Econometric Optimization of Fresh Start Policy

"What is the optimal consumer bankruptcy law?" Now that's an abstract first line that grabs my attention! I've thought about this question for most of my academic career, and I've struggled to find solid bases for an answer. Now, Indiana Univeristy economist Gray Gordon offers an intriguing if difficult to understand possibility. In his paper, Optimal Bankruptcy Code: A Fresh Start for Some, Gordon actually quantifies the sweet spot: (1) an optimal system offers a discharge of debt (a constant refrain in policy papers, e.g., here and here), (2) it does so for households whose debt is 2.6 times their endowment, and (3) this optimal system results in a welfare gain of 12.2%. The conclusion is nowhere near as confusing for a non-economist (like me) as the proof, expressed in inscrutable Greek-symbol-filled equations which occupy the bulk of the paper. But this paper offers a rock solid answer to a question that has plagued Europe, in particular, for many years--how does one define "overindebtedness" and therefore the proper entry criterion for personal debt relief. Gordon's answer is very powerful, though I wonder how compelling the econometrics are behind these hard numbers. I'm not at all qualified to critique Gordon's proofs, but I'm a bit skeptical in light of Gordon's observation that "[r]elative to the U.S., the optimal policy results in a four-fold increase in the bankruptcy filing rate and a thirty-fold increase in debt." Whoah! Anyone care to comment on what could be a really groundbreaking new approach?

Comments

[mt]

Thanks for calling this to the audience's attention. The degree of specificity --2.8 -- shuld make one skeptical. All stylized models like this are extremely susceptible to gerrymandered assumptions, usually signaled by words like "for simplicity, I...". Two that jump out at me from his model are:" ... full information. Because of it, there are no bad borrowers: Borrowers prone to default are known and the risk is perfectly priced."

Um, no, not in any world inhabited by carbon life forms. Especially because his model also says the planner need not impose a bad credit record once people have filed.

And: "For simplicity, I assume households
may not default on savings > 0."

Again, it's his model but in the real world, unless you eliminate all exemptions, there are in fact debtors with assets having a positive value.

Last example: here is the population of his world:
"To simplify computation, the baseline model period is taken to be three years (as in
Livshits et al., 2007). The risk-free price q is set to give a 2% annual real interest rate.
... Households age 1 (real age 23) to R􀀀1 (real age 61) have one endowment process,
representing working age earnings risk, and households aged R (real age 64) to T (real
age 85) have another, representing guaranteed retirement income."

Yup, that is certainly simple. And not remotely like the world we find today. Right now the risk free rate for 3 years is about 60 bps. In past years it's been 10X that. And amazingly enough, the earnings power and spending obligations of households varies not just by age.

I imagine someday some good might come out of a stylized model, but if I had to promote someone to an economics faculty, I would focus instead on someone doing big data analysis of the real world -- I think s/he will produce more useful results sooner.