Skip to content
The UK's only independent think tank devoted to higher education.

Why higher education should take the ‘signalling critique’ seriously – and what that might look like

  • 19 October 2021
  • By Chris Percy & Dr Aveek Bhattacharya

This blog was kindly contributed by Chris Percy and Dr Aveek Bhattacharya, regarding their recent publication with the Social Market Foundation. They are both on Twitter: @chris_percy and @aveek18.

A couple of weeks ago, HEPI published a paper by Lord David Willetts, making a forceful case for continuing to support and expand higher education. There is a lot that we agree with in that paper. Education can be a transformative experience, improving graduates’ earning power and wellbeing, with spill-over benefits for the rest of society. Willetts’s recommendations – more universities, better funded, alongside greater investment in further education and lifelong learning – is a particularly clear example of what seems to be the orthodox position among educationalists: more is better.

Yet as Willetts recognises, we are in a time of great scepticism regarding the value of higher education. One concern he flags is the declining graduate wage premium in the UK, which we chart in the report (see the excerpt Fig 3 below). The think tank Onward argues that many degrees offer a ‘poor return on investment’ because of how little their graduates earn. There are clear indications that the Treasury wants to recoup some of the billions of pounds of public money that goes into the system. 

One argument critics of higher education often make is the so-called ‘signalling critique’. This is the claim that a significant proportion of the wage benefit of education comes from demonstrating a person’s pre-existing skills and abilities, not from enhancing those skills and abilities. In short: a meaningful share of the value of the degree comes from the fact you were able to get into a particular university rather than what you learned there.

Some signalling is probably unavoidable – and not worth worrying about – but no-one funds universities in a desire to fuel an arms-race to signal capabilities that everyone had in the first place. From a fiscal and a moral perspective, we should all want as much of education to be focused on long-lasting, transformative experiences (including those that increase productivity overall), rather than just making it easier for employers to adopt lazy recruitment filters (and does little for productivity or anything else).

Productivity and economic benefits are, of course, only one motivation for taxpayer subsidy and individual investment in higher education – but they are at the centre of the fiscal arguments with the Treasury that Lord Willetts hopes to win and a primary motivation for many students, especially in England (see citations in our report, on pages 8 and 9).

The higher education sector has typically chosen not to engage with the signalling critique at a policy level. Three examples: Lord Willetts mentions it and dismisses it without providing any evidence for or against (p.20), focusing instead on such concerns as the culture war and voting trends. A recent research synthesis from UCL Centre for Education Policy and Equally Opportunity gives the evidence on signalling a fairer hearing, but prefers not to place a number on how important it is for wage gains, arguing instead that any impact it has is ‘limited’.

A third example can be found in Professor Danny Dorling’s book review for HEPI, arguing against the most bombastic case for signalling: Professor Bryan Caplan’s The Case of Education. The review presents one counter piece of research but focuses most of its attention on Professor Caplan’s upbringing, character and motivations. The counter evidence is welcome (and well put), but a single piece is not sufficient to overturn a synthesis study, no matter who the author is or why they wrote it.

At the same time, signalling advocates share some of the blame for a lack of constructive dialogue. Caplan’s primary policy recommendation is for sharp ‘educational austerity’, such an extreme suggestion that it acts as a conversation stopper. His other ideas, like increased emphasis on vocational education, can subsequently struggle to get a fair hearing. A defensive backlash may not be sensible, but it is understandable.

In our view, both sides have a point: much of education is extremely valuable, but there are aspects to it that are socially wasteful. By trying to bring them together, we can better direct the time, energy and resources of our education system to things that really benefit learners and society at large. That is where our recent paper, Signal Failure, comes in.

We do not know for sure how much of the wage premium to higher education is due to signalling alone, rather than productivity enhancing gains like new knowledge and skills. It is fiendishly hard to accurately estimate and the academic debate has raged for decades: it is easy to cherry-pick studies to support either side of the debate. However, our review of five different econometric techniques suggests that it is a big enough issue to merit taking seriously: a conservative 20 per cent to 40 per cent of the wage premium is likely to be signalling (see the Table 1 excerpt from our report below).

