(with Hannes Wagner and Stefan Zeume)
at The Review of Financial Studies
We exploit one of the largest data leaks to date to study whether and how firms use secret offshore vehicles. From the leaked data, we identify 338 listed firms as users of secret offshore vehicles and document that these vehicles are used to finance corruption, avoid taxes, and expropriate shareholders. Overall, the leak erased $174 billion in market capitalization among implicated firms. Following the increased transparency brought about by the leak, implicated firms experience lower sales from perceptively corrupt countries and avoid less tax. We estimate conservatively that
one in seven firms have offshore secrets.
Components of balance sheet asset growth which are related to earnings management contributed to the asset growth anomaly in the past. These components of balance sheet asset growth are no longer related to returns since 2002 and this has contributed to the disappearance of the asset growth anomaly. I provide evidence that the Sarbanes-Oxley Act reduced earnings management and improved the integrity of accounting information: earnings manipulation has decreased, earnings predictability has increased, and analyst forecast errors have decreased. Further, the cross-sectional relationship between the accrual accounts used to manage earnings and analyst optimism has reduced. The evidence suggests that the asset growth anomaly was driven by mispricing in the past and that this mispricing has decreased. More broadly, these findings point towards changes in the regulatory environment as a novel driver of equity market anomalies.
Affiliated Stocks: Business Groups, Control, and Stock Returns
(undergoing major revision)
(with Massimo Massa and Hong Zhang)
We study how firm affiliation to a business group affects stock prices. We argue that the more important a firm is to the ultimate owner to retain the control of the group (“central”) the less sensitive it is to informational and cash-flow shocks. The market prices this higher stability and lower informational sensitivity by requiring a lower return on the stock. We test these predictions using a novel dataset of worldwide ownership of publicly listed firms for the 2001-2013 period for which we have information not only on firm characteristics and their prices, but also full ownership structure. We develop a measure of the degree to which firms are used by the ultimate owner to retain control of the group (“centrality”). We document that central firms display lower sensitivity to information and cash flow shocks. This limits the informational advantage of informed investors, reducing short-selling, increasing liquidity, and lowering idiosyncratic volatility. These effects are priced as central firms command lower stock returns (gross and factor-risk-adjusted). High-centrality portfolios deliver a 75 (72) basis points lower three-factor (four-factor) alpha than the low centrality ones. A centrality factor helps to explain the cross-section of stock returns in business groups. Centrality is most appreciated at the times in which market uncertainty is highest.
Work in Progress
Equilibrium Computation with Heterogeneous Firms: An Alternative Approach
The algorithm of Krusell and Smith (1998) adapted to heterogeneous firms by Khan and Thomas (2008) is often used to find the equilibrium in a market with a representative agent and heterogeneous firms. Recent applications in asset pricing include Gomes and Schmid (2016) and Chen (2017). A key feature of this algorithm is the inner-loop outer-loop structure: the firm optimization problem is solved given a forecasting rule for aggregate states and the forecast rule is updated given simulations from the optimal firm policy until a fixed point is found where the forecast rule does not change and the policy is optimal given the forecast rule. I propose an alternative algorithm that uses a histogram based approximation of the firm distribution of Young (2010) as well as the first order conditions of the agent’s problem to solve market clearing and optimal firm policies simultaneously. An advantage of this approach is that each solution point can be considered an equilibrium realization conditional on the learning rule. This is similar to the idea of an “Experience Based Equilibrium” in Fershtman and Pakes (2012) where forecast rules are determined as a function of the data the agent could have learned through optimal play. The algorithm can be used to understand off equilibrium dynamics and the impact of learning rules that bound rationality in the sense of the number of state variables or the number of past periods included in learning of decision rules.
Asset Pricing with Heterogeneous Firms and Entry: An Investigation
Many asset pricing anomalies are found to be larger in small, fast-growing firms. Furthermore, default risk is also related to anomaly strength e.g. Avramov, Chordia, Jostova, and Phllipov (2013). I investigate the ability of a heterogeneous firm model with equity issuance, risky short-term debt, and entry to rationalize these relationships. The key risk mechanism relies on fixed costs of issuance and decreasing returns to scale which is magnified by a leverage effect in bad times when firms are closer to default. I consider the general question of what anomalies can be rationalized in this framework, and what additional assumptions are needed to jointly match key anomalies.
Merger Financing and the Information Content of Option-Implied Moments
(with Cal Muckley and Conall O’ Sullivan)
We show that the options market plays an important role in price discovery prior to corporate merger and acquisition events. Changes in the option implied model-free skewness have the greatest predictive capacity for principally cash financed target firms. A rationale for this pre-announcement option trading is (i) that the primarily cash financed target firms exhibit the largest discontinuous price impact at announcement and (ii) this finance mechanism, unlike in primarily equity financed deals, does not facilitate equity hedging of deal risk with an equity position in the counterpart deal firm. We formalize this intuition by simulating a simple model to show the information assimilation in changes in implied moments with respect to announcement returns.