Research

Published and Forthcoming Articles (Selected)

2025

1. Published

   How Digital Platforms Affect Local Entrepreneurial Activities: Evidence from the Staggered Entry of Craigslist

   Daehyun Kim, Yongwook Paik, and Namil Kim

   Strategic Entrepreneurship Journal (FT50), 2025

   Abstract Journal

   Access to markets is critical for entrepreneurial success, yet many face significant barriers. This study argues that digital platforms functioning as peer-to-peer marketplaces can facilitate entrepreneurial activity by improving market access for new ventures. Using the staggered entry of Craigslist into various markets, we find that Craigslist’s entry increased local entrepreneurial activities, especially in low-income regions. Moreover, individuals from lower-income households, as well as Black and immigrant communities, were more likely to create new ventures following Craigslist’s entry. Firms founded after Craigslist’s entry were also smaller in size, potentially reflecting more efficient management, and primarily operated in B2C rather than B2B sectors. This study highlights how multisided platforms can empower individuals with limited market accessand resources, ultimately democratizing entrepreneurship in the digital age.

2. Published

   Anti-Labor Environments and Employee Entrepreneurship: Evidence from Right-To-Work Laws

   Daehyun Kim, Namil Kim, and Haemin Dennis Park

   Strategic Entrepreneurship Journal (FT50), 2025

   Abstract Journal SMS

   We explore how changes in labor union and related labor environments affect employees’ likelihood of starting a new business. Using Michigan’s and Indiana’s adoption of right-to-work (RTW) laws as a natural experiment, we demonstrate that the likelihood of employees becoming self-employed increased by 54.6% as compared to that in states without RTW laws following the passage of the law. However, this tendency is more pronounced for less-educated and blue-collar workers. Moreover, newly created-firms are mostly small businesses with limited scalability. These findings offer novel insights on the relationship between anti-labor environments and employee entrepreneurship.

3. Published

   When Automation Hits Jobs: Entrepreneurship as an Alternative Career Path

   Daehyun Kim, Taekyun Kim, Wonjoon Kim, and Hyejin Youn

   PLOS One, 2025

   Abstract Journal

   This study investigates the relationship between occupational automation risks and workers’ transitions to entrepreneurship using data from the Current Population Survey. We find that employees facing automation-related job displacement are inclined to shift toward unincorporated entrepreneurship, emphasizing entrepreneurship as a viable alternative career path. Noteworthy variations emerge when examining specific automation technologies, revealing a positive association between industrial robots and entrepreneurial transitions, whereas artificial intelligence displays a negative relationship. Gender disparities are observed, with female workers exhibiting a lower likelihood than males of transitioning into entrepreneurship. This study also shows a heightened prominence of entrepreneurial transitions during the early stages of the COVID-19 pandemic. By illuminating entrepreneurship as a response to job displacement, our results offer crucial policy insights into the labor market implications of automation.

Manuscripts Under Review

current

1. 3rd R&R

   How Does Publicly Available Knowledge Affect Interfirm Hiring?

   Shu Deng, Haemin Dennis Park, and Daehyun Kim

   3^rd Revise and Resubmit (minor), Journal of Management Studies (FT50)

   Abstract SSRN

   Increased availability of public knowledge may diminish the necessity of hiring inventors for codified knowledge, but it may heighten the demand for inventors with tacit knowledge to utilize public knowledge. By using the enactment of the American Inventor’s Protection Act (AIPA) as a shock that increased publicly available knowledge through mandated patent disclosure, we find that firms with greater exposure to AIPA hire fewer inventors who possess less complex knowledge while hiring more inventors who possess highly complex knowledge. Further analyses reveal that both the substitution and complementation effects are less pronounced for firms with highly novel technological portfolios, whereas they are more pronounced for firms with greater technological capability.