Some of the studies we cite that land within the 20 per cent to 40 per cent range conclude that this level of contribution is unimportant, perhaps because it is not the ‘primary’ driver, being less than 50 per cent. We disagree. If it were only 5 per cent or below, it might be ignorable – but it is hard to believe that any good faith synthesis study could arrive at such a point estimate. Given the amount of time and money going into higher education, even the low end of the range is worth our collective attention.

So what to do about it? We don’t think the answer is broad-based defunding of education (at any of its levels). Indeed we fear that the sector’s current response to the signalling critique makes such an outcome more, not less, likely.

Here are three initial ideas for ways the higher education sector could try to limit (not remove) the role of signalling:

  1. Shift the emphasis towards degree subject and grade. The emphasis on the institution of study in graduate recruitment places weight on successful admission (a signal of priority ability) rather than what is achieved in the course (the human capital gain). More granular degree classifications that are more comparable between institutions would give employers much more to work with.

  2. Make it easier to stop, start and switch. Where signalling prior ability is a key benefit, we should help students gain it more quickly, without sacrificing difficulty lest the signal lose credibility. Shorter courses, micro-credentials and modular courses would all help. Students could be actively helped to stop, start and switch courses, allowing for multi-year breaks with less stigma, underpinned by excellent career guidance and employer engagement to navigate the newfound flexibility. When students come back (and surely many would) they would do so with a much clearer idea of the value being gained. Such a shift would help a culture of lifelong learning and engaging mature students, who are also plausibly more likely to pursue education with genuine human capital benefits.

  3. Pick up the measurement challenge. Capture learning gain directly. Track the role of signalling consistently over time and for subgroups. Use this as part of a constructive engagement with policymakers. We must not let the perfect be the enemy of the good. Inflation and social inequality are critical social issues that are very imperfectly measured today – and yet we still make important policy gains from tracking proxy metrics and understanding the metrics’ limitations.

Signalling is not just down to the higher education sector and other actors have a role to play as well. Our report suggests ideas for schools and colleges providing career guidance, employers’ hiring habits and the system infrastructure overseen by policymakers.

Some of these proposals are familiar – and are attractive for a range of different policy outcomes. This is not a bad thing. We might hope the signalling lens can both strengthen the case for such policies and take the heat out of a critique that has such potential to harm the sector if it continues being exploited for political ends.

We do not pretend to have all the answers. The ideas above are the start of the conversation, not the end of it. Some might not turn out to be effective or feasible, but one or two might mark a genuine improvement – and there will be other ideas to bring into the tent as well. Please get in touch if you would like to publish a formal response with us, join a group discussion, or simply share a reaction.

Get our updates via email

Enter your email address to subscribe to this blog and receive notifications of new posts by email.


  1. conor king says:

    Perhaps time to talk of the negative impact of not having completed a post school qualification rather than the advantages of it. Given in most advanced economies (more than western) the majority do gain some post school qualification.

    There is one reference here (that I saw) to senior secondary. Surely the argument applies to any level of education – it merely signals some are more capable than others etc.

    Yet a person with no education might teach themselves to read etc but otherwise no matter the intrinsic ability would struggle?

    Is the evidence more to do with the rise in general economic wealth – which aligns with growth in education outcomes. With the influence in both directions – the education helps produce more wealth; more wealth allows expenditure on education (and other services).

    To extent it is signalling then should people do what interests them over do what a profession requires?

  2. albert wright says:

    Trying to measure the “value” of ” formal education” at different stages of human development for different individuals at different ages is a very difficult task.

    Firstly we need a clear definition of “value” and “formal education” and a clear understanding of how the measuring will be done. If the purpose is to measure the added value of HE, we need to know the starting value possessed by the individual (including how that value has been measured to date) and what the outcome targets are that relate to the value being measured at the end of the educational period.

