How much does research cost?

The average research cost investment per publication for 2014 was about $45,000 but varies enormously. The group of eight (high profile universities) were the least cost effective publishers, averaging over $75,000 in research income per publication. The science, research and innovation sector has experienced a 13% decline in funding over the last 4 years. However since 1996 the collective funding for the sector has increased by 38%.

 

My friends and colleagues in academia are not going to thank me for this post which overly trivialises the entire multi-billion dollar Australian research industry. However I have lived through funding cuts to research and have always been exasperated by claims that we must aspire to have a ‘knowledge economy’ one week followed by cuts to research funding the next. This post is focused on getting a high level overview of just how big the research sector is, how funding has changed and how well our universities do with the funding they receive.

Funding changes for the Science, Research and Innovation sector

My sense from my academic friends is that the collective science and innovation budget has declined through time. The historic budget tables (converted to 2015 dollars) shows that in 2015 dollars the federal contribution to science, research and innovation is down by nearly 13% from a peak of $10.97 bn in 2011-12 (Figure 1). A 13% drop over 4 years is pretty dramatic for an industry which is heavily dependent on government support. However to put this in some perspective, despite the recent drop, the overall government investment in the sector is up 38% from 1996-97. The short term view therefore shows some recent pain, but over the longer term this has certainly been a growth sector.

Figure 1: Science, research and innovation funding through time ($2015) across socio-economic fields.

 

The individual socio-economic objectives in Figure 1 tell a different story. There was a rearrangement on how these where attributed in 2001 (however they have been consistent since then). The key winners with more than a doubling in funding are political and social systems (whatever that is), energy, transport, telecommunications, health and exploration and exploitation of space. The environment (my area) has seen an 82% increase in funding since 2001. Unfortunately culture, recreation, religion and mass media is down by 72% over the same period.

Figure 2: Winners and losers: changes in funding from 2001 to 2015 for different socio-economic objectives.

In 2015-16 the Australian Government provided $9.7 billion in support of science, research and innovation (that is about 2.4% of the Federal Budget). This funding is spread across many streams and includes the Governments R&D tax incentive program ($3 billion) which is focused mostly on the private sector but it also includes the university performance based block funding and Australian Research Council funding. Figure 3 shows how the Science, research and innovation budget has tracked as a proportion of the overall budget. You can see the peak in 2010-2012 at about 3% – and now at about 2.4%.

Figure 3: Science, research and innovation funding as a percentage of the federal budget.

How effective are our Universities?

The previous section deals with the entire science, research and innovation sector. The next section is only focused on academic research.
There are 43 Universities reported under the Federal Government’s Science, Research and Innovation Tables. Whilst these institutions provide undergraduate teaching I am just focused on research output here. The basic currency of research output are research publications (books, chapters, journal and conference papers). The research income and publications data for 2014 is a compilation of the collective research focused funding across all sources (public and private, competitive and non-competitive grants) as well as the collective publication output from each institution.

So how much does it cost to produce a research publication? We have excluded Torrens University (South Australia) as they don’t feature in the funding tables, and the Bachelor Institute of Higher Education (Northern Territory) due to their small size. We have used the total weighted publications (which adds a factor of 5 weighting for books). For the remaining 41 institutions (Table 1) the most cost effective publishers are the University of Notre Dame ($13,063 per publication) and the most expensive research institution is Charles Darwin University at $139,276 in research funding received for each publication (ouch). The Charles Darwin Uni cost per publication numbers seem outrageously high and I suspect some error in reporting the publications.

