risk Premia and the VIX Term Structure

The shape of the VIX term structure conveys information about the price of variance risk rather than expected changes in the VIX, a rejection of the expectations hypothesis. A single principal component, Slope, summarizes nearly all this information, predicting the excess returns of S&P 500 variance swaps, VIX futures, and S&P 500 straddles for all maturities and to the exclusion of the rest of the term structure. Slopes predictability is incremental to other proxies for the conditional variance risk premia, is economically significant, and can only partially be explained by variations in observable risk measures.

A Simple Multimarket Measure of Information Asymmetry

We develop and implement a new measure of information asymmetry among traders. Our measure is based on the intuition that informed traders are more likely than uninformed traders to generate abnormal volume in options or stock markets. We formalize this intuition theoretically and compute the resulting multimarket information asymmetry measure (MIA) for firm-days as a function of unsigned volume totals and without estimating a structural model. Empirically, MIA has many desirable properties: it is positively correlated with spreads, price impact, and absolute order imbalances; predicts future volatility; is an effective conditioning variable for trading strategies stemming from price pressure; and detects exogenous shocks to information asymmetry. This paper was accepted by Lauren Cohen, finance.

Asymmetric Trading Costs Prior to Earnings Announcements: Implications for Price Discovery and Returns

This study examines the link between earnings announcement premia (i.e., higher returns in announcement periods) and changes in liquidity prior to the announcements. Motivated by prior research, we model market makers as holding positive inventories and show they asymmetrically raise costs of providing liquidity to sellers, relative to buyers, to reduce inventory risks ahead of earnings news. This asymmetry gives rise to the announcement premium by increasing the relative cost of trading on negative news. Consistent with our friction-based hypothesis, we show that equity prices predictably rise in the week prior to announcements and gradually decline following announcements. Our model also yields implications of this friction for trading activity, price dynamics, and the information content of prices, all of which we validate in our empirical tests. JEL Classifications: G10, G11, G12, G14, M41 ∗We thank SP Kothari and seminar participants at Cornell University and MIT for helpful feedback and suggestions. Corresponding authors: Travis Johnson, travis.johnson@mccombs.utexas.edu, 2110 Speedway Stop B6600, Austin, TX 78712 and Eric So, ESo@mit.edu, E62-677 100 Main Street, Cambridge MA 02142. Earnings Announcement Premia: The Role of Asymmetric Liquidity Provision 1

Distortions Caused by Asset Managers Retaining Securities Lending Income

Using newly-mandated disclosures, we show fund managers often retain a fraction of securities lending income by employing in-house lending agents at above-market rates. This retention incentivizes fund managers to overweight stocks with high lending fees. In a heterogeneous agent model, we show this incentive distorts equilibrium portfolio choices, fund performance, and asset pricing. We confirm our model’s predictions empirically: fee-retaining active mutual funds overweight high lending fee stocks, underperform, and charge lower management fees. Our model also offers a new explanation for the negative relation between lending fees and future fee-inclusive returns. ∗We thank Andres Almazan, Aydoğan Altı, John Hatfield, John Griffin, Garrett Schaller, Clemens Sialm, Laura Starks, Sheridan Titman, and seminar participants at The University of Texas at Austin for their helpful comments, and Tim Park, David Xu and Qifei Zhu for excellent research assistance. Send correspondences to travis.johnson@mccombs.utexas.edu, 2110 Speedway Stop B6600 Austin TX 78712. State Street is fighting two separate US lawsuits over claims it . . . took an “unreasonably large” 50 per cent share of the net income generated from lending securities owned by the trust, amounting to “fiduciary self-dealing and enrichment in violation of the defendants” fiduciary obligations. – Financial Times article “State Street battles two US lawsuits,” 2/10/2013 BlackRock, the world’s largest asset manager, has systematically “looted” securities lending revenues from investors, according to a lawsuit filed by two US pension funds. The suit alleges a number of BlackRock’s US-listed iShares exchange traded funds operated a “grossly excessive” fee model that allowed an affiliate . . . to retain 40 per cent of the revenue generated by lending stock. – Financial Times article “US pension funds sue BlackRock,” 2/3/2013 Mutual funds and exchange traded funds (ETFs) often lend assets to short sellers, generating a total of just over

