Buying Debt as Effective Altruism?
A collection of Occupy activists recently bought over $14,000,000 in personal debt for $400,000.
Normally, debt-buying companies do this with the intention of collecting the money from the debtors—Occupy did not, and I was struck by the lopsidedness of the figures.
A number I see often in the high-impact philanthropy world is $2300 to save a life (with plenty of caveats). At Occupy’s rates, that would buy roughly $80,000 in debt—enough to get two or three families out of a hole that would otherwise render them bankrupt.
By itself, this isn’t enough to be better than mosquito nets or deworming. But the thing about personal debt is that, thanks to interest payments and stress, it prevents people with high earning potential (compared to an average African) from making decisions that would optimal were they debt-free—like finishing college or buying a used car so they can take on a higher-paying job.
My idea, though it’s a tentative, spur-of-the-moment thing:
Why not found a charity that acts like a combination of Vittana and Giving What We Can, freeing people with good prospects from debt in exchange for their signing a contract to donate a small portion of their future salary to charity?
A few issues that come to mind:
1) Occupy bought a lot of medical debt, which this company wouldn’t, and other types of debt might be harder to buy.
2) People who have decent earning potential have more valuable debt, since they’re more likely to pay it off later. (On the other hand, freeing them of interest payments might help them get into a better position for repayment.)
3) The idea is a lot like micro-lending, and organizations that offer that service don’t have a great track record (though some have been successful).
4) People just freed from debt might not be in a position to donate much salary/might be unreliable. (Deferred payments until college is finished/the new job is had could be helpful here.)
5) There might be (well, almost certainly are) difficult legal issues with finding information on people in debt before you actually own their debt.
Are there any other obstacles you all can think of? Other features of the charity that might make it more effective? How does it sound as an intervention that increases the world’s productivity in the long run, stacked up against other such interventions?
The argument for microlending is “access to credit allows people to invest in themselves and their business, but the cost for interacting with people on the $100 level is such that it eats any profit, making banks uninterested.” Thus, the service can be run primarily as a nonprofit, or by offloading some of the administrative work and risk to the borrowers (i.e. the lending circle idea).
That’s because the market expected that the net return on the debt would be about 3%. So it might look like you’re saving people from $80,000 of heartache, but you’re only actually increasing their financial prospects by about the $2,300 you transferred to them. Yes, they don’t get called by debt collectors, and yes, they don’t go bankrupt- but artificially inflating someone else’s creditworthiness is a problem for them and everyone else. It doesn’t seem like you could get more than $2,300 out of them long-run, because otherwise the debt collection companies would be able to get more out of them!
Or… taking out more debt!
When helping people to get out of their problems, there is always a risk of perverse incentives. (Should I take a lot of debt, hide the money, and contact the debt-buying charity?) Which reasonable steps could decrease the risks?
A policy of helping each person only once in a lifetime. If they take more debt, it’s their problem now.
Selecting the people to be helped randomly. For example, if we are able to help n people, we choose 2n best candidates and then help a randomly chosen half of them. This could discourage the “get into debt, they will help me” scenario. -- Okay, I am not sure about this. A stupid enough person would take the risk, precisely because they are stupid.
Mandatory participation in “elementary rationality” exercises which would teach people to not take more debt. Increases the costs of the program.
Any other ideas?
If you’re going to randomize the recipients, you should also track them to see what effects the debt forgiveness has.
The OP seemed to be hypothesizing that the mere fact of being in (so much) debt was diminishing productivity; it might be that you can get more than $2300 of value out of them if, say, their being inspired to work for some noble cause (or for higher-status forms of self-interest than debt reduction) makes them work oodles harder or smarter than if they’d been doing analogous work to pay off debts.
It also costs debt collectors some amount of money to collect debt. Presumably, if a business buys debt from a bank, it’s because they think they can collect the debt for a non-trivially lower cost. Otherwise, the bank wouldn’t be willing to sell the debt.
I don’t know how significant this is.
For the same reasons that no one wanted to buy the debt for 13 million dollars, that apparent 13.6 million dollar “profit” is largely imaginary. Debts that are unlikely to ever actually be repaid and so forth. There’s probably some profit to the system in terms of administration and debt collection that goes unhandled but that’s really charity for the previous debt owning institution than it is to the people who are in debt. As you noticed, people who are good investments (eg likely to pay off their debts and then be profitable) are not going to have debt that cheap.
On the whole this seems strictly worse than well-aimed scholarships.
Even if you don’t repay the debt having no debt collectors that go after you is valuable.
How valuable?
You might decide against taking a legal job if you can count on the fact that all income will go to debt collectors anyway.
Not in any real world I know. Indentured servitude has been abolished and I haven’t seen any debtors’ pits around lately.
Specifically, the only problem is debt not dischargeable in bankruptcy, that is, in the US, just student debt. If you have, say, credit card debt—just go declare bankruptcy, I don’t see any useful consequences of a charity helping you avoid the costs of going bankrupt.
