University Productivity - UNSW vs UTS

Full Article:

The appointment of Professor Attila Brungs as vice-chancellor of the University of New South Wales is a strong one, and no doubt will be welcomed by the UNSW community. Any move brings the new incumbent challenges, but our data suggests that he will have a productivity challenge at UNSW compared with the institution he currently leads, the University of Technology Sydney.

We have developed the Research and Education Efficiency Frontier Index (REEF) methodology to analyse the productivity of universities because its underlying econometric technique is uniquely suited to organisations with more than one mission, especially where joint and common costs mean that there is no valid and reliable means of separating the missions for accounting purposes without making heroic assumptions about cost allocations and cost-drivers. Universities, with their dual missions of education and research, are a prime example of this, as many staff are appointed to do both teaching and research and many overheads cannot be unambiguously divided between the two. The REEF methodology charts the relative productivity of an institution as between its two missions. It can then be compared against itself over time, or with the sector as a whole or with specific institutions.

In earlier pieces we have used the technique to show a relative lack of diversity in the university sector (The Australian, 9 June) and the benefits of specialisation, in the case of Victoria University (The Australian, 30 June). The move of a vice-chancellor from one institution to another provides a further occasion to harness the REEF Index’s utility.

UTS has been a stronger performer during this century. In the graph below we see a near vertical rise in research productivity, from about 0.6 publications per million dollars of expenditure (in 2019 dollars) in 2001 to about 4.4 in 2019. There has been almost no productivity increase in education, illustrated by the 2019 dot being barely to the right of the 2001 dot, although at about 34 students per million it is quite productive, as a later graph shows.

This contrasts with the productivity picture for UNSW.

The graph below shows a steady rise in research productivity from 2001 until 2016, but then a noticeable fall to 2019. As with most of the sector, there was also a moderate rise in education productivity in those years: the dots from 2001 to 2016 move a little to the right, but they too then decline from 2016 to 2019 as the dots move back to the left. At all times, UNSW has educated fewer students per million dollars than UTS.

Now we can look at the two universities in the context of the Australian university sector. This is shown in the graph below which shows university productivity integrating outcomes in both education and research using by expenditure in millions of dollars as the input.

If we take the two institutions and show their position in the sector as a whole in 2019, we can see that UTS (the arrowed red dot) is closer to the efficiency frontier than UNSW (the arrowed orange dot). UNSW produced slightly fewer publications per million dollars spent and educated fewer students per million dollars spent.

The major 37 public universities are shown in the graph below, by reference to their affiliations (Group of Eight, ATN etc) with UTS and UNSW highlighted.

What might account for the contrast? The student/staff ratio is one possibility. In 2019, UTS had about 24 students per academic compared with about 17 per academic at UNSW. This is demonstrated in the graph below, which also shows that publications per academic were marginally higher at UNSW at 3.3 and 3.4 respectively.

That might suggest that the academic staff at UNSW were more research-active but the cost base of UNSW was also higher so when research publications are divided into million dollars spent the number is still lower than UTS. A key reason for a higher cost base could be the discipline mix at each institution. Fields such as medicine and agricultural science are more expensive as compared with, say, business and law.

While space does not permit a full analysis here of the many nuanced differences that are present, the two measures of productivity – one calibrated using expenditure and a second using academic staff resources – it is important to note that, while there are similarities in the productivity past trajectories of each of the two universities as measured by the two productivity measures; there are also striking differences.

As at 2019, for example, UTS is at the REEF frontier when measured by reference to academic staff, but not as measured with expenditure. UNSW is closer to the frontier as measured by academic staff but considerably further away using the expenditure measure. There are implications for these differences including the approach one might take in terms of priorities for finding productivity gains in the future.

Another further factor one needs to consider is “quality”. All universities are regulated to ensure they meet threshold standards in education and research so there is no suggestion of there being a problem of quality at either institution, but it is possible that a university goes through a period of aiming for fewer but higher quality research outputs. The results shown are the ‘base level’ REEF results without recalculating research productivity to examine varying levels of research quality. These base level results can be thought of as a ‘fair average’ research quality. When calibrated in this way, the research productivity trajectory of UTS has an edge.

Conversely, a university could decide to spend more time and attention on students to improve their learning experience. Without access to internal data we can only speculate, but we know of no reason to think that either of these explanations is a material one in the current contrast.

A further source of explanation is spending money on other things. If a university invests in, say, a new campus, or a restructure, or consultants etc, then it spends more millions on things other than research and education – and if those things are reported as expenditure and not capitalized assets or investments (as would be the case in many circumstances) – then this will result in lower result for publications and students per million dollars of expenditure. Productivity might then pick up again in future years, if these expenditures are wise resulting in future productivity gains in research or education or both. If it should turn out that these investments were not wise, however, then ground will have been lost in productivity without a compensating gain.

Concluding Remarks

The productivity outcomes of universities have, historically, been challenging to assess without overt or hidden preferences or biases in respect of the teaching/research intensity mix.

All Australian universities both teach and research. How productive the inputs provided to universities are converted to education and research outputs is of interest to a range of stakeholders – including, but certainly not limited to, incoming vice-chancellors as well as the wider community of universities.

The current positions of UTS and UNSW differ significantly in respect of the REEF productivity measures – both calibrated using expenditure and academic staff resources. Their trajectories over recent years also vary markedly. No doubt the outcomes from 2020, which should be available as soon as the data from Canberra is available, will likely show a very different picture, but the underlying trends will likely be best seen in these pre-COVID effected data.

What will the future bring to UTS and UNSW? It may be that the strong productivity growth seen in UTS will transfer to UNSW. Time will tell and the REEF methodology will be able to detect any ongoing productivity shifts.