Job Market Paper
Switching costs that arise in repeated purchases of durable goods cause buyers to face conflicting incentives: facilitating competition among potential sellers leads to lower prices while restricting competition among them allows buyers to avoid the disruption associated with introducing a new brand. I study this trade-off in an auction environment with bid preference that allows buyers to favor certain sellers. I construct novel data on fleet renewal by municipal bus operators in Poland who use a common format of scoring auctions to implement bid preference. Consistent with their incentive of avoiding switching costs, the operators favor incumbent bus producers. Motivated by this finding, I develop and estimate a structural model of public procurement with bidder favoritism to quantify the main driving forces of the trade-off. Estimates suggest that bid preference programs can balance the trade-off if an auction attracts sufficiently many bidders, whereas forcibly promoting competition while ignoring the underlying lock-in relationship between buyers and incumbent sellers would lead to counter-productive outcomes. Therefore, the design of public procurement should not only target achieving low prices but also account for other aspects contributing to buyers' welfare.
with Anusha Nath and Scott Dallman, working paper
Education services in the United States are determined predominantly by non-market institutions, the rules of which are defined by state constitutions. This paper empirically examines the effect of changes in constitutional provisions on education outcomes in the United States. To show causal effects, we exploit discontinuities in the procedure for adopting constitutional amendments to compare outcomes when an amendment passed with those when an amendment failed. Our results show that the adoption of an amendment results in higher per-pupil expenditure, higher teacher salaries, smaller class size, and improvements in reading and math test scores. We examine the underlying mechanism driving these results by studying the actions of the legislature and the courts after an amendment is passed. We find that, on average, the legislature responds with a one-year lag in enacting education policies satisfying the minimum standards imposed by the amendment, and there is no increase in the number of education cases reaching appellate courts. Using school finance reforms, we also show that in situations where the legislature fails to enact education policies, courts intervene to enforce constitutional standards to improve outcomes. This enforcement mechanism is more impactful in states that have higher constitutional minimum standards. Taken together, the causal effects on education outcomes and the patterns in legislative bill enactments and court cases provide a novel test of the hypothesis that a strong constitutional provision improves the bargaining position of citizens vis-à-vis that of elected leaders. If citizens do not receive education services as mandated in the constitution, they can seek remedy in court.
Estimating effects of large scale government programs: new ML approach and application to labor supply reaction to child benefits
under major revision, project supported by Polish National Science Centre grant #2016/N/HS4/03637
This paper studies the evaluation of large-scale government programs in which violations of parallel trend assumptions invalidate the use of standard econometric methods. Instead of relying on a potential outcomes framework, I obtain counterfactual outcomes from a general model of discrete choice. The model implicitly accommodates individuals’ per-period payoffs as well as expectations regarding future outcomes and is estimated non-parametrically using recent advances in machine learning. The estimation procedure requires conditioning on a potentially large set of observed variables which is typically available to researchers in recent years. I apply the method to the introduction of a universal child benefit program in Poland and show that it led to a drop of 2–4 percentage points in the labor force among eligible females. The effect results mainly from changes in women’s perceived trade-offs and beliefs that discouraged labor market participation.
with Ewa Stanisławska, published in Eastern European Economics, 55.1 (2017): 3-28
This article explores the relation between consumers’ inflation expectations and their spending and saving behavior, proxied by their buying and saving attitudes. The analysis, using a large micro dataset, suggests that inflation expectations negatively affect saving attitudes, especially in the group of consumers characterized by a very good financial situation. Moreover, the role of inflation expectations has increased since the global financial crisis. The results for buying attitudes are somewhat puzzling—especially if one tries to interpret them together with the results for saving attitude—as they also suggest a negative, although very weak, link to inflation expectations.