One of my fellow grad students just sent me the following email:
[We] were discussing whether the ability to option into a loan where you legally cannot default (like student loans) on all loan types (car, home, etc.) would be a good thing or bad thing. We both decided that we didn't know, but that there would be potentially really bad sorting mechanisms put into place. Do you have any thoughts on this?
Once I started typing my answer, I realized it would make a good blog post, because it illustrates a lot of important topics. This is a question where different assumptions will give us very different answers. Economic theory can tell us what will happen under different assumptions, but only empirical data from the real world can tell us which assumptions are true.
One set of assumptions is that people are rational and that the default risk is mainly based on the character of the borrower. In this case, such loans would only be a good idea. They would give people the option of credibly revealing their type. Banks would charge lower interest rates on the loans where people could not default, because those loans loan would be less risky. If you were a responsible person and you knew you would repay, you would sign up for the no-default option to get the lower interest rate. If you thought you might default, you would pay a higher interest rate for the right to do default. The current system is a pooling equilibrium, where everyone with the same credit score pays the same interest rate. This new system would generate a separating equilibrium, where different people make different choices and pay different rates. This is the 'sorting mechanism' that my friend mentioned. The current system means that good borrowers pay for the deadbeats, and this option would eliminate that. It would generate a higher interest rate for deadbeats, meaning that they would take out less loans and less resources would be wasted on them.
But if money problems can happen to good people at random, for example if they get laid off because of a change in technology, then the situation changes. There are good reasons that our society got rid of debtor's prisons and allowed people to go bankrupt and eliminate their debts. If you suppose that everyone is the same, and faces some small chance of job loss or other misfortune, then the ability to default is best seen as a kind of insurance. If bad things happen, you get to erase your loan, which is like giving you money. We know that buying insurance will improve the expected utility, or well-being, of someone who is risk-averse, and most people are like that. The current default rules mean that everyone gets insurance, which makes society better off. If people had the option of not buying this insurance, then most of them would not bother. Everyone would take the option to default.
Both of these assumptions are simplistic. In reality, 'good' people have a small chance of default and 'bad' people have a larger chance of default. Modeling this situation properly would require some heavy-duty economic analysis and game theory, and I am not going to spend the time to do so, but I have seen similar situations and models in my Financial Economics class so I have a good feel for what would happen.
The current pooling equilibrium gives valuable insurance to everyone, while also causing good people to subsidize the deadbeats. A separating equilibrium resulting from the no-default option would erase both of these effects. Good people would no longer pay for bad people, but if misfortune hit, they would be in very bad shape.
However, we know that a separating equilibrium can only be generated if the 'good' people are better off choosing the no-default option. If the insurance is valuable enough to them, they will choose the default option, even with the higher interest rates that result from being pooled with the bad people.
It is hard to say if a separating equilibrium would emerge. It depends on how much people value the insurance and how many 'bad' people there are. But I am pretty confident that if a separating equilibrium did emerge, it would be better for society. Under rational actor assumptions, giving people the choice of a no-default loan will allow them to be better off, even though they would lose their 'insurance'.
However, this conclusion changes once you add in psychological factors like the overconfidence bias and limited information. This means moving from neoclassical economics to behavioral economics. If people underestimate the chance of misfortune happening to them, then they will not buy insurance when it would be in their best interest to do so. In this case, too many people would choose the no-default option, and end up suffering when misfortune hits and they cannot erase their debts. If you believe these assumptions, the current 'paternalistic' system where everyone must buy insurance with their loans is the best system.
The best option for society would be one that kept the insurance but eliminated the subsidy to 'bad' borrowers. This could be done by the creation of more complete insurance markets. If everybody bought layoff insurance the way they buy house insurance, then the no-default loan would not have any disadvantages. When good people got laid off, they would be able to use the insurance payments to keep paying off the loan. There would be a separating equilibrium where 'bad' borrowers pay more, and society would be better off.
This is another reason to remember the targeting principle. If we want people to have insurance against misfortune, then we should provide it directly, rather than bundling it with loans. Forcing the loans to come with insurance generates unwanted side effects in the form of subsidies to 'bad' borrowers. Our bankruptcy laws were written before most individuals had access to insurance. Before the 1800's, individuals did not buy insurance of any kind. There was no market for it and there were no companies providing it. Now that this has changed, our bankruptcy laws may be obsolete. It is a very good thing to give people insurance, and the side effects were unavoidable back then, but we can do better now.