Paying academics bonuses for their research encourages them to join more productive colleagues, thereby concentrating scholarly talent and boosting output, according to a study of German universities.
The research bolsters the argument for introducing controversial performance-related pay, a measure recommended by a Treasury-commissioned report in February.
But it also warns that by concentrating the best academics into fewer places, people in some areas of the country could be deprived of a decent scientific education.
The study used data from Germany, where from 2005 a change in the law meant that all new academics would be eligible for performance-related bonuses on top of a basic wage.
Under the new system, known as “W-pay”, bonuses are paid for research performance and winning research funds, as well as for taking on management duties, or are used to attract and retain academics.
Universities have discretion over how they award bonuses, which can be worth more than €5,000 (£3,650) a month. In some universities, an annual “prize pot” is divvied up at the end of the year according to each academic’s relative performance. On average, just over a quarter of a German academic’s salary comes through bonuses.
After the reforms, the top performing researchers, as measured by publication rates, clustered together more than they had previously, according to “Lone stars or constellations? The impact of performance pay on matching assortativeness in academia”.
High-quality departments hired far more productive researchers than their lower performing counterparts. They also got rid of poorer performing academics, while lower quality departments lost more productive scholars.
This kind of clustering is known as “positive assortative matching”, explains the paper, which was presented on 26 August by its author Erina Ytsma, a doctoral candidate at the London School of Economics, at the European Economic Association’s annual congress in Mannheim.
In other words, the best scholars seek out similarly talented colleagues to work with in order to boost their productivity, just as highly educated or wealthy people look for mates with similar traits.
Concentrating workers into high-productivity and low-productivity clusters results in higher production overall than when groups contain a mixture of talents, the paper argues.
“A greater total scientific output may boost technological progress, so to the extent that there are positive productivity spillovers in academia, this calls for a concentration of the most productive academics,” Ms Ytsma’s paper concludes, although it warns that this may come at the cost of providing “good scientific education to many people, all over a country”.
The results from Germany play into the long-standing debate over how far to concentrate research funding into a few select institutions, and whether to pay academics in relation to their performance.
Earlier this year a Universities UK review, Efficiency, Effectiveness and Value for Money, recommended phasing out automatic annual pay rises in favour of performance-related pay.