Student Debt and the Freedom of Choice
This was meant to be a short post. But it got a bit longer.
A friend suggested I listen to a NPR podcast on student debt, called Throughline. The podcast created a wonderfully interesting background to the current government provision of student loans. It provided lots of background on institutional history and the evolution from government loan guarantees through SallyMae to the current program of direct loans. Most of it derives from the Higher Education Act of 1965 and its many amendments.
But there is an obvious undertone of discontent. At some point, the word “forced” is used to describe a decision to take out a loan. This perspective on student loans bothers me, quite a lot.
First a tiny bit of economics and then the words to go along. Here is the model. Someone is choosing how much education to obtain, with education represented by e. Education has a return in terms of income, represented by w(e), if you get a job commensurate with your education. Else, you get a job which requires no skill with compensation w(0). To make matters simple, this of these income as being the discounted present value of your lifetime earnings in each period of your working life.
The key thing is that the type of job you are going to get given your education is random. It is not known at the time you choose your education. To make that concrete, let q represent the probability that you get a job that fits your educational attainment.
With 0<q, there is a chance to get a job that fits but as long as q<1, there may be no return to education.There is a cost of education, which we represent as c(e). This notation just means tha the cost of education depends on how much you choose to acquire. The cost includes tuition and the opportunity cost of time: i.e. lost wages while attending school. So costs would naturally increase in e.
The model does not have student loans, just this cost of school that includes tuition. If you were to borrow to pay tuition rather than paying it directly, then c(e) would capture the discounted present value of the payments made over the life of that loan. If the banking industry was competitive, then this discounted present value would be the same as someone paying tuition out of pocket at the time of attending college. So it is ok for now to think of loans as underlying the cost of education function.
In economics we focus on the choice problem of an individual or a firm or a government. Here the education choice is made by individuals (or groups such as households). Think of that as someone choosing the education level e to maximize
qw(e)+(1-q)w(0)-c(e) The first two terms represent the expected discounted value of income, which depends on education. The last term is the cost. So the objective here is to maximize the net benefit of obtaining education.
The optimal choice will balance expected marginal benefits and marginal costs. The expected marginal benefit of education is that with probability q you get a job at a higher wage, as we assume that wages increase with e.
Just a small aside on this last point. It seems like an easy matter to say that wages increase with education. But to actually quantify this is very difficult. The reason gets to the heart of an important challenge in economics. We do not run experiments like, say, a biochemist might. That is, we cannot randomly assign some people to college and others to no-college and then study the income differences created by this assignment. Instead, as in this model, education is a choice. So to study the return to education we have to study the choice of education at the same time. Or we have to find a way to interpret the data as a natural experiment whereby people are randomly assigned to college. Either way, this remains a challenge for economists.
With this choice problem in mind, let’s get back to student debt. The point is to understand this notion of being “forced” to do something, like to go to college and accumulate a lot of debt.
From the perspective of this problem, it is easy to understand the situation of someone who went, let’s say, to law school, and was unable to find a job as a lawyer. As long as q<1 (even for lawyers), there is a chance this might happen. On top of that, imagine this individual had taken out a loan to finance law school.
If you met and interviewed that person today (the press is full of these examples), then for sure individual may regret the decision to attend law school and/or the decision to borrow. But this is a ex post (after the fact) regret: if only I had know I would not have obtained a job as a lawyer, then I would not have gone to law school.
Sure, fine, but that type of foresight was not available when you made the education choice ex ante (before the uncertainty was realized). Suppose that this individual did not have a lot of resources available at the start of law school to pay tuition and that is why they took out a loan. Then if you ask this person would they have been better off, again ex ante, if the opportunity to take a loan was not there, they would say no. Likewise, if you asked them if they wish they did not have an opportunity to choose to go to college, they would also say no.
Of course this must their answer as they benefited from the opportunity to go to law school and to borrow to pay tuition. How do you know that? By a simple revealed preference argument: they chose law school over other options. So, despite their ex post regret, ex ante they would not be in favour of eliminating the option to borrow to finance tuition and living expenses.
For those who took the gamble and won, there is still debt to be paid. And sure, like anyone else in debt, they wish they did not have to make those payments. But again, they chose to borrow and obtain the education that now supports their relatively high income. So we are back to freedom of choice and revealed preferences.
So then what is all the fuss about? Is it just ex post regret because someone took a gamble and lost? There must be more to this story.
Well what about q? Where does it come from? In Economics we often resort to a connection between beliefs and reality, called rational expectations.
In this context this means that the individual’s choice is based upon the same q as the actual outcomes in the economy. So if say 85% of individuals with a certain level of education get a job commensurate with their educational attainment, then everyone deciding on education would incorporate this 85% into their assessment of the marginal benefit of that level of education.But this is for sure not always the case. There are examples and remedies directly associated with charges of colleges overselling the returns to a degree and student taking out loans based on these promises. In our little model, this is just convincing someone that q is higher than it is in reality. It is like being in a casino and being led by a host to a high stakes blackjack with a guarantee that you will come out a winner! Or one of those cold calls from a broker with a sure winner of a stock to buy. Yeah, right.
But this is fraud. It is not a problem of student loans per se but rather illegal activity surrounding those loans. This is an argument for better enforcement of laws to prevent this type of behavior. It is not an argument to shut down student loan programs.
In the end, there is a simple enough economics point. Except for the problem of fraud, the outcome we see of individuals with large quantities of student debt is a consequence of individual choice. These decisions involved uncertainty and sometimes the outcomes are not the most desirable ones.
But the gamble was worth it, otherwise you would not have made that choice. Preventing you from taking that gamble is not in anyone’s interest.
Here I am being a little vague about whether or not q depends on e. And also I am being vague about whether e is continuous or from a discrete set. I hope those details matter for the points I am trying to make.