There is a simple question sitting underneath many higher education funding debates: what does it actually cost to teach a student?
It is simple to ask, but much harder to answer.
Most systems need categories. They are administratively necessary. Governments cannot fund every class, course, campus arrangement or teaching model as a unique case. So higher education funding tends to rely on broad groupings: discipline, level of study, mode, student type or combinations of these.
But the risk is that the categories become more convincing than the evidence behind them.
In our recent paper “Analysing the costs of teaching in universities through structural cost profiling,” we examined this issue using Australian data. We analysed nearly 190,000 unit-of-study records from 12 universities, looking at teaching costs alongside student load, staffing, delivery mode, campus type and other operational indicators, such as student fees.
Funding versus cost
We developed and applied a novel approach to the data, which we called “structural cost profiling.” Instead of asking whether isolated variables explain differences, it asks how several features of teaching delivery come together in recurring cost patterns.
We found that teaching costs vary, but that variation has a pattern that’s not obvious at first look. Staffing numbers, campus location, delivery model and student mix all matter, but do so in particular combinations.
This has implications for policymakers and university managers, where funding rules and internal accounting might be misaligned to cost drivers.
Higher-cost patterns were associated with delivery at small regional campuses and more labour-intensive postgraduate provision. Lower-cost patterns were more often associated with large-scale online delivery but not always, as some online provision was not uniformly inexpensive, particularly where it was embedded in smaller or more resource-intensive delivery models. This is especially the case where academics are located in rural campuses.
Funding categories should not be treated as cost categories.
Funding systems should remain relatively simple, but should recognise structural cost drivers where the evidence shows they are persistent, particularly scale, location, staffing intensity and mode of delivery.
These findings might sound obvious at first. But the underlying patterns matter.
A common policy reflex is to treat some areas of study as expensive and others as cheap. There are reasons for this. Clinical, laboratory-based, and studio-intensive forms of provision often have higher costs. But once we examined unit-of-study level data across multiple dimensions, the picture became less tidy.
One example of this is the high cost of “regional” delivery. Our analysis shows that this comes from a combination of structural factors: smaller cohorts, staffing intensity, and more complex provision, which is easy to miss without looking at the whole cost picture.
The implication is seemingly neat cost divisions, such as by areas of study or discipline. Think the lab versus the class room, can be deceptive and do not tell the full story.
And this is relevant beyond Australia.
Boomeranging back to England
In England, for example, teaching funding recognises high-cost subjects through broad price groups and targeted allocations.
These categories are necessary, but the wider financial problem facing universities is not only about whether subject categories are right or wrong. It is also about whether those categories are being asked to do too much work.
If a funding model recognises laboratory costs but not the diseconomies of small-cohort regional delivery, it may miss one kind of structural cost. If it assumes online delivery is automatically cheaper, it may miss the staffing and support required to teach well in particular contexts. If it treats course and subject groupings as stable proxies for cost, it may not see how the same field can be delivered through quite different operating models.
The sector is being asked to do several things at once.
Universities are expected to widen participation, support student success, respond to skills needs, sustain regional and local provision, improve teaching quality and remain financially viable. These goals are not cost-neutral, they change the structure of teaching. This matters particularly in England’s current financial climate. With universities under sustained pressure from rising costs and volatile recruitment, unclear assumptions about teaching costs can lead to blunt cuts or the withdrawal of provision that is structurally expensive but publicly important. Better cost evidence would help government target support more deliberately, and help universities distinguish genuine inefficiency from the cost of sustaining access, quality and regional provision.
The uncomfortable implication is that some provision may look inefficient because the funding model is not seeing the right structure.
Regional delivery is a good example. Smaller cohorts and distributed campuses can weaken economies of scale. But if regional access is a public good, then the additional cost is not simply inefficiency. It’s part of the cost of delivering the policy goal.
The same applies to student support and access. Widening participation may require more intensive teaching, transition support, academic guidance and local infrastructure. These are not peripheral costs. They are part of what it means to deliver higher education well to a broader student population.
Our findings also suggest caution around easy claims about online learning. Scale-based online models can be lower cost. But online delivery is not a single cost structure. A large, centralised online unit is different from a small online unit embedded in a regional or professionally intensive program.
Don’t forget about the price tag
The broader lesson is that funding policy needs to distinguish between price, cost, and purpose.
Price is what students or governments pay. Cost is what is required to deliver teaching in a particular configuration. Purpose is what the system is trying to achieve: access, equity, skills, quality, regional presence, research-led teaching, or a mix of these.
Problems emerge when funding formulas assume the cost structure but do not make the purpose clear.
None of this means funding should become impossibly granular. A system that tries to price every teaching activity separately would be unworkable and probably undesirable.
Universities also need flexibility to manage cross-subsidy and local priorities but simplicity should not depend on weak assumptions.
A more practical path is to keep funding models relatively simple while adding structural recognition where there is strong evidence. This might include regional delivery, persistently small cohorts, placement-heavy provision, staffing intensity, or other features that repeatedly shape the cost of teaching.
For policymakers, the message is not “make the formula more complicated,” it’s make the formula more honest about what it is trying to support.
For universities, the message is also important. Better cost evidence should not only be used for cuts or efficiency narratives. It should help explain why some teaching is more costly, why some provision matters despite being expensive, and where genuine scale efficiencies may be possible.
In periods of crisis or transformation, that distinction supports more targeted decisions, rather than across-the-board reductions that risk weakening access or quality.
Teaching costs vary, but that variation has a pattern.
If funding systems do not recognise that pattern, they risk treating structurally different forms of provision as if they were the same. And that is where policy misalignment begins.
