Harvard's Fryer found paying kids for grades fails — paying for inputs works

Economist Roland Fryer ran a series of paid-incentive experiments in public schools in Chicago, New York, Dallas and Washington, D.C. — $6.3 million paid out to 38,000 students across 261 schools, covering second graders through ninth graders. NPR's Planet Money reported the results in February 2011.

The headline finding inverts a basic economic idea: you should pay for output (grades) and let workers figure out the best inputs. Direct cash for good grades didn't work. But paying for inputs — doing homework, attending class, reading books — did seem to boost achievement in several cases.

The explanation is the leading theory Fryer's paper offers: students 'do not understand the educational production function and, thus, lack the know-how to translate their excitement about the incentive structure into measurable output.' Input payment dissolves that problem — a paid reader 'simply needed to know how to read... or sit still long enough to collect their short-term incentive.' Paying second graders to read books looked particularly promising: it lifted reading comprehension relative to unpaid peers, and the improvement persisted a year after payments ended.

The design isolates the mechanism: same kids, same schools, same money — only the payment target (input vs output) changes, and only input pay moves results.

It explains a whole class of failed incentive programs: paying for outcomes presumes the payee already knows the causal recipe.

The persistence of the reading gains after payments stopped shows input pay can bootstrap a habit rather than rent one.

Pay for outputs only when people already know how their actions produce them; when they don't, pay for the actions themselves.

The studies — Fryer's paper is linked from the NPR piece — became the standard citation for why education incentives must target inputs; paying second graders to read remained the most promising variant reported.