So, student debt. Is there empirical evidence that a lot of people choose not (legally) work because they have high student debt levels? Does that happen in reality? Is it a really big problem worth focusing one?
As a practical matter, also note that there are legal limits to how much of your salary can be garnished for unpaid debts. It is not 100%.
Well, the IRS does things like this, but their debts aren’t for sale.
In the United States, only 25% of after-tax income on income over the equivalent of a thirty-hour minimum wage workweek, excepting alimony, child support, tax debt, and certain chapter 11 bankruptcy court agreements (none of which are types of debt a charity can purchase for pennies on the dollar).
Most types of debt that can be purchased for pennies on the dollar are also dischargible through bankruptcy, and have not reached the garnishment stage or reached the garnishment stage in ways that mean very little income is going to debt collectors.
I’m not aware enough of other national laws to comment on how things look elsewhere.
There’s some value to it, but you’d have to target the matter very carefully to get reasonable returns.
I agree that’s it’s not easy to decide. If one is interested in spending money that way you would have to see how many people get free from debt and their more specific life situations.
Cannot upvote this enough to show how much this changed my mind: OP now seems like a clever solution to the Parfit/Newcomb-type problem outlined in parent. Can we find this problem elsewhere in economics?
If nothing else, the value of the time of the debt collectors. There’s probably something useful they can do instead.
Become TSA agents..?
At least within the United States, forgiving debt counts as a taxable event, on the year that the event occurs. This Occupy group has largely sidestepped the issue, and the IRS is unlikely to audit the beneficiaries in this particular case, but the larger the system becomes and the longer-term the system becomes the more likely your patrons will be looked at.. If you’re not very careful about how you build the forgiveness, this can even push people into higher tax brackets or out of welfare benefits that they need to live. See some criticism of the Bush-Income-Based-Repayment plan for details on this matter.
Debt collection companies are also likely to adjust the cost of debt around your purchasing power, if your charity persists for long, which may slightly decrease the rate of return. ((Contrawise, there are reasons to keep uncollectable debt on your own books, even if it’s obviously uncollectable, due to current accounting practices.))
To expand on your point four, the debt is valued at less than its paper number because these people are the ones that are generally hardest to discuss matters with and/or get money from. This may not always be a matter of income, but a matter of the people not having reliable contact information or living in areas where the courts tend to be very suspicious of debt contracts. Garnishing someone’s wages involves drastically different costs depending on your state.
You’d also need to get a fairly good rate of return for this to be the more efficient than typical investment strategies. If eliminating the debt it itself a value thing, that’s less of an issue, but it’d be tricky to evaluate that. If you want too much of someone’s future wages, they’re likely to discount opportunities to increase their future wages (especially due to how marginal tax rates go around the poverty line to four-time the poverty line).
+1
Under the US tax law is someone forgives your $50,000 debt, that $50,000 is added to your income and you owe all taxes on it.
Yes, but having the forgiveness happen in a low-income year would result in less taxes. So perhaps the charity could forgive debt in a way that is conditional on later income being donated to effective causes.
In the United States, marginal tax rates at the poverty line can, due to welfare or tax deductible cutoffs, be pretty punishing or even negative. Worse, because this income does not go through conventional channels, it can cause someone’s annual tax return to suddenly switch from a rebate to a payment—a payment that is difficult to delay and near-impossible to escape through bankruptcy, and that can only be challenged in the IRS’s very own home courts.
The rate is likely to go so blatantly negative in countries that lack phaseouts or means-tested welfare, but even those exceptions will still have that surprise tax bill.
It’s not a fatal problem, and there may be ways to limit or negate it—although any serious attempt will face serious and unfriendly tax scrutiny—but it’s a very significant problem.
Lots of U.S. government benefits for poor people are tied to income and wealth.
Seems like the only difference between buying someone’s debt and paying it off for them vs them just defaulting is that if they just default, their credit rating takes a hit. So it seems like the major side effect of this scheme would be to make credit ratings less reliable.
My understanding is that people who use credit ratings aren’t interested in defaults on medical debt. Thus this project makes credit ratings more reliable.
OK. But even if that were true, that would essentially be the main net effect of the scheme, right?
So, let’s see.
First you need to identify people who defaulted on their debt payments. This is easy.
Second, you need to filter them and select “people with good prospects”. This is considerably harder. I am not even sure what does that mean. Presumably you’ll ignore the credit card debts of a black single mother of nine kids but will buy the student debts of white suburban girls graduating from college?
Third, what you are doing, in economic terms, is refinancing their debt. You give them the lump sum now in exchange for a promise of a series of payments later. That’s just a normal loan transaction. What makes it special—that you will be willing to loan money on favorable terms to people already in default? Well, I guess, but that doesn’t strike me as particularly effective or increasing “the world’s productivity in the long run”.
