England and Australia have long had parallel trends in ÖйúAƬ policy. In England, the teaching excellence framework (TEF) will ultimately link student fee increases to university performance. In the May 2017 Australian budget, the government announced that, in future, part of its teaching grant will also depend on university performance.

The details of Australia¡¯s scheme will be worked out in consultation with universities. But Simon Birmingham, the education minister,?has indicated that student satisfaction, retention, completion and employment will be among the performance indicators.

Student attrition has been rising in Australia. In 2009, 12.5 per cent of domestic bachelor¡¯s degree students did not return to study after their first year at university. By 2014, that figure was 15.2 per cent. This increased attrition is partly a side-effect of policy. After the lifting of previous caps on undergraduate numbers, commencing students increased by a third between 2009 and 2014. Compared with other students, these new arrivals were on average less academically prepared for university study, and this is one of the largest known attrition risks.

Encouraged by government policy and funding, universities have taken many more students from disadvantaged communities. Most complete their degrees, but they are over-represented in multiple attrition risk categories: not just the less well prepared, but also mature, part-time and online students.

ÖйúAƬ

Institutional factors matter, too. Government analysis suggests that much of the national attrition increase is accounted for by a small number of universities, some of which are doing worse than other universities with similar student profiles. Performance funding would penalise such poor outcomes.

But designing the right retention incentives is very difficult. Some students will know only after enrolment whether university is right for them. And even if it is, there are often good reasons for abandoning or delaying studies. The major Australian student experience survey asks its respondents whether they are considering leaving university without finishing, and if so why. The top seven of 30 possible reasons cluster around financial, health and workload problems, none of which universities can solve directly.

ÖйúAƬ

¡°Health and stress¡± issues are by far the most common reasons, cited by 41 per cent of people thinking of dropping out in 2016. Several recent studies show that anxiety and depression levels are higher among students and young people than among the general population.

Universities could do more. In a recent survey of students who had left an unnamed Australian university, more than half said that there were things it could have done to encourage them to stay. But that doesn¡¯t mean that these things should always be done. The danger with policies that give universities financial incentives to increase their retention rates is that they can encourage both good and bad behaviour.

On the positive side, universities may take more care with their admissions processes and improve their student support. On the negative side, they may promote retention through passing students who should fail, or by pressuring students considering leaving to stay. If it turns out that university probably isn¡¯t a student¡¯s best option, it is better for them to save their money and move quickly on to something else; unfortunately, recent cohorts are staying longer before leaving, making their mistake more expensive. And while university counselling services can often assist with mental health issues, some time out from anxiety-inducing study deadlines and exams could be even more therapeutic.

Governments can promote good practice, at lower risk of unintended consequences; the Australian government is already publishing improved statistics on completions, which can be built on to provide information that guides students¡¯ and institutions¡¯ decision-making. They can help with research on and promotion of university policies and practices that reduce attrition. We can learn from the universities with higher retention than other institutions with similar student profiles.

ÖйúAƬ

But policymakers also need to accept that desirable trends are not all going to arise in perfect harmony with each other. If we want a more open and accessible ÖйúAƬ system, all other things being equal, one price of that is going to be higher attrition. It should be kept as low as possible, but, ultimately, it is a price worth paying.

Andrew Norton is the ÖйúAƬ policy director at the Grattan Institute in Melbourne, Australia.?