Why do people pursue academic courses that in no way set them up with the skills needed for their future careers? One obvious answer might be because they enjoy the topic they study. Education is a good thing in itself. I’m all for that. Yet, that seems to be far from a complete explanation. Maybe it is just me, but occasionally I have students in my classes who don’t seem to find listening to me what they were born to do. What else is non-vocational education for if it isn’t for the sheer joy of it? Here we can turn to work on signaling and education.
Signaling And Education
The challenge of signaling goes well beyond the field of education but work, such as the classic 1973 Michael Spence paper that I discuss in this post, often focuses on signaling in education. This makes sense. There are reasons to think signaling is especially important in this field. (And academics know a lot about the field).
Signaling solves the same problem as a person’s reputation. How do we know about another person when what we are looking for is not obvious? If you already know people well they don’t really need to signal as they have a sufficient personal reputation. Consider a famous actor, we already know their work well. As such, no one needs to ask what acting school they trained at. We already have enough information to know whether we think they are good or not.
Signaling matters when you really don’t know much about the person you are meeting. A natural example is a job interview, especially for more junior posts. You get lots of candidates to interview. You can’t really get to know them all too well. How then does a candidate signal (consciously or unconsciously) that they are the best choice?
…the primary signalers are relatively numerous and in the market sufficiently infrequently that they are not expected to (and therefore do not) invest in acquiring signaling reputations.
Spencer, 1973, page 355
Signals And Indices
Signals are distinct from indices. Unalterable attributes are indices. You are a certain height. It is hard to change that so it doesn’t reveal much about the decisions of the person.
A signal on the other hand can be sent, or not sent, at the discretion of the sender. The presence, or absence, of a signal can tell us something. The sender can choose to go to university. These signals can help the receiver (for example, the hiring manager) determine something about the person they are interviewing, or even merely thinking of calling to interview. If you could ask the person and they would always tell you an honest answer then you wouldn’t need signals. Signaling applies when it is hard to trust the sender, here the job applicant. Who is going to honestly say that they really are a bit lazy at an interview? As such, the hiring manager needs something harder to fake.
What Makes For A Good Signal?
The key consideration is that it can’t just be any random signal that is sent. The signal must be relatively cheaper to send for those who will be good at the job. This is so those who find it cheap to send the signal will be more inclined to send the signal.
Why signaling and education? This works if the ability to get an education correlates with the ability in the workplace. This is not an outrageous assumption.
- Those who don’t mind working hard will likely do better at university.
- Those who don’t mind following rules (even if the rules are sometimes a little silly) will likely do better at university.
- Hopefully, those with stronger intellects will likely do better at university.
The point is that if those candidates with the above characteristics are what you are looking for then looking for graduates is often a good idea. This is true even if the graduate didn’t learn anything relevant during their degree. (My Hellenistic Studies M.A. didn’t obviously prepare me for my career in accountancy).
The good news is that the cost of the signal encourages honest signals. (Cost is defined widely, including psychological costs; discomfort, unhappiness etc..) Those who, for example, hate hard work will find it much more challenging to get a degree. They will simply not see getting a degree as worthwhile, given how painful the work is for them to get the degree. This is despite the fact that they could get a higher-paid job with a degree. Better for them to take a lower-paid job, but without the cost of the degree.
Clearly, this isn’t the only reason people don’t go to university. E.g., some who would do well at university never get the opportunity. While this suggests actions are needed for society to improve access it doesn’t break signaling theory which suggests that the signal works if there is a correlation. The theory doesn’t need there to be a perfect correlation, i.e., that all who go to university will be better at the job than all who do not.
Does This Mean Education Doesn’t Benefit The World?
This leads to questions about whether education is a social benefit. Education may not teach useful skills. It might just be a test of resolve that can lead to some people, who get it, having an advantage over those who don’t. Merely determining the order of who gets what reward is not a wider social benefit. (For a discussion of this from a pessimistic viewpoint see here). In general, I wouldn’t want to press too far on that idea though. Not all non-vocational education is completely useless in people’s lives beyond academia; skills can transfer. What is more, some students will get pleasure out of the process of learning, raising the world’s happiness. Basically, I don’t think merely redistributing benefits is the vast majority of what education does (although there is likely some of that).
I do have a positive spin on the idea that education can benefit the student, merely because it is hard, even if the education is pointless. As I tell my students every year, they gained benefits just by attending class and showing that they were willing to put up with me. If they actually learn anything worthwhile that is a bonus. It certainly takes the pressure off the teacher to think of education that way.
Read: Michael Spence (1973) Job Market Signaling, The Quarterly Journal of Economics, Vol. 87, No. 3. Aug., pp. 355-374