Who Leaves, Where To, And Why Worry?: Employee Mobility, Employee Entrepreneurship, And Effects On Source Firm Performance Benjamin A. Campbell Ohio State University Martin Ganco University of Illinois April M. Franco University of Toronto Rajshree Agarwal University of Illinois 2009 Joint U.S.-Canadian Census Research Data Center Conference October 5, 2009
The smallprint… The research in this paper was conducted while Ben Campbell and Martin Ganco had Special Sworn Status as researchers of the U.S. Census Bureau at the Chicago Census Research Data Center. Research results and conclusions expressed are those of the authors and do not necessarily reflect the views of the Census Bureau. This research has been screened to insure that no confidential data are revealed.
Motivation • Where does productivity come from? • Recent research points to the importance of human capital and human assets • Entrepreneurs’ previous work experience helps to determine the success of their ventures • But several questions are yet unanswered: • Which individuals are most likely to leave a firm? • Which individuals are most likely to go to a spin-out? • What is the impact of such movement on the incumbent firm?
Klepper and Sleeper, 2004 Franco and Filson, 2006 Agarwal, Ganco and Ziedonis, 2008 Agarwal, Echambadi, Franco, and Sarkar, 2004 Elfenbein et al., 2008 Klepper and Thompson, 2009 Somaya et al., 2007 Wezel et al., 2006 Phillips, 2002 Groysberg et al., 2007 Employee mobility in knowledge intensive industries High-Technology Manufacturing Human Capital Intensive Services
Firm organization and employee mobility in the services sector • Garicano and Hubbard, 2007 • Levin and Tadelis, 2005 • Rebitzer and Taylor, 2007
Main Ingredients • Bargaining Power • Relative importance of complementary assets to production • Ability to transfer/recreate complementary assets • Create New Opportunities • Ability to reconfigure complementary assets to be more productive
Who Leaves, where do they go, and does it matter?? • H1: There is a negative relationship between earnings and the likelihood of employee mobility. • H2: Conditional on mobility, individuals with greater earnings are more likely to join spin-outs than join established firms. • H3: The adverse impact on firm performance due to employee mobility is greater for employee mobility to spin-outs than for employee mobility to incumbents. • H4: The adverse impact on parent firm performance due to employee mobility to spin-outs relative to employee mobility to established firms increases with the earnings of the moving individual.
Context • We test the hypotheses in the legal services industry: • Human capital is easily transferrable (within state borders). • Overhead costs are low, and wage bill is the dominant cost, hence the aggregate wage bill is a good proxy for revenues (Gilson and Mnookin, 1985). • Data • A custom extract of the Longitudinal Employer-Household Dynamics (LEHD) Project available at the Census Research Data Centers. • The data are longitudinal spanning over 10 years and covering 10 large states. • The custom extract includes all workers who have ever worked in the legal services industry and all firms that have ever reported operating in the legal services industry.
Key Variables • Mobility to Spin-out/Mobility to Incumbent • longitudinal records of employment history allow us to track employee mobility and employee entrepreneurship. • Employee Earnings • Earnings include all forms of taxable compensation that are received in the calendar year, including salary, bonuses and other reported income. • Firm Performance: By summing the earnings of the universe of employees inside the firm, we re-construct the total revenues earned by the firm. We divide by employment to capture the revenue generated per employee (in $10,000s). • Firm Mobility Measures. We measure mobility in three ways: • Number of employees leaving • Aggregate Payroll of employees leaving • Number of employees leaving in different payroll classes • $0-$100K • $100K-$300K • $300K-$5000K Individual Firm
Universe • For the individual-level analysis: • a random 25% sample of the data • individuals who earn more than $25K annually • Individuals who are employed at a firm of more than 5 people • Individuals that are employed at a firm that does not exit the data that year or in the subsequent year • For the firm-level analysis: • firms with more than 5 employees • firms do not exit in the current or subsequent year • firms that have revenue per worker of between $10K and $1M • firms that do not lose greater than 20 workers in any payroll class to an established form or to a startup in a given year
Empirical Strategy • Stage 1 (Who Leaves? To Go Where?) • First, identify the individual characteristics that are related to employee mobility in general. • Second, identify the individual characteristics that are related to mobility to spin-out conditional on mobility. • We estimate a series of linear probability model with firm-year fixed effects and robust standard errors. • Conditional logit is computationally infeasible with our sample size. • Out-of-sample predictions are extremely rare indicating that the linear probably model performs acceptably. • Stage 2 (And Does it Matter?) • Firm performance is a function of the intensity of different types of worker mobility and firm characteristics. • We estimate a series of linear regression models with firm fixed effects and year dummies.
Who leaves, to go where? ? H1: supported H2: supported
And Does it Matter?Mobility to spin-out vs. Mobility to incumbent H3: supported The average 85-employee firm faces a $22,865 loss when an employee moves to a spin-out!
And Does it Matter?Value appropriation and mobility to spin-out If the employee earns between $100,000 and $300,000 and moves to a spin-out, the firm faces a $193,000 loss in revenues! H4: supported If the employee earns between $300,000 and $1,000,000 and moves to a spin-out, the firm faces a $1,000,000 loss in revenues!
Summary • Workers with higher earnings are less likely to move (H1) • If workers with higher earnings do move, they are more likely to go to startups (H2). • Mobility to spin-out has a larger adverse impact on parent performance than mobility to incumbent (H3). • The adverse impact on firm performance of mobility to spin-out increases with earnings of the mover (H4).
Contributions • We study the implications of knowledge transfer mechanisms where knowledge is a rival good • We provide a direct comparison between mobility to incumbent firms and mobility to spin-outs, both at the individual and at the parent firm level
Follow-on Projects • Working on: • Impact of Teams • Does it matter if a superstar leaves with a team or as an individual? • Does a superstar hurt the firm more if she goes to a start-up? • Compensation practices and spin-out/incumbent performance