2. 2nd R&R

   Unpacking the Entrepreneurial Process after Failure: The Role of Attribution and Behavioral Responses in Technology-Based Startups

   Daehyun Kim, Haejun Jung, Aaron Chatterji, Wonjoon Kim, and Sunghyun Cho

   2^nd Revise and Resubmit, Strategic Entrepreneurship Journal (FT50)

   • Award: Amore-Pacific Scholarship from MK-AKMS, 2021

   Abstract

   This study examines why entrepreneurs perceive the causes of failure differently and how these attributions relate to post-failure processes. Using a unique survey of failed entrepreneurs, we find that those with prior external financing are more likely to attribute failure to internal factors, since receiving such critical resources for startup success makes external causes less plausible. We further show that such internal attribution is negatively associated with reentry. Next, among serial entrepreneurs who do reenter, attribution is related to strategic change: internal attribution is associated with revising the business model, whereas technology change occurs mainly when shortcomings are perceived as the cause. Finally, entrepreneurs who revised their business model or technology after failure were more likely to secure venture capital through such internal changes.

3. Rej & R

   Technological Uniqueness and Venture Capital Investment

   Daehyun Kim, Haemin Dennis Park, and Shu Deng

   Reject and Resubmit, Journal of Business Venturing (FT50)

   Abstract SSRN

   We examine how venture capital (VC) investors evaluate the technological uniqueness of startups and the conditions under which this evaluation varies. Using a text-based approach that maps startups in the technology space, we find an inverted U-shaped relationship between technological uniqueness and the likelihood of receiving first-round VC investment, indicating that startups with moderate uniqueness are most likely to attract capital. This pattern varies with the degree of startup–VC alignment. When an investor’s existing portfolio overlaps with the startup’s technological domain or industry, the evaluation penalty at higher levels of uniqueness is attenuated, shifting the peak of the inverted U toward greater uniqueness. These findings suggest that prior exposure to related technologies or industries helps VC investors interpret unfamiliar technologies and reduce the uncertainty that otherwise discourages investment at the upper end of uniqueness. Our study contributes to research on optimal distinctiveness and entrepreneurial resource acquisition by drawing attention to investor-side characteristics that shape how technological uniqueness is evaluated. It also offers a text-based approach for measuring startups’ technological positioning that future work on innovation and early-stage investment can apply.

4. Under Review

   When Employee-Friendly Institutions Challenge Human-Capital Continuity in Startups: Evidence from Paid Family Leave Policies and Venture Capital Investment

   Daehyun Kim, Yongwook Paik, and Shu Deng

   Under Review, Organization Science (UTD24, FT50)

   • Award: Best Paper Award, Korean Academic Society of Business Administration (KASBA), 2025
   • Media Cover: Maeil Business Newspaper

   Abstract SSRN

   Employee-friendly institutions can expand worker protections, but their organizational consequences for young firms remain less understood. We examine this issue in the context of U.S. state-level paid family leave (PFL) policies, asking how PFL implementation affects startups’ access to venture capital (VC) investment. We argue that, by enabling employees to take paid caregiving leave, PFL can reduce startups’ continuous access to critical employees, weakening the execution and innovation signals that VC investors use to assess investment readiness. Leveraging the sequential implementation of PFL policies in New Jersey and Rhode Island, we use a staggered difference-in-differences design with coarsened exact matching to study VC-backed, growth-oriented startups. We find that PFL implementation reduces both startups’ likelihood of receiving VC investment and the amount they receive. The decline is stronger among startups with a higher pre-PFL female employee share, consistent with greater exposure to likely PFL uptake, and weaker among startups with greater pre-PFL occupational concentration, consistent with more internal substitutability. Mechanism analyses show increased female employee departures and reduced quality-adjusted innovation output. Additional analyses indicate that the effect is concentrated among smaller startups and early-stage investment deals. Supplementary robustness checks provide consistent support for our theoretical framework linking PFL implementation to reduced VC investment through disruptions to human-capital continuity that weaken execution and innovation signals. Together, these findings develop human-capital continuity as a boundary condition for understanding employee-friendly institutions and show how worker-protective policies can shape entrepreneurial resource acquisition indirectly by altering organizational signals used in external evaluation.

