As opposed to the ATL article, I don't want to call people stupid - maybe "not paying attention" or "not understanding the consequences" would be better phrases. However, when one looks at the data, it really does seem pretty apparent that there is an increase in the asset price (college degree) that has become decoupled from the underlying economic reality - at least for most college degrees. The article cites the recent decline in starting salaries for college graduates vs. the increase in tuition in this regard.
Further, it also seems clear that there is acknowledgement of the risk by the agencies evaluating risk. As the article points out, Moody's Analytics has pointed out that:
"many students will be unable to service their loans as income growth falls short of borrowers' expectations."and
"Fears of a bubble in educational spending are not without merit"This is different from the situation in the housing bubble where there really wasn't much appreciation of the system risk until it was pretty much too late.
What are the consequences of ignoring the problem? The article cites a study from the Chronicle of Higher Education that finds that 1 in 5 government student loans that entered repayment in 1995 has since gone into default. I remember 1995 - 1995 was not that bad a year, economically speaking. We were coming out of the recession in the early 90s. It wasn't go-go time, but not bad. Additionally, student debt was much, much less - as recent studies point out, college costs have tripled from 1990 to 2009.
So if a 20% default rate is what you get in 1995 when you have a) better overall economic circumstances than now, b) less than one third the cost of going to college, and c) increasing starting salaries for college grads rather than decreasing salaries, what is the 2011 default rate going to be? Absent some other factor, it would obviously be more than 20% - likely far, far more. Maybe as high as 50%. That would be terrible and it would be quite obvious that our student loan system was in serious need of reevaluation.
However, what about those "other factors"? Well, as LSTB has pointed out to me, most federal student loans given today qualify for Income Based Repayment (IBR) which will allow those with student loans to not pay anything if they are not working (in certain situations). Additionally, the loans may be forgiven after 25 years (although there may be a large tax consequence if current law is not changed). The forgiven loans would be a taxpayer expense.
IBR does strange things to the "default rate." Suddenly failing to make payments to your student loans because you are out of work is not "default" - instead it is just "participation in the IBR program." Consequently, (as governments measure things) the "default rate" is likely to be calculated only based on those who are somehow not included in the IBR program. For example, students have to take affirmative steps in the IBR program to establish how much they are paid - failing so, the remaining debt is amortized over 10 years and students have to pay the (typically higher) payment. I suppose it would count as a default if someone failed to notify the IBR program and then ignored the new letters - and I am sure that of the millions of people taking the loans there will be a percentage that will do so. However, based on the difficulty of actually defaulting under IBR, I would put that percentage at about 2%. Frankly, under IBR, "default" no longer has anything to do with the student's underlying economic situation.
That is, IBR effectively hides the default rate. It makes "default rate" into something that is no longer a useful barometer of the actual economic outcomes of people taking student loans. This is by no means a scientific measurement, but the 80% non-default rate for the class of 1995 leaves one with the impression that college was a successful investment for 80% of those taking loans in 1995. It's a gross, intuitive measurement, but it likely has at least some correlation with actual outcomes.
Alternatively, IBR pushes the accounting acknowledgement of the loss of money to the program due to lender non-repayment to the 25 year mark when it would presumably be possible to determine the amount of student loan debt forgiven. And here's the big fat stinker - we have $110 B increase in student loans over the last 3 years ($36.6B/year). Based on the current bag economic situation and the increased cost of college relative to earnings, it is highly likely that we will have a default rate greater than 20% - and I will estimate a reasonable value of 50%. That would mean at least about $7B - 18B/year - and that's just for ONE YEAR.
Additionally, the actual amount forgiven would be considerably higher considering the 7.9% interest rate attached to the loans. (People may or may not want to include the total loan amount because can we really consider losing money that we never had to be a loss? That is, when recognizing loss is our "loss" just the loan capital given or does it also include the interest we never got.) Regardless, we are going to be looking at a big, big loss. However, the loss may not be acknowledged for decades.
So, the answer to the question "When is a default not a default?" Answer - when it's an IBR.
Update: Here's another article on the debt crisis from The Atlantic. They also draw parallels with the housing bubble.