I wonder if one could focus on something that often goes untapped like innate programming ability. Have the person take a test that sees if they can learn to program, and if they can, forgive their debt and enroll them in a program to train them and get them employed.
I don’t understand why forgiving someone’s debts is necessary for giving a test for programming ability and enrolling him into a training program.
So maybe hold onto the debt indefinitely and offer to forward any repayments to charity? That might work, but it seems like if their income increases later, it might not be as advantageous to forgive it then for tax reasons. Also, there might be a goodwill factor associated with debt-forgiveness that isn’t there with repayment. The person may even feel the debt was unjustly accrued (e.g. medical bills for botched procedures) and feel repayment is a bad thing overall.
Is the clock running? Loans are rarely made at zero interest rate, as the time goes by does my total obligation increase?
Also, what is my incentive to make any payments?
How is this relevant to anything?
It could be zero interest, if the primary purpose for holding onto it is to remind the person of their obligation and produce good feelings when they return the favor.
If you hold onto a debt, it shows on your credit report. Paying it off could improve your credit. But apart from that, there’s the matter that it is functionally identical to donating to effective charity.
The subjective feeling of obligation with regards to the original debt might affect probability of repayment over time.
Having debt without being in default improves your credit score.
So my incentives to pay off the debt are exactly the same as my incentives to donate to some charity that I didn’t pick?
Interesting thought. What metrics does this improve more than deworming or mosquito nets? If your goal is something like “increase area under the curve of “expressed IQ of humanity”, this does some good. I suspect it does less overall good, but it moves a different part of the curve than the very basic life-giving charities do.
I suspect it also has some negative effects (like encouraging less debt forgiveness/reduction than might otherwise occur: if there are charities that buy debt for more than it could be sold to vulture collectors for, that raises the overall value of creating and holding this kind of debt.)
This approach isn’t necessarily about improving the same metrics as global disease control, but the idea is that, when those whose debt has been relieved pledge to give, they’ll wind up giving more than was spent to help them, after accounting for the lost value of a potential investment and inflation and whatnot.
Vittana’s approach involves mico-lending for student loans—so that students who then graduate college have plenty of income to pay Vittana back. This approach is similar, but since forcing people to pay us back makes us essentially debt collectors (and student practices are much more established in the first world than in places Vittana services), I thought a giving pledge might be a nice alternative: “We’ll help you if you help others.”
It wouldn’t be the same metrics, which was my point—I wonder what metrics you’d be focusing on when determining the value of this expression of altruism. You couldn’t directly compare against action that helps a very different populace in different ways, but you will need to indirectly compare and you’ll want to directly compare against alternate more similar interventions (like buying used cars for the already-bankrupt or whatnot).
I guess I’m looking for a way to move from “nice alternative” to “effective alternative for X and Y goals”.
This is similar to the idea of paying for a student’s college education in return for a percentage of their future earnings. Word on the street is that this is impossible because slavery is not legal. I don’t know if that’s true, but I do know people keep reinventing that idea, and no one has ever made it work in America.
I don’t know what you standard is for “made it work,” but it has been done for school in America and people are selling equity for other purposes right now. Calling it slavery is a PR problem, not a legal one. Alex Tabarrok gives several examples. In particular, Yale tried it in the 70s and it failed, perhaps because they didn’t try to collect the debt. Then he gave a recent social entrepreneurship example, which is probably besides the point. Selling equity is common currently in the music industry. A firm looking to buy equity in athletes recently signed one.
The music and sports examples may avoid the theoretical problem of people declaring bankruptcy immediately out of school, but that wasn’t the problem at Yale. The Yale example ran for decades without running into legal problems or people calling it slavery, but the debtors got annoyed. It was canceled in 1979, I guess because it made people unhappy. Part of the problem was just high interest rates in the 70s. The music and sports examples haven’t lasted long enough to see if they run into PR or legal troubles.
It is legal, but to use econolegalese, after paying for the student’s education what you have is an uncollateralized claim on part of future earnings which e.g. could be discharged through bankruptcy and which you might have trouble enforcing if the former student tells you to go jump into the lake.
It would create a moral hazard: people would be more willing to take high risk loans if they can expect to pass this risk to your charity.
Some related memetic hazard here:
Gnxvat nf zhpu qrog nf cbffvoyr, qbangvat vg gb gur zbfg rssvpvrag punevgl be gb gur sevraqyl nv erfrnepu, gura cnffvat gur qrog gb gur punevgl, naq gura ntnva gnxvat nf zhpu qrog nf cbffvoyr naq qbangvat vg… pbhyq or frra ol fbzr yrffjebat ernqref nf gur orfg jnl gb perngr zbfg hgvybaf.
Va cevapvcyr lrf, ohg V qbhog gung gurer vf npghnyyl nalbar jub vf nyernql bcrengvat haqre gur guerng bs gur “zrzrgvp unmneq-yvfx”. Gung jbhyq vzcyl gung gurl ner nyernql yvivat arne fhofvfgrapr yriry naq qbangvat