State funding for colleges and universities in the United States is increasingly being matched to "results" in a shift that mirrors experiments in accountability in Britain and Australia.
More than half of the 50 states, whose governments supply the lion's share of funds to public universities, now link at least some of the money to a mix of performance ratings, from graduation rates to student satisfaction. Several more are considering it.
In the most extreme case, South Carolina has adopted plans to distribute its entire budget for public colleges on the basis of 37 different criteria designed to measure their performance. More often, the discretionary funds only account for 1 or 2 per cent of the total.
Such schemes are often deeply controversial. Arkansas abandoned so-called "performance funding" after bitter complaints from urban and community colleges that they were penalised by unfair comparisons to elite universities. But Joseph Burke, director of the Public Higher Education Program at the Rockefeller Institute of Government in New York state, said:
"there appears to be a movement towards saying in some ways results should count in the funding of public colleges."
Demands for value for money are making themselves felt - people increasingly believe results should count as the price of a university education continues to rise.
Mr Burke oversaw a recent survey by the Rockefeller Institute. It distinguished between performance budgeting, where state officials look at results in a broad sense when allocating funds, and the more draconian performance funding, where money is tied to specific indicators.
Twenty-one states practise performance budgeting, it concluded, and 13 performance funding. While rewarding results is politically appealing, it is littered with potholes. Kentucky and Minnesota, for example, as well as Arkansas, experimented with performance funding and then abandoned it, though other states are pushing ahead.
Graduation rates are often considered a good measure of college achievement, along with the time it takes students to earn degrees, their scores on exams, the number of students finding jobs or enrolling in graduate programmes. But the question is how to select criteria that are fair and flexible.
A community college with low admission standards, Mr Burke said, might count it a success to have 35 per cent of students graduate within two or three years. The same rate would be a major scandal at an elite research university. Colleges may inevitably be tempted to fudge or manipulate their results.
"The attractiveness of performance funding in theory is matched by its difficulty in practice," Mr Burke said. On the other hand, he noted, "even fudging results is an advance on not thinking about results at all".