CMU Portugal Program: In the abstract of your dissertation, you say that one of the results of your research shows “a negligible effect of voluntary quits on the parent firm, suggesting that job sorting among workers and firms occurs without much ado for the firm." What method and data did you used that led to this result?
Cristina Carias (CC): Our results are strengthened by the fact, that, for the first time, we used a country wide data-set (Quadros de Pessoal) aggregating firms operating in different industries, and econometric methods tailored to the intricate problem of estimating the effects of employee quits on the employer. Econometrically, this problem is a challenge, as there are multiple confounders that may bias the results. We used several econometric techniques, including propensity scores, well-known methods to take into account fixed effects, and the usual ordinary least squares regression.
CMU Portugal Program: What is the impact of your findings?
CC: Entrepreneurs, business, and policy makers will find our results highly relevant. For policy makers and business, our results suggest non-compete agreements (or work clauses employees need to sign curbing their ability to work for competitors after they leave their current employer) may be of questionable importance. Policy makers may want to consider that, given the importance of new hires when starting successful ventures, and the relevance that these ventures have for the economy, curbing the employee’s ability to work for other firms after they leave their current employer may actually be detrimental to society. For entrepreneurs, we estimated that founders that hired people they knew well, and that started their company in a region and industry they had experience on were more successful in the long run.
CMU Portugal Program: You started your doctoral studies in 2007. In looking back, can you identify which were the most challenging stages and greatest learning experiences?
CC: A doctoral degree is always challenging the student’s limits. First, you learn the advanced techniques in your field, and then you need to learn how to use them to solve your problem. The greatest challenge for me was to find a problem that was both interesting and solvable with the techniques and data set I had available. The greatest learning experience was characterizing problems so they became tractable, and then adapting techniques to find the solution. This was the most important thing I learned from my advisors: how to frame the problem in terms of a research question, and then answer it in a way that was insightful beyond the specific question itself.
CMU Portugal Program: You have worked at Portugal Telecom before starting your Ph.D.. Did the experience of working at a company gave you another perspective of your research?
CC: Having worked on the private sector (in Portugal Telecom, the largest telecommunication provider in Portugal) before joining the Ph.D. program was extremely useful in that it allowed me to understand the organizational challenges faced by large firms, and kept my work more grounded in reality.
CMU Portugal Program: After finishing your Ph.D., what opportunities came up?
CC: I am assigned at the Centers for Disease Control and Prevention in Atlanta (to the Office of the Director of the National Center for Immunizations and Respiratory Diseases). In my current work, I evaluate vaccination programs, and quantify problems of interest to decision makers in emergency responses (which are set up to respond to outbreaks caused by a previously unknown pathogen such as new flu strains, or the Middle Eastern Respiratory Syndrome coronavirus; or large outbreaks such as the current Ebola outbreak in West Africa). I use a range of techniques and concepts that I previously studied, but my current position also requires me to adapt new techniques very quickly. At the same time, I had the opportunity to be a visiting Professor in Emory University (Atlanta), and to teach econometrics to undergraduates.
“Worker exit, New Firm Formation, and the Effects of Quits on the Parent Firm”
We analyze the literature on startups, regional networks, and employee turnover to i) understand which employees are more likely to exit and to where; ii) analyze how new employees use knowledge about previous hires to find a new firm; and iii) describe four different ways in which employee voluntary quits affect the parent firm: increase in strategic definition, alliance formation, increase in competition and adjustment costs.
We hypothesize and estimate that managers are more likely to exit the firm to find a new firm than other employees, and that specialized workers, following exit, are more likely to join incumbents in the same industry. We then theorize and estimate that startups entering the same industry of the founder’s previous employer are more likely to be located near the founder’s previous employer, more likely to employ former colleagues of the entrepreneur, and more likely to performing better than other startups.
We finally estimate the effects of voluntary quits on the parent firm. While voluntary quits are not observed, we define a voluntary quit as an exit followed by immediate entry in the dataset and salary raise or entry in entrepreneurship. We use several estimation methodologies, including General Method of Moments and regressing on the Propensity Score, to take into account firm heterogeneity, unobserved effects, and endogeneity bias. We further complement the analyses with regressions tailored to firms of different sizes, ages and ability managers. Our results show a negligible effect of voluntary quits on the parent firm, suggesting that job sorting among workers and firms occurs without much ado for the firm.
Committee: Lee Branstetter (CMU; Chair), Rui Baptista
(IST-UL), Serguey Braguinsky (CMU) and Brian Kovak (CMU).