The way a system is designed tends to become invisible once it has been running long enough. The arrangements are settled, the assumptions harden and the decisions made at the outset stop looking like decisions at all and start looking like the shape of the world. In English higher education, the architecture established after the Dearing report has been in place for close to three decades. For most who work in the sector, it is the only system they have known professionally, and that longevity carries a particular risk. When a design becomes familiar, it ceases to be available for examination.
A recent paper from the Post-18 Project attempts that examination. It takes a diagnostic framework originally developed to analyse school systems, which identifies a set of control factors that must be held in coherent alignment for any education system to function, and applies it to English higher education. What it finds is not a failure of effort or of intent. It is a failure of architecture. In the settlement that followed Dearing, student choice was not one control factor among several. It was positioned as the single mechanism through which the distinct management of at least five separate functions was expected to occur. These included funding distribution, quality signalling, the shaping of provision, regulatory accountability, and the direction of investment. All of these were, in one form or another, routed through the aggregate decisions of 17- and 18-year-olds choosing where to apply and what to study.
This is a striking observation, perhaps easier to see now than at the time. The original design had a certain internal logic. If fee-paying students were directly selecting providers, their choices could plausibly be said to shape the system. If they were signalling preferences through their applications, those signals could be treated as information about quality. If providers responded to demand, the distribution of provision would follow the distribution of interest. The mechanism was elegant in theory. The difficulty is that it was being asked to do the work of several distinct mechanisms at once, and there are limits to how much any single lever can carry.
What happens when a lever carries too much
When a single mechanism is expected to perform multiple functions, those functions begin to interfere with one another. A signalling mechanism needs information to be clear. A funding mechanism needs distribution to be stable. A quality mechanism needs judgements to be disaggregated from price. An investment mechanism needs direction to be coordinated rather than aggregated from uncoordinated individual decisions. These requirements do not always align, and when they are compressed into a single point of control, the tensions between them are absorbed rather than resolved. What appears from the outside to be a coherent system may, on closer examination, be a set of competing pressures held in unstable equilibrium by the mechanism meant to organise them.
The paper suggests something similar when it notes that per-student funding has returned to its pre-Dearing level in real terms, despite the fee system’s purpose of increasing it. It notes that sticker price bears no relationship to quality. It notes that the widening of access, often presented as an achievement of the post-Dearing architecture, was arguably not produced by the system at all but by public demand for growth, which the system then struggled to convert into equitable outcomes. Each of these observations, in isolation, might be explained away. Taken together, they suggest something about what happens when one mechanism is expected to carry a load it was never engineered to bear.
What coherence looks like
The paper also identifies a part of the system where things have gone rather well. Research funding, governed by expert-mediated accountability and the Haldane principle, has largely been left alone through successive reforms. It approaches coherence, in the sense that the different elements of how research is funded, evaluated, governed and directed operate through distinct but aligned mechanisms. And the outcomes, whatever criticisms might be made of particular features, are genuinely world-leading. This contrast is worth sitting with. It suggests that coherence, when achieved, produces distinctive outcomes, and that the sector is capable of designing coherent systems when it chooses to. The difficulty is not that coherence is unknown. The difficulty is that it is demanding, requires sustained attention to how different functions interact, and tends to be the first casualty when political pressures compress multiple objectives into a single reform.
The difficulty of seeing from inside
There is a further observation worth sitting with. It concerns those who have worked within the current settlement for most or all of their careers. When a system has run for thirty years, its design choices no longer feel like choices. They become the conditions within which decisions are made, rather than decisions themselves. Senior leaders are rarely invited to examine the architecture; they are asked to deliver within it. Boards receive papers framed around the targets the system produces, rather than the system itself. Strategic plans assume the existing settlement and ask how best to respond to it.
This is not a criticism. It is a description of how long-running systems entrench themselves. The frame becomes the ground. Examining the frame requires a deliberate act of stepping outside a way of seeing that has become habitual, and the discipline to do so without collapsing into nostalgia for what preceded the settlement or into false certainty about what should replace it. Both collapses are tempting, because each relieves the discomfort of holding open a question whose answer is not yet available.
What the paper appears to be asking, and what warrants reflection from those in senior positions in the sector, is whether the moment has arrived for that deliberate act of stepping outside. Not to produce a new settlement overnight. Not to adjudicate between the competing versions of reform already in circulation. But to notice the architecture. To name what it has asked its central mechanism to carry. And to consider what a system designed with a more distributed load-bearing might need to do differently.
It is not obvious which forms of inquiry are best suited to that work, or who would be trusted to lead them. It is not obvious whether the sector has the appetite, at this moment, to sit with the discomfort of examining something that has shaped careers and institutions for three decades. And it is not obvious how to distinguish, in advance, between genuine structural reflection and the kind of surface-level consultation that tends to confirm what is already in place. These may be the questions worth holding on to.