A Fresh Look at Return Predictability Using a More Efficient Estimator

1bn in gross lending fee revenue in 2017, equivalent to 7.7% of total management fees. A fraction of lending fee revenues goes to lending agents, rebates to borrowers, and other costs rather than back to investors. As illustrated by the above examples, these costs have proven controversial as some investors argue they are excessive due to ‘self-dealing’ whereby the fund manager uses an affiliated lending agent and pays them above-market rates. Until recently, such claims were difficult to systematically evaluate because fund managers did not disclose the gross amount of lending revenues they generate, instead only being required to disclose the net amount returned to investors. In 2016, though, the SEC amended Regulation S-X to require mutual fund managers disclose gross lending revenues, a cost breakdown, and the identify of their lending agent starting with fiscal year 2017. Using the newly-mandated disclosures, we show that self-dealing is common. For all mutual funds and ETFs reporting lending income in 2017, only 76% of gross lending revenues were returned to investors in 2017, with the other 24% going to costs. Furthermore, many large asset management companies used affiliated lending agents, collecting

Weighted Least Squares Estimates of Return Predictability Regressions

30mm of lending agent fees, an average of 8.4% of gross lending revenue for mutual funds using an affiliated

Time Will Tell: Information in the Timing of Scheduled Earnings News

Time varying volatility causes substantial heteroskedasticity in return predictability regressions, making OLS estimates less efficient than least squares estimates weighted by ex-ante return variance (WLS-EV). In small sample simulations, I show that using WLS-EV instead of OLS results in large efficiency gains, fewer false negatives, and avoids the bias associated with ex-post weighting schemes. Using WLS-EV also changes several important conclusions based on OLS estimates: traditional predictors such as the dividend-to-price ratio perform better in- and out-of-sample, whereas WLS-EV estimates of the predictability afforded by proxies for the variance risk premium, politics, the weather, and the stars are not significant.I assess time-series return predictability using a weighted least squares estimator that is around 25% more efficient than ordinary least squares (OLS) because it incorporates time-varying volatility into its point estimates. Traditional predictors, such as the dividend yield, perform better in- and out-of-sample when using my estimator, indicating the insignificant OLS estimates may be false negatives driven by a lack of power. Some newer predictors, such as the variance risk premium and the president’s political party, are insignificant when using my estimator, indicating the significant OLS estimates may be false positives driven by a few periods with high expected volatility. Received March 31, 2018; editorial decision September 26, 2018 by Editor Jeffrey Pontiff. Authors have furnished an Internet Appendix and supplementary data and code, which are available on the Oxford University Press Web site next to the link to the final published paper online.

Asymmetric Trading Costs Prior to Earnings Announcements: Implications for Price Discovery and Returns: ASYMMETRIC TRADING COSTS PRIOR TO EARNINGS ANNOUNCEMENTS

Time varying volatility causes substantial heteroskedasticity in return predictability regressions, making OLS estimates less efficient than least squares estimates weighted by ex-ante return variance (WLS-EV). In small sample simulations, I show that using WLS-EV instead of OLS results in large efficiency gains, fewer false negatives, and avoids the bias associated with ex-post weighting schemes. Using WLS-EV also changes several important conclusions based on OLS estimates: traditional predictors such as the dividend-to-price ratio perform better in- and out-of-sample, whereas WLS-EV estimates of the predictability afforded by proxies for the variance risk premium, politics, the weather, and the stars are not significant.I assess time-series return predictability using a weighted least squares estimator that is around 25% more efficient than ordinary least squares (OLS) because it incorporates time-varying volatility into its point estimates. Traditional predictors, such as the dividend yield, perform better in- and out-of-sample when using my estimator, indicating the insignificant OLS estimates may be false negatives driven by a lack of power. Some newer predictors, such as the variance risk premium and the president’s political party, are insignificant when using my estimator, indicating the significant OLS estimates may be false positives driven by a few periods with high expected volatility. Received March 31, 2018; editorial decision September 26, 2018 by Editor Jeffrey Pontiff. Authors have furnished an Internet Appendix and supplementary data and code, which are available on the Oxford University Press Web site next to the link to the final published paper online.