Historically, increases in commencing bachelor-degree students flow through into increased completions in the three to five years afterwards. And initially the demand driven boom of 2009 to 2014 looked like previous patterns. The increased commencing cohort sizes, shown lagged by four years by the orange line in the chart below, are evident in larger completing cohorts between 2012 and 2015 (blue line).

But then growth in completions slows to a near stall in 2017, which had 0.3% more completions than in 2016. In 2018 there were 2.2% more completions than in 2017, but this still looks surprisingly low. If there had been the same relationship between completions and commencements four years later in 2018 as there had been in 2008, nearly 26,000 more people would have finished their degrees in 2018 (grey line in the chart above).

Last year a consulting firm noticed a changed relationship between enrolments and completions, and suggested that people not finishing their degrees was a problem. Their estimates of drop-out rates are too high, but the chart below shows that there was an increase from 2012. Between 2008 and 2015, the share of the commencing cohort that had disappeared without a degree four years after commencement increased by 3.6 percentage points. But in trying to understand a 10 percentage point decline in the four year completion rate this does not provide a full explanation.

Another possible cause of surprisingly low completions is that students are just being slow. According to the Department’s completions data the proportion of people still enrolled without completing at the four-year point was 1.1 percentage points higher for the 2015 commencing cohort than it had been for the 2008 commencing cohort. But there have been no major recent changes in the proportion of students still enrolled after 4 years. So while delay will increase the average time it takes to complete a degree, it should not have a big impact on numbers of completions over a multi-year period.

I think most of the balance of the explanation comes from course and institution changing and students enrolling in second degrees. The chart below does not have an exactly right count of the repeat customers, as the unique student identifier (the CHESSN) only starts in 2005. With this caveat, the share of commencing students who had definitely been enrolled in higher education before was 19% in 2008, but more than 30% in 2016. Course changing students are counted twice as commencing students, but only once as completing students if they eventually finish a course.

Using unique students, the relationship between numbers of commencements and numbers of completions after four years is still changing, but there is a 5 percentage point rather than 10 percentage point decline in completions as a percentage of commencing students four years previously. Increased rates of dropping out and slower completions (due to lost time from changing courses, as well as some recent increase in part-time study) should explain the remaining difference.