5. Under Review

   Entrepreneurial Training and Startup Internationalization: Evidence from the German Accelerator Program

   David Heller, Daehyun Kim, and Dietmar Harhoff

   Under Review, Journal of Business Venturing (FT50)

   • Award: European Investment Fund Best Paper Award on Highest Policy Impact, 2025
   • Note: CEPR Discussion Paper No. 20888

   Abstract CEPR SSRN

   International expansion represents an important growth strategy for new ventures, but entering foreign markets exposes startups to unfamiliar institutional and market environments. We examine whether and how entrepreneurial training programs can help startups overcome these challenges. To answer these questions, we use unique data from the German Accelerator (GA), a government-funded program designed to support German startups entering foreign markets. We find that startups participating in the GA secure more funding from local investors, greater local employment, and more trademark filings in the target market, while we find no corresponding effects on domestic market activities. These effects are particularly strong when the training provides experiential learning opportunities through structured learning, in-person mentorship, and sustained immersion in the target market. We link our results to the literature on organizational learning and confirm them with qualitative evidence from a survey of GA alumni. Our findings extend research on entrepreneurial training into international contexts, showing that these programs can facilitate foreign-market expansion with experiential learning being a key mechanism behind these effects.

Manuscripts Under Development

Correct authoring will be determined later

current

1. In Progress

   Cheaper to Build, Harder to Stand Out: Generative AI and Startup Similarity

   Daehyun Kim

   Preparing manuscript

   Abstract

   A growing share of new firms now build both their products and their organizations around large language models. Recent work shows that AI-oriented startups have proliferated since the release of ChatGPT and that the teams founding them have grown smaller. What this evidence establishes is that generative AI has lowered the cost of entry. Yet what it leaves open is which firms walk through the door. We ask whether the technology that lowers the cost of entry also narrows the range of ideas entrepreneurs pursue. Our sample covers startups founded between 2018 and 2025, drawn from PitchBook. For each venture we recover the historical homepage from the Wayback Machine and embed the text with SBERT, which lets us measure similarity between firms using how they describe themselves. We estimate a difference-in-differences model that compares firms founded in industries highly exposed to generative AI with firms founded in less exposed industries, before and after the release of ChatGPT, where exposure is built from the occupational composition of each industry using Revelio data. We find that firms entering high-exposure industries after ChatGPT resemble one another substantially more than firms entering low-exposure industries. The convergence is not only lexical. Removing AI-related terminology attenuates the estimate but leaves it significant, which suggests that ventures are converging in what they propose to do rather than only in the words they use. Resembling one’s peers also carries a price: the more similar a venture is to others founded around it, the less likely it is to raise a first round of venture capital. Generative AI has made it cheaper to start a company, but the companies it makes cheap to start are increasingly the same company.

2. In Progress

   Untangling Technology Concepts: An Integrative Review and Multidimensional Framework

   Carolyn Crichton, Daehyun Kim, Dietmar Harhoff, and Marc Gruber

   Preparing manuscript

   Abstract

   Technological change is reshaping organizational landscapes and opening new opportunities for the emergence and scaling of tech-based ventures. Yet research on technology and technology entrepreneurship remains highly fragmented. Concepts and associated labels such as high-tech, low-tech, hard-tech, soft-tech, essential-tech, emerging-tech, and deeptech are widely used, but they have often been developed separately and are sometimes combined without clear specification of their domains, attributes, boundaries, or relationships. This lack of comparability limits theoretical and empirical progress. This paper develops an integrative review of technology labels across academic and practitioner literatures. Drawing on a family-resemblance approach, it examines what each concept captures, where concepts overlap or diverge, and how they can be organized into a multidimensional framework for classifying technologies and technology ventures. Rather than treating these concepts as fixed or mutually exclusive categories, the framework conceptualizes technologies and technology ventures as profiles across four dimensions: temporal development, complexity, form, and human necessity. This approach clarifies how different concepts highlight distinct aspects of technology ventures and how those aspects shape venturing processes, resource needs, ecosystem support, and inspire future research opportunities. By improving conceptual precision, the paper provides a more accessible and cumulative foundation for scholarship on technology entrepreneurship, established organizations, and ecosystems.

3. In Progress

   Entrepreneurial Growth with Generative Artificial Intelligence

   Daehyun Kim, David Heller, and Asam Dominik

   Data Analysis
4. In Progress

   The Impact of Generative AI on Founding Team Diversity

   Daehyun Kim, David Heller, and Jae Cho

   Data Analysis