How the Study Knows: Measures, Matching, and Counterfactuals

The specific problem here is how to decide whether the Brazilian judges' post-leave output recovery reflects a real pattern or merely a convenient artifact of measurement and comparison. The study's answer has two parts. First, it builds a measurement system in which employment, official work products, case-level outcomes, and career records test different claims. Second, it compares judges who become parents at different times in a way that creates a plausible counterfactual without pretending that fertility timing was random. Understanding those choices lets the reader judge both the strength and the limits of the productivity result.

A causal estimate is only as useful as the path from an event to an observable outcome. For this question, that path is unusually demanding: the researcher must know when parenthood occurred, observe the same person before and after it, measure work rather than merely job status, and check whether the worker's tasks or opportunities changed. The Brazilian judicial setting helps with each step, but it does not remove uncertainty. It makes the question more answerable; it does not make the answer automatic.

Measurement is part of the argument

The first advantage comes from the work itself. Judges perform cognitively intensive tasks that generate recorded decisions. The number of sentences issued each month is the judiciary's dominant objective performance measure in practice and is related to the number of cases closed. It is also tied to career incentives: productivity helps determine some merit-based promotions, even though judges receive stable salaries rather than being paid per case. That makes sentence counts relevant to both public-service output and the worker's professional incentives.

But the count is not the same as total productivity. A sentence count may miss preparation, legal reasoning that is not visible in a simple total, management of clerks, and other work that does not end in a recorded sentence. The study therefore uses additional official measures—settlement agreements ratified and hearings held—to see whether the main pattern extends beyond one task. Later, case-level measures add timeliness, appeals, reversals, workload characteristics, and ruling behavior. Each measure answers a different question, and the combination is more informative than an undifferentiated claim that “performance” stayed the same.

The second advantage is the way work is allocated. Within jurisdictions, cases are assigned across court units of the same type, and when multiple judges work in one unit, cases are typically divided mechanically. This limits the possibility that judges or supervisors systematically select easy cases for a parent or remove difficult cases from them. The rule does not mean that every case is equally demanding or that individual output is fixed. Judges differ in how much they produce, and the system has room for variation. It means that a post-birth change in output is less easily explained by a visible change in the kind of docket the judge receives.

The third advantage is retention. Strong employment protection and the profession's near-universal return after childbirth allow the study to follow the same judges across the relevant states.