A Note on Signaling in Education

Signaling is not enough to make claims about the social optimum

The usual way that people think about the implications of the return to education being due to signaling, as opposed to human capital accumulation, is that the greater the share due to signaling, the less we would want to subsidize education. The value of a signal is diminished if everyone has it; we are stuck in an arms race of acquiring more and more degrees. 

It was therefore a bit of a shock to learn that there are instances where the whole return of a degree can be due to signaling, and yet attaining the social optimum requires more subsidies for education. That is the point of Joe Stiglitz’s 1975 paper “The Theory of ‘Screening’, Education, and the Distribution of Income”. He argues that there is no reason to expect free individuals to come to the optimal level of screening, and depending on the case that can be above or below optimal. It’s possible that there is no equilibrium whatsoever! This really shouldn’t be all that surprising, to be honest. Quality of worker is information, and information breaks our rules for perfect competition. 

A quick example from the paper. Suppose that productivity is reduced by variance in worker’s abilities. One can imagine a production line where each task is dependent on a step done before, and if workers are of varying speed, there will be wasted ability. Everyone would prefer that variance be reduced, but no individual would find it worthwhile to take on the cost of screening. The employer could screen, but they are not the whole market – any other employer can copy the information discovered by the first employer. 

I think this should clarify that there are some unexamined presumptions in the case against education. The subsidization of education has led to more people passing, but it needn’t. It is not sufficient to show that education is due to signaling, but also necessary to show that standards are being lowered to pass more people. I do think this is broadly true, of course – more people have a college degree now, at the same time that the average IQ of people who are in college has fallen. Unless there has been a great revolution in our ability to impart information and skills to people that I missed, I have to believe that standards have been lowered.

And we shouldn’t expect government to fill in the hole perfectly. Later on Stiglitz discusses what might happen when we have a majority voting system deciding how much screening to do. We will assume that everyone votes in their self-interest, in order to be able to say anything meaningful; if this is the case, then if voters are perfectly informed of their own abilities and the distribution is skewed, they will vote alternately for too much or too little screening. There may be no equilibrium, as before education the median voter received the average product. As screening increases, if the average is higher, their income may decline, and then increase later as productivity improves. As Stiglitz writes, “In recent years economists have shown an increasing awareness of ‘market failures’ and have increasingly called upon government intervention to correct these failures. But to turn over an allocation process to the public sector is to make it subject to ‘political laws’ which may be no less forceful – and even less efficient – than the ‘economic laws’ which previously governed the allocation process. The fact that these political laws are less understood … than the corresponding economic laws is not an excuse for relying on the mythical ‘benevolent despot’ who plays the central role in most economists; models of the public sector”.