https://smile.amazon.com/dp/0691174652
People with more education earn more money. How might the causal structure look?
- Education works ie education => ability => money
- Education signals ability AND ability leads to more money ie ability => education, ability => money
- Education signals ability which leads to more money ie ability => education => money
Obviously, none of these are 100% true, but we can talk about the relative contribution.
The book beats up on 1 pretty early eg:
- Most high-school graduates don’t even recall elementary science or history facts
- Lambskin effect - most of the bump in money comes from graduating - if you quit in the last month you might as well not have gone
- College students typically go out of their way to avoid learning, gravitating to subjects with easy passes - the question is always “will this be on the exam” not “will this be in my job”
It acknowledges that some core skills (reading, writing, arithmetic) are effectively learned. Also that the contribution may vary for technical subjects (eg engineering), but even there the teaching methods are inefficient and there is a large opportunity cost vs eg apprenticeships.
The bulk of the book is spent arguing for 3 over 2. The core idea being:
- Traits like conscientiousness (will this person slack off after we hire them) and conformity (will this person follow the rules) are hard to detect before hiring
- Firing people is expensive and unpleasant
- Companies have strong incentives against giving negative signals (eg giving a bad reference) so even if you get managed out you can still be successful in the next job hunt
This means that if you can fake the signals to get in you can make it in the job world - you won’t be detected and filtered out. 3 makes education a zero-sum competition to produce the best signal, leading to a successful arc relatively independent of actual ability.
But the bulk of the argument doesn’t seem that important once we’ve discard 1. 1 is the sole argument for education in it’s current form. If all we care about is signaling conscientiousness we can just make people copy books by hand for four years at much less expense to society.
A more interesting question is how the signaling argument relates to new schools like Minerva. Is it possible to improve education, or does signaling entrench the current players?