    Starting with the entry to University, the value being measured is in the currency in terms of the UCAS points that were required to gain entry.

    At the end of 3 years the currency being used is the class of degree (how that can be related to UCAS points I am not sure).

    However, this currency (Class of degree, is complicated by study subject and rating of University – in a world where there is no nationally recognised currency as each University decides what constitutes a first or second class degree.) If this has any benefit at all, it can only be understood by the individual student and University.

    In addition, we are now in a different “country” the buyer of the graduate product – the new employer- uses an entirely different currency, namely money, expressed by the salary offered to the new graduate employee. The employer has a flawed “currency converter” when turning the degree class to money and many employers have abandoned this measuring method and now prefer new currencies like psychometric profiles, interviews, group observations etc.

    When it comes to “society” and measuring the added value of HE education we use a different currency again depending on what outcomes we have as targets.

    The concept of “Signaling” and the use of Table 1 above, whilst interesting, seems to me rather inappropriate in the circumstances.

  3. Chris Percy says:

    Conor, thank you for these thoughtful comments. Much of what you say resonates with the research. The general theory of signalling applies to all levels of education; we focus on higher education in the note because it’s where the policy challenge is most live in the UK (no-one is calling to defund primary).

    The importance of reducing drop-outs is clear in a number of studies (e.g. the Futuretrack longitudinal study in the UK and the extensive studies in the US, where drop-out rates have typically been higher than the UK). And as you say, where many people have a certain qualification, the stronger signal relates to those without it.

    Your point about the bidirectional relationship with wealth is also picked up in the full paper – such features make it hard for macroeconomic research to disentangle the productivity value-add of education from the goal of investing in education as a good in its own right, using the wealth generated by other activities.

    Your final point hits to the core of the issue. It may be rational/sensible for individuals to invest in signalling if that’s what their chosen profession expects. But it’s adding little value societally and can lead to an education arms-race. if those signals could be obtained more cheaply (not more easily) or employers could find better ways of testing those prior abilities, then everyone would be better off from a financial point of view. Our paper is arguing that it is worth investigating whether there are ways of saving this money – even if we’re not certain whether they might work at this point.

    (Chris Percy)

  4. Chris Percy says:

    Dear Albert, thank you for engaging with this topic. You raise important contextual questions about this avenue of work. It sounds like the efforts on “learning gain” might resonate well with your thinking around value/formal education. It’s proved tricky for the sector to make progress on learning gain, but we hope that seeing it as one way of responding to the signalling critique will help motivate renewed efforts in this direction. I like your metaphor of currencies and imperfect currency conversions. In the report, we also call for employers to build on the experimentation with different hiring methods – crude degree/institution based measures still appear dominant for much knowledge work in the UK at least. (Chris Percy)

  5. albert wright says:

    Thank you.

    As they say, “Every little helps”.

    Employers are making slow progress with new hiring techniques like the psychometric tests and many more are looking at applications blind – without names or ethnicity.

    Historically, our elites came from kings and their family crowds. Being good at killing people and winning battles against those trying to take stuff from you were the key skills required to succeed.

    There will always be leaders and followers and those with more of the talents and skills than others in an ever changing world.

    Cream rises to the top. I am not sure that holding a specific degree from a specific university at a specific time is the cause of success and high income but it is a relatively easy way to make hiring decisions without too much effort or disasters.

    Long after they ceased to be the best, buying a computer from IBM was considered good practice.

    For me, any causal connection between HE and success remains thin.

  6. Conor King says:

    Albert it is clear most positions are not the top – if you define success as such then no education system is about that group alone.

    From your analysis it is about how capable and productive the everyday soldier of fortune is, the workers in the company, the small self run business.

    It used to be easy to claim a degree was ticket to being on top – now that most have a ticket of some sort, we can analyse about the ongong heirarchy or attempt to consider whether the group as a whole and most within are better off than previously, and why.

Leave a Reply

Your email address will not be published. Required fields are marked *