The First Pattern: Value Concentrates at the Employment Margin

The Census threshold evidence, the broader within-score comparison, and the unresolved signal-versus-experience question now fit together into a conditional pattern. A temporary public job appears most valuable when it changes the path of someone who might otherwise not work, but the evidence does not say that every marginal worker should be placed into any available public role. The practical judgment is narrower: evaluate the worker's counterfactual, the credibility of the job, and the costs of delivering it before treating a large employment estimate as a general policy promise.

The value of a job depends on the path it replaces

The sharpest contrast in the U.S. study is not between public and private employment. It is between applicants with different outside options. For Census applicants near the score-70 threshold, hiring raised the probability of a positive W-2 by about 61 percentage points in 2010, and the estimated difference was still about 28 points in 2023. For a broader within-score population, the peak 2010 increase was about 22 points and dissipated by 2011. The same temporary public employer reached both groups, yet its additional value was very different.

That difference is what a counterfactual adds to the story. If a higher-scoring applicant would probably have worked without a Census position, the job may substitute for another job or briefly rearrange the timing of work. The public position still has a wage and may produce public output, but it adds less employment than it would for someone who otherwise would have had no W-2 record. By contrast, if a low-score applicant was close to remaining outside formal employment, a temporary position can change whether they re-enter a cycle of reported work at all.

The score gradient makes the point without turning test performance into a measure of human worth. In the broader Census design, the estimated 2010 effect was about 46 percentage points for scores 70–80, 23 points for scores 82–90, and 16 points for scores 92–100. The test score is useful here because it organizes the study's hiring process and correlates with prior labor-market position; it is not itself the policy objective. What matters for the employment effect is the distance between the treated person's actual path and the path that would have occurred without the opportunity.

This is the first principle to carry forward: the benefit of a temporary job is an incremental benefit. It is not the number of people who worked during the program, and it is not necessarily the average quality of the jobs participants later held. It is the difference the program made relative to the next-best—or no—path for the people it actually reached.

The principle also clarifies why the U.S. result can be both large and local. The threshold design compares applicants just above and below 70, whose observed characteristics and pre-hiring outcomes are broadly similar. It gives a credible estimate for the people whose hiring probability changed at that boundary.