Table 1: Cost of publications

Institution	Total research funding (2014)	Total publications 2014 (weighted)	Cost per publication $ (weighted)
The University of Notre Dame Australia	$2,266,068	173.47	$13,063
Charles Sturt University	$13,905,636	912.42	$15,240
MCD University of Divinity	$2,778,299	166.05	$16,732
Bond University	$4,926,947	264.27	$18,644
Federation University Australia	$5,792,610	309.3	$18,728
University of Western Sydney	$26,230,746	1340.78	$19,564
Central Queensland University	$8,642,595	399.4383	$21,637
Swinburne University of Technology	$22,907,038	1015.8	$22,551
University of Canberra	$14,527,927	611.29	$23,766
University of the Sunshine Coast	$8,736,547	323.91	$26,972
Deakin University	$47,831,018	1719.12	$27,823
Victoria University	$18,016,560	624.78	$28,837
Griffith University	$67,774,267	2305.86	$29,392
University of Technology, Sydney	$46,868,921	1583.27	$29,603
University of Southern Queensland	$11,352,852	374.98	$30,276
Australian Catholic University	$15,401,125	493.87	$31,185
Southern Cross University	$12,006,209	380.07	$31,589
Macquarie University	$57,928,838	1832.85	$31,606
Edith Cowan University	$16,031,869	504.08	$31,804
RMIT University	$61,845,691	1689.49	$36,606
University of Wollongong	$63,713,054	1708.81	$37,285
Murdoch University	$23,357,652	591.61	$39,482
The University of New England	$32,230,612	788.36	$40,883
University of South Australia	$71,810,567	1607.04	$44,685
Queensland University of Technology	$91,474,123	1956.9617	$46,743
Flinders University	$53,883,850	1150.89	$46,819
La Trobe University	$55,381,434	1157.56	$47,843
Curtin University of Technology	$80,457,602	1486.67	$54,119
The University of Newcastle	$92,960,318	1623.7	$57,252
James Cook University	$64,369,546	1081.75	$59,505
The University of New South Wales*	$344,876,605	5368.5659	$64,240
The University of Sydney*	$356,548,641	5501.62	$64,808
University of Tasmania	$97,052,806	1370.13	$70,835
The Australian National University*	$272,410,232	3631.945261	$75,004
The University of Adelaide*	$182,082,440	2362.15	$77,083
The University of Western Australia*	$204,561,471	2544.46	$80,395
The University of Queensland*	$377,307,509	4683.37	$80,563
The University of Melbourne*	$412,378,080	5118.11	$80,572
Monash University*	$355,948,869	4319.47	$82,406
Charles Darwin University	$46,077,776	330.837	$139,276
Grand total	$3,744,811,264	65441.07816	$57,224

* Group of eight
What is really striking is how poorly the ‘group of eight’ most prestigious universities perform. Arguably they must do much more expensive research and invest in major research related infrastructure that others do not. I’m sure that they would also argue that their research has a larger impact (the fewer papers is made up for by better journals) than other institutions. However, eight of the ten least cost effective publishers are the group of eight universities.

Figure 4: Distribution of the cost per publication (dashed line in each box is the mean)

What about the CSIRO?

The Commonwealth Scientific and Industrial Research Organisation (CSIRO) is a federal government funded research organisation. They don’t feature in the research funding and publications tables so we cannot directly compare them with the universities. However, to give some context – based on their 2014-15 annual report the CSIRO’s government funding for 2014-15 was $745.3M (total funding including external sources was $1,214M). The total number of journal papers for 2014 was 3,176 (or if you include all publications including technical reports it was 3,552). If we consider just the government funding against the journal papers you get a publication cost of $234,666 per paper. If you count all funding sources (as we have done for the universities) and all output (even reports) you get a cost per publication of $341,779 (that seems pretty high). In their defence, the CSIRO does a lot more than just produce publications, there is a lot of infrastructure to maintain and blue sky research that doesn’t yield publications.

Caveats

This summary is superficial. It is not just total publications, but broader impact which incorporate the underlying quality and reputation of each journal that should count toward the university publication record. The broader educational role of the universities in producing postgraduate (higher degree research) students is part of the costs included in the research quanta. So those universities with high numbers of postgraduate students may have relatively low publication output, but they contribute to scientific endeavour by creating the academics of the future.

About the data visualisations

The visualisations in this post are contained within their own Dashboard. You can link to the Dashboard, or use the short code from each viz to embed or link to specific Viz’s in another web page. You can simply ‘like’ any of the visualisations to post them on your social media feed.
The viz’s are easy to make. All you need to do is create a free Truii account to create and publish your own data visualizations.
Don’t forget to sign up to Truii’s news and posts (form on the right).