Published or Forthcoming Papers
We investigate the strict exogeneity assumption, a necessary condition for estimator consistency in many finance panel settings. We outline tests for strict exogeneity in both traditional (non-IV) and IV settings. When we apply these tests in common traditional finance panel regressions, we find that the strict exogeneity assumption is often rejected, suggesting large inference errors. We test for strict exogeneity in specific finance IV panel settings and illustrate the potential for these tests to help confirm, or rule out, the validity of common panel IV estimators. We offer a set of recommendations to address the strict exogeneity issue in finance research.
♦ Lead Article in JFQA
♦Presented at the 2015 London Business School Symposium on Causal Inference, the 2016 Financial Research Association conference in Las Vegas, Iowa University, and the University of Nebraska
Corporate Investment and Innovation in the Presence of Competitor Constraints, Review of Financial Studies (2019), with Zack Liu
We study the relation between investment behavior and competitor financial constraints. Using inter-firm patent citations and text-based product market similarities to identify intransitive competitor networks, we find that firms increase investment spending, patenting activity, and employee poaching when competitor constraints become more binding. In addition, firms shift their investment composition (product market and patent portfolios) towards competitors who experience a relative tightening in constraints. These effects are robust to controlling for selection and correlated effects across competitors. To mitigate endogeneity concerns, we exploit the 2004 AJCA tax holiday and the 1989 junk bond crisis as exogenous shocks to competitor constraints and find similar effects.
♦ Presented at the 2017 American Finance Association meetings in Chicago, Texas Christian University, the University of Texas at Austin, Michigan State University, and the 2015 Midwest Finance Association
Rational and behavioral asset pricing theories offer conflicting interpretations of the covariance structure of asset returns. Return comovement beyond what prespecified empirical factor models can explain is often interpreted in favor of frictions or behavioral explanations. However, we show that randomly grouped assets exhibit “excess” comovement that is ubiquitous and indistinguishable from the comovement of economically motivated groupings advanced in the literature. Our finding is consistent with the presence of a latent factor that could be derived from multiple sources of systematic variation, including rational sources. We propose new statistical tests that account for latent factors when detecting excess comovement.
♦Scheduled for Presentation at the 2020 AFA in San Diego, also Presented at Emory University, Michigan State University, Rice University, Southern Methodist University, Texas Christian University, Tulane University, the University of Arizona, the University of Houston, the University of New Orleans, and the University of Texas at Dallas
Firm clustering is positively correlated with productivity, and it exhibits significant cross-sectional variation across industries. Thus, it is important to understand the industry characteristics that drive firms’ decisions to co-locate. We develop a model of knowledge sharing and derive the prediction that riskier and more complex industries experience the greatest gains from knowledge spillovers. Using tests that account for the nonrandomness of location decisions, we find a strong positive relationship between industry risk or complexity and the clustering of: 1) firms’ headquarters, 2) patent inventors, and 3) R&D expenses. Customer–supplier proximity is also significantly and positively related to industry risk and complexity.
♦Presented at the 2019 Urban Economics Association Conference in Philadelphia, the 2016 Regional Science Association International Conference, Michigan State University, and Tulane University
Product Differentiation, Benchmarking, and Corporate Fraud, with Audra Boone, Rachel Li, and Parth Venkat
We find that firms with less product market differentiation exhibit significantly lower rates of fraudulent activity. This relationship is more pronounced for complex firms and is robust to controlling for various measures of competition and industry heterogeneity. Exploiting IPOs by product-market rivals as a shock to a firm’s information environment, we find that greater publicly available information from comparable firms facilitates the detection of fraud. Moreover, this effect is stronger for firms with ex ante less similar public rivals. These findings suggest that greater product market overlap with rivals disciplines firms by providing benchmarks for auditors, regulators, and investors.
♦Best Paper Award at the New Zealand Finance Meeting (2018). Also presented at Alabama, Clemson University, Drexel University, Michigan State University, the Sixth Annual Conference on Financial Market Regulation, Southern Methodist University, the University of Nevada Las Vegas, the 2018 Financial Management Association (best paper semi-finalist), the SEC, and the Australasian Finance and Banking Conference
Work in Progress
We adapt methods from spatial econometrics to explore the industrial scope of regional network effects in corporate decisions. These methods circumvent well-known challenges in estimating and interpreting empirical models of externalities. We find that cumulative regional network effects are roughly 4-5 times as important as own-firm effects for explaining corporate investment decisions, but only 1.2 and 1.5 times as important for financing decisions and firm performance, respectively. To explore whether these effects operate primarily within or across industry boundaries, we implement a novel structured network regression that allows for a (convex) combination of networks based on product market and supply chain relationships, in addition to regional proximity. We find that regional effects operate primarily within industry boundaries, and that outside of industry regional effects are only present (yet highly attenuated) for corporate investment decisions. These findings lend support to the urban specialization view of agglomeration economies proposed by Marshall (1920).
This paper uses the introduction of Transportation Networking Companies (TNC), such as Uber, to show that access to flexible sources of income facilitates entrepreneurship. TNCs give prospective entrepreneurs the ability to earn a substantive wage, which provides the financial assurance to start their new businesses. Because hours are set independently, entrepreneurs also maintain the temporal flexibility to work the unpredictable hours that characterize new ventures. We employ a difference-in-differences approach that exploits variation in the timing and location of TNC introductions and find evidence that the introduction of TNCs increases young firm employment and proprietorship. We also present survey-based evidence consistent with these findings; TNCs have provided artists, musicians, engineers and other entrepreneurs the financial and temporal flexibility to pursue their passions.