NATALIE MILLAR

Job Market Paper

"The Labor Market Returns to Customized Job Training"

Abstract: Economic theory dating back to Becker (1962) predicts that employers should pay for training workers in specific, non-transferable skills while governments should fund more general training. Customized job training (CJT) programs, which exist in many U.S. states, defy this logic by using public subsidies to teach workers a range of skills, including those that are firm-specific. Are governments naive or mistakenly subsidizing training that firms would pay for on their own, or does CJT generate benefits that justify public investment? I answer these questions using unique hand-collected data from Tennessee on firms' applications and trainees' enrollment linked with rich administrative education and earnings data. I exploit quasi-random rationing among equally propitious firms and equally eligible prospective trainees. My results indicate that CJT generates persistent earnings increases of 3%, comparable to the returns from an additional year of work experience. To explain these findings, I classify the skills taught in each program by mapping program descriptions in firms' applications to O*NET. Despite the fact that CJT programs include training in many transferable skills across industries, occupations, and more generally, CJT yields benefits to workers and the government through higher earnings that more than offset the cost of training to both firms and the government, yielding an exceptionally high marginal value of public funds relative to other job training programs.

Working Papers

''Business Incentives and Economic Activity: Evidence from Alabama''

with Rebecca Lester and Joshua Rauh

Abstract: Business incentives are a central policy tool of state and local officials to attract and retain firms. However, prior academic work suggests that these incentives come at a large fiscal cost, with approximately 40% of state corporate tax revenues and 70% of local economic development funds spent on these incentives. Earlier studies on the effectiveness and overall impact of incentives on job creation produce mixed results, potentially due to the reliance on incomplete data. In partnership with state and local authorities in Alabama, we assembled a dataset of state and local business incentives for the period 2012 through 2024. Of the 3,500 incentives granted during this period, only approximately 3% were previously observed in academic research. In total, these incentives equal $6.2 billion, with an average incentive value of $4.1 million. Using administrative employment data at the firm-county level, we find that receiving a business incentive increases employment at recipient firms by an average of 61 jobs as compared to non-recipient firms. This effect increases to 247 additional jobs for firms receiving the largest incentives, whereas we find no employment effect for the bottom 30th percentile of incentives. We estimate an average cost per job of $77,000, which is equal to 2.5 years of average per-capita income in Alabama and is lower than estimates from prior work. Our findings provide new insights into the prevalence, characteristics, and economic impacts of business incentives, contributing to the literature on the effectiveness of incentives as an economic development policy tool.  

Media Coverage: Office of the Governor, State of Alabama, Yellowhammer

Policy: Researchers: Transparency a Key Step for Incentives, Innovative Alabama: A Report by the Hoover Institution

Selected Works in Progress (Data Usage Agreements Signed)

 ''The Impact of Prisoner Education on Recidivism, Labor Market Outcomes, and Public Assistance Utilization'' 

with Valentin Bolotnyy and Michael Lovenheim - Data analysis stage

Abstract: This paper examines the impact of career technical education (CTE) prison education on recidivism, labor market, and public assistance outcomes. The motivation for this paper is that recidivism rates are very high, likely due to weak labor market outcomes for former inmates. Prior research has focused on high school level basic education which has low returns generally and for this population. CTE programs have the potential to strongly support labor market success of formerly incarcerated people by training them for specific occupations that are in-demand. We worked with Tennessee to create a novel longitudinal data set that links together administrative records on education, corrections, labor market outcomes, and public assistance. Prior research designs generally do not allow for a causal interpretation of the underlying estimates. We use variation in program offerings across facilities and variation across inmates in access to these programs to estimate the causal returns to prison CTE training. Our paper contributes to the literature by providing evidence on the causal effect of CTE on recidivism, labor market outcomes, and social service use, which we will then use to conduct cost-benefit analyses of these programs.

Grants: Awarded W.E. Upjohn Institute for Employment Sponsored Grant

''Are Firms' International Capital Markets Efficient? Evidence from Regional Government Subsidies'' 

with Lisa De Simone, Rebecca Lester, and Aneesh Raghunandan - Data analysis stage

Approved U.S. Census Bureau project

Abstract: This study investigates whether firms efficiently allocate capital across internal divisions when faced with exogenous, location-specific cash flow shocks. We use state and local government subsidies—cash grants and tax exemptions—as plausibly exogenous shocks to divisional cash flows for publicly traded U.S. firms. Merging detailed subsidy data from Good Jobs First with establishment-level Census microdata, we examine how firms reallocate employment, investment, and production activity across jurisdictions following receipt of a subsidy. We distinguish between cash grants, which provide immediate and fungible liquidity, and tax exemptions, which lower local operating costs, to test whether firms expand activity in the subsidized region or redirect funds to higher-return projects elsewhere. Our approach yields large-sample, firm-region-year evidence on the efficiency of internal capital markets—whether capital flows toward jurisdictions with greater growth opportunities. Complementary analyses benchmark internal efficiency using capital expenditure deviations and Tobin’s q-based measures. Beyond contributing to research on corporate finance and public economics, the project enhances Census Bureau data quality by reconciling capital expenditure definitions across establishment- and firm-level surveys. The findings inform both theories of internal capital allocation and the economic incidence of regional subsidy policies.

''The Impact of Online Training Certificates on Educational and Labor Market Outcomes'' 

with Michael Lovenheim and Joshua Rauh - Data merger stage

Abstract: This paper evaluates the educational and labor market impacts of non-traditional online certificates, focusing on Google Career Certificates offered through Coursera in Texas. Despite rapid growth in the online credential market, little is known about who earns these certificates, how they interact with traditional education, and whether they improve labor market outcomes. We construct the first dataset linking administrative education and workforce records from the Texas Education Agency, the Texas Higher Education Coordinating Board, and the Texas Workforce Commission to individual-level certificate data from Google and Coursera. Using individual fixed effects models, we estimate the causal effect of earning a Google certificate on educational attainment and earnings, comparing post-certificate outcomes to pre-certificate outcomes for the same individuals. We examine heterogeneity across certificate types, demographic groups, and institutional contexts where certificates are embedded in traditional education programs. The findings will provide novel evidence on whether online certificates act as complements or substitutes to traditional education, their relative costs and returns, and the extent to which they accelerate entry into the labor market. 

Grants: Texas 2036 Sponsored Research Grant

Selected Works in Progress (Data Usage Agreements in progress)

''Can Education and Training Grants Meet Workforce Needs? Evidence from the Idaho LAUNCH Program'' 

with Valentin Bolotnyy and Nick Gebbia - Data collection stage

Abstract: Since the passing of the federal Workforce Innovation and Opportunity Act (WIOA) in 2014, federal and state funding incentivizing student enrollment in in-demand job fields has proliferated. In this project, we study the effects of one such program: Idaho’s LAUNCH scholarship, which offers up to $8,000 to each student enrolled in an eligible postsecondary education program connected to an in-demand job field. We use administrative data to study the effects of LAUNCH awards for students who receive an award in a DID-IV framework, which uses the interaction of post-LAUNCH with baseline propensity for take-up to instrument for receiving a LAUNCH award. We study the effects of LAUNCH on two primary sets of outcomes: (1) education outcomes, including years of postsecondary education, degree attainment, and field of study; and (2) labor market outcomes, including employment, income, and field of work.

''Documenting and Decomposing State Returns to Higher Education''

with Jacob Light - Data collection stage

Abstract: States are the primary funders of U.S. public universities, an investment they recoup through the increased productivity, earnings, and social outcomes of college graduates. Graduates who attend public universities but leave the state benefit from a subsidized education without repaying the subsidy, potentially incentivizing states to invest in ways that diverge from the levels that maximize national welfare. This project studies how institutions differ in their contributions to both earnings and graduate retention to measure the return on investment (ROI) to the state. Using administrative data linking enrollment at all Texas public colleges and universities to earnings for individuals who remain in-state, we estimate institution-level value added to earnings and residency using a selection-on-unobservables approach following Dale and Krueger (2002). Combining these estimates with data on state appropriations, we identify which institutions yield the highest fiscal returns. In a descriptive extension, we decompose these returns to isolate the contributions of field of study, institutional resources, and degree completion.

Grants: Arnold Ventures Foundation

''Doubly Robust Instrumental Variable Estimators''

with Jun Zhao - In preparation

Teaching 

Instructor of Record

Intermediate Microeconomics (Fall 2021, Spring 2022)

Principles of Microeconomics (Summer 2021)

Teaching Assistant

PhD - Econometrics (Fall 2020)

MBA - Policy Evaluation (Spring 2020)

MBA - Managerial Economics (Spring 2021)

Undergraduate - Honors Intermediate Microeconomics (Fall 2019, Fall 2020)

Undergraduate - Intermediate Macroeconomics (Fall 2019)

Undergraduate - Development Economics (Spring 2020)

Undergraduate - Legal Environment of Business (Fall 2018, Spring 2019)