The decline in international student numbers has many people worried about the future of university research in Australia. A recent report from the Chief Scientist predicted that 7,000 research jobs could go due to reduced teaching profits, philanthropy and corporate funding.
In this post, I estimate how reliant research is on international student profits. It combines data from multiple sources. None of them were designed to calculate this profit, so my result should be taken as being in a plausible range rather than as a precise total. But it can give us a sense of the scale of reliance on international students.
According to the 2018 ABS higher education research report that was released yesterday, in 2018 universities spent $12.158 billion on research. The ABS also gives sources of research funding, but these only explain 44 per cent of the total, with the rest coming from ‘general university funds’.
I am not going to use the ABS funding source categories, because the revenue sources reported in the Higher Education Research Data Collection can account for slightly more (50 per cent) of total research spending and include much more detail. This detail will help me avoid double counting when I use revenue line items from the 2018 Finance report to investigate other possible sources of funding (a few institutions are in the ABS survey but not the Finance report, but they are small).
The table below spells out my assumptions, all of which could be challenged. As universities give research a high priority, and there is a large amount of spending to be explained, it would be unrealistic to think that sources of potential discretionary revenue are not being used heavily to support research.
On the other hand, I need to leave revenue for spending on things other than teaching and research and to produce the reported surplus for 2018. On the ABS figures, 38 per cent of university expenditure is on research, and on Deloitte’s analysis 52 per cent is on teaching, so 10 per cent of expenditure is on other things.*
Combining research-specific funding sources with possible non-teaching sources of revenue gets me to about $7.5 billion, or 62 per cent of research spending. That still leaves about $4.6 billion in research expenditure with unexplained funding. The only major source of revenue left is teaching.
Earlier in the week I estimated total teaching profits for 2018 at $5.4 billion, which is enough to cover the unexplained research expenditure. Of that total, approximately $1.3 billion came from domestic students. Assuming 100 per cent of that was spent on research, 73 per cent of research spending is now explained – leaving $3.3 billion of spending looking for a funding source.
On fairly conservative assumptions, this means that 27 per cent of total research expenditure, or about $3.3 billion, relies on profits from international students.
——————————–
*I am assuming there is no double counting in the teaching and research estimates, but given that each of them relies on estimates of how teaching and research staff spend their time, and how dual purpose facilities are used, that may not be the case.
Andrew Norton 
I’ve often wondered how accurate the ABS data is – given it is based on survey data populated by university research offices. The numbers line up approximately with a 40/40/20 split of academic time. This can be calculated by a chain of assumptions… 69%/31% split of salaried person years of effort (PYE) across academic and non-academic staff…. 69% of ABS research labour costs gives you $3.6b spent on academic salaries… $3.6b as a % of academic salary expenditure from DESE HE Finance report gets you close to 40% of academic salaries spent on research. The 40/40/20 split is a conventional distribution of research / teaching / engagement time but I hear nothing but grumbling from academics that their allocation for research is no where near the 40%. convention. The question arising from this is whether the ABS survey data averages out at 40% across the sector for those that are high and low on research, or whether it mainly embeds an assumption that academic salary costs come in at 40% irrespective of the actual costs given workload allocations etc. If the data embeds a 40% assumption then many of the downstream findings on research expenditure might be inaccurate.
LikeLike
The footnote about double counting academic time was alluding to concerns such as these. We know from the much more detailed Deloitte teaching report that the quality of the data is likely to be variable between institutions. On the other hand, the trend passes reality tests such as numbers of publications.
LikeLike
Interesting. Makes me wonder if the data you have available allow you to go back in time before international student revenue was a big line item in uni budgets (e.g the 90s) to see if you could potentially account for all research spending using the method you’ve outlined here.
LikeLike
Sort of, although there some methodological issues created by the transition over 25 years from a general block funding system for teaching and research to one in which money is ‘clawed back’ (as they used to say) and redistributed based on purely research-based criteria. Also, there is no teaching cost data from the 1990s to help derive a teaching-research split.
LikeLike
Thanks for this breakdown. A query on the list of revenue sources. In the ABS data about $1bn of the $12bn is capital spending, presumably financed by a mix of operational revenue and borrowings in many cases. In the breakdown, where do borrowings sit as a revenue source?
LikeLike
Geoff – A good question, as they are not included in the financial summary I used. So far as I can tell sector level net borrowing in 2018 was only around $300 million, but it is quite possible that the situation was very different at specific unis. From when we did a similar exercise at Grattan, one issue is the that the ABS uses an essentially cash methodology while the Department uses accrual. This may help explain why some line items that seemingly should match between the two sources are often out.
LikeLike