Published and Accepted Work

  1. Sharma, S., Frake, J., and Watson, J. Symbolic vs. Substantive Support: The Impact of Black Lives Matter on Black-Owned Businesses. Marketing Science.

    This study examines the impact of the Black Lives Matter (BLM) movement on consumer behavior toward Black-owned businesses, using George Floyd’s murder as a shock to the salience and support for BLM. Utilizing a difference-in-differences design with data on reviews, revenues, and foot traffic from Yelp and SafeGraph, we analyze changes in symbolic (review frequency and valence) and substantive (revenues and foot traffic) support. Our findings show a significant increase in symbolic actions for Black-owned businesses following Floyd’s murder, especially amongst White and Democrat-leaning consumers. However, we observe no substantial evidence for a corresponding increase in substantive support, indicating that while the BLM movement heightened visibility and prompted short-term symbolic gestures, it did not seem to translate into large, lasting economic benefits for Black-owned businesses. Additionally, we observe an increase in "Not Recommended" reviews on Yelp following Floyd’s murder, suggesting that one explanation is that consumers are leaving inauthentic reviews without visiting the stores. These results highlight the complexities of sociopolitical consumerism and suggest that social movements like BLM, while important, may be insufficient on their own to address the economic disparities faced by minority-owned businesses.


  2. Frake, J., Byun, H., and Kim, J. (Forthcoming). The Effect of Financial Resources on Misconduct: Evidence from Lottery Ticket Sales. Organization Science.
    Download paper

    We investigate the influence of financial resources on a firm's propensity for misconduct. Previous studies offer conflicting predictions regarding the relationship, and much of the empirical evidence suffers from issues like selection, measurement error, reverse causality, and omitted variable bias. Leveraging a difference-in-differences design, we first examine quasi-random fluctuations in retailers' financial resources resulting from large windfalls from selling winning lottery tickets. Our results suggest that an increase in financial resources from selling a large winning lottery ticket reduces retailers’ tobacco sales to minors. Next, to rule in strain theory and to rule out alternative explanations, we leverage a second natural experiment, heterogeneity analysis, an alternate measure of misconduct, and an online randomized experiment. In doing so, we provide plausibly causal evidence on the relationship between an organization’s financial resources and its propensity for misconduct and provide potentially useful findings for policy makers and regulators.


  3. Frake, J., Hagemann, A., and Uribe, J. (2024). Collider Bias in Strategy and Management Research: An Illustration Using Women CEOs Effect on Other Women’s’ Career Outcomes. Strategic Management Journal.
    Download paper

    Collider bias can cause spurious correlations when researchers condition on a variable that is caused by—or shares a common cause with—both the outcome and the exposure variable. Despite its threat to inference in many settings, we document that papers published in top strategy and management journals discuss collider bias at far lower rates than those in economics and sociology journals. We distinguish colliders from other threats to identification and estimation and illustrate its importance with replications of published work suggesting that having a woman CEO reduces the career outcomes (compensation and representation) of other women executives. After accounting for collider bias, we find no evidence that women CEOs damage the career outcomes of other women in their organizations. We close by providing generalizable approaches to identify and mitigate the risk of collider bias in applied research.


  4. Hurst, R., Lee, S., and Frake, J. (2024). The Effect of Flatter Hierarchy on Applicant Pool Gender Diversity: Evidence from Experiments. Strategic Management Journal.
    Download paper

    This paper investigates how job seekers' perceptions of an employer's formal hierarchy affect the size and gender composition of its applicant pool. Building on the literature on gendered organizations and organizational design, we develop opposing perspectives on these relationships. To arbitrate between these perspectives, we first conduct a field experiment in partnership with a hiring firm. We find that featuring a flatter hierarchy in recruiting materials does not significantly affect the size of the applicant pool, but significantly decreases women's representation within it. Our follow-up survey experiment identifies several potential mechanisms (e.g., perceptions of career progression, informality, workload, and fit). Our findings imply that firms' growing tendency to adopt flatter hierarchies could inadvertently undermine efforts to attract a greater proportion of women applicants.


  5. Frake, J. and Harmin, D. (2023). Intergenerational Transmission of Organizational Misconduct: Evidence from the Chicago Police Department. Management Science.
    Download paper

    This paper investigates how organizational misconduct is perpetuated through intergenerational transmission. We theorize that early exposure to a subculture of misconduct imprints newcomers with the belief that misconduct is normal, which is then carried by these individuals into managerial positions and passed down to their subordinates. We test this using longitudinal administrative data from the Chicago Police Department from 1980 to 2017. We exploit a random lottery that assigns applicants to training cohorts to demonstrate that officers exposed early on to a subculture of misconduct not only engage in more misconduct over their entire careers, but also increase the misconduct of their subordinates after they become managers. We also find that this intergenerational dynamic is stronger when subordinate officers were exposed to a subculture of misconduct themselves, are earlier in their tenure, and have not yet received their annual review from their manager. Taken together, these findings reveal a bottom-up dynamic whereby beliefs about misconduct are developed in an organization’s lowest ranks, carried by these individuals over time, and passed down to future generations. This study expands our understanding of how organizational misconduct is perpetuated as well as offers important policy implications for addressing the problem of police misconduct.


  6. Starr, E., Frake, J., and Agarwal, R. (2019). Mobility Constraint Externalities. Organization Science, 30(5): 869-1123.
    Download paper

    Covenants not to compete are often included in employment agreements between firms and employees, justified by each party’s voluntary “freedom to contract.” However, noncompetes may also generate externalities for all individuals in the market, including those who have not signed such agreements. We theorize that enforceable noncompetes increase frictions in the labor market by increasing uncertainty and recruitment costs, and by curtailing entrepreneurship. We find that in state-industry combinations with a higher incidence and enforceability of noncompetes, the unconstrained receive relatively fewer job offers, have reduced mobility, lower wages, and are less satisfied with their jobs. The results offer policymakers a reason to restrict noncompetes beyond axiomatic appeals to a worker’s “freedom of contract” and highlight labor market frictions that may impact firm-level human capital strategies.


  7. Byun, H., Frake, J., and Agarwal, R. (2018). Leveraging Who You Know by What You Know: Specialists, Generalists, and Returns to Relational Capital. Strategic Management Journal, 39:1803–1833.
    Download Paper

    This paper investigates the interaction effects of specialization and relational capital on performance. We distinguish between upstream and downstream relational capital and theorize that higher levels of specialization will buffer against decreases in upstream relational capital, because of deeper domain expertise and stronger downstream relational capital. Conversely, higher levels of generalization permit greater gains from increases in upstream relational capital, due to leverage across a more diversified downstream portfolio of activities. We test and find support for these hypotheses in the context of the US lobbying industry. Our study contributes to the strategic human capital literature by isolating the dimension of specialization and relational capital embodied within individuals and providing performance implications of the interactions.


  8. Frake, J. (2017). Selling Out: The Inauthenticity Discount in the Craft Beer Industry. Management Science, 63(11): 3930-3943.
    Download paper

    This paper investigates why audiences devalue organizations that behave inauthentically. One explanation is that inauthenticity leads to lower perceptions of product quality. This stems from the audience's doubt of an inauthentic actor's capability and commitment to producing high-quality goods. Another explanation is that audiences discount the symbolic value — or what the good represents — of goods from inauthentic actors. I empirically test each of these mechanisms in the craft beer industry. First, I exploit exogenous variation in consumers' knowledge of craft brewers' inauthentic identity (whether they are owned by a corporate brewer) to empirically demonstrate an inauthenticity discount. Next, I decompose audience evaluations to show that knowledge of a producer's inauthenticity does not have a statistically significant impact on evaluators' sensory experience of the product, but that it does affect audience evaluations of the product's symbolic value.


Working Papers

  1. Frake, J., Gibbs, A., Goldfarb, B., Hiraiwa, T., Starr, E., and Yamaguchi, S. From Perfect to Practical: Partial Identification Methods for Causal Inference in Strategic Management Research. Minor Revision at Strategic Management Journal.
    Download paper

    The use of difference-in-differences (DD) and instrumental variables (IV) designs to identify causal effects has grown in strategy and management research. However, these methods rely on untestable identifying assumptions to interpret the results as causal. When authors, reviewers, or editors suspect that these assumptions might be violated, they may conclude that the approach has a “fatal flaw” or that the results are “not identified” and can no longer be interpreted as causal. We argue, however, that potential violations of the identifying assumption need not invalidate a causal interpretation of the results. Instead, we highlight that “partial identification” techniques allow researchers to draw causal inferences from imperfect identification strategies by quantifying how severe of a violation is needed to meaningfully change the direction of the results. We explain how these tools work in the context of DD and IV designs, provide practical guidance in deploying them, and illustrate their use in an empirical example that investigates how patenting affects inventor mobility. In doing so, we advocate against a binary view of causal inference and emphasize the role of theory, context, and judgment when deciding how strongly to infer causality from an empirical result.


  2. Frake, J. Hurst, R., and Kagan, M. Office Parties: Partisan Sorting in the United States Labor Market. Under Review.
    Download paper

    Despite a growing literature on geographic sorting along political lines, there exists no large-scale estimate of the degree to which Americans are sorted by political partisanship when at work. This is despite the large and important role the workplace plays in cultivating cross-partisan contact. We address this gap using an novel dataset we create by merging voter registration data with over 16 million online employee profiles covering 14 million unique workers. We present three main findings. (1) Partisans are sorted by workplace. We estimate that the average Republican worker's co-workers are 12% more Republican than the average Democrat's co-workers (and vice versa). Accounting for factors correlated with partisanship—geography, occupation, and industry—reduces this estimate to 4 percentage points, which is similar in magnitude to our estimates of workplace sorting by gender and race. (2) Since 2017, political sorting has increased among new joiners to firms. (3) Because Democrats comprise a greater share of the workforce, sorting means that Republicans experience a significantly larger share of out-partisan coworkers. We outline avenues for future research regarding the origins and consequences of workplace partisan sorting.


  3. Kagen, M., Frake, J., and Hurst, R. Measuring Workplace Partisanship.

    A growing literature considers the role of workers' personal politics in driving a range of key organizational outcomes. To date, research has measured employees' political beliefs using public records of political donations. In this paper, we introduce an alternative approach (VRscores) which measure political affiliation by merging voter registrations with online worker profiles. We compare our novel measure with prior measures based upon campaign finance donations and demonstrate several empirical improvements. First, voter registrations capture a much larger share of the workforce, as very few employees donate. Second, the voter registration measure captures a more representative sample, as employee donors are demographically unrepresentative of the overall employee population. Third, the larger sample makes it possible to measure employee political partisanship at a broader set of firms, as many firms have few if any political donors. In addition to empirically comparing the coverage, we also draw upon recent literature in political science to make a number of theoretical arguments for why voter registration is a more conceptually valid and reliable measure of political partisanship for rank-and-file employees. We provide recommendations for applied researchers interested in studying workers' personal politics on what measures may be appropriate for different research questions. We conclude with a short demonstration of how applied researchers may use VRscores.


  4. Gordon, S., Frake, J., and Harmon, D. Beyond Deterrence: Unintended Consequences of Punishment in the Chicago Police Department.

    This paper explores the impact of punishment on peer misconduct. While previous research implies that punishment serves as a deterrent, we propose that it may increase the misconduct of the punished individuals’ peers. Moving beyond a purely rational view of punishment, we consider circumstances where emotional reactions may outweigh rational ones. Specifically, we argue that when an individual’s guilt is uncertain, their peers will perceive them to be less guilty of wrongdoing than would independent third parties. Consequently, peers may perceive any punishment—even objectively “just” punishments—as unjust treatment by the organization, increasing the likelihood that they react emotionally and engage in misconduct in response. We test this argument using longitudinal data from the Chicago Police Department and a random lottery used to assign police-academy peers. Our findings demonstrate that punishment causes an increase in the subsequent misconduct of peers. Consistent with our theory that such reactions are emotional, we find that the effects are short-lived and become stronger as the severity of the punishment increases. Taken together, this study indicates that prior understandings of punishment as an effective deterrent are incomplete. By provoking an emotional reaction in peers, organizations may unintentionally encourage misconduct when punishing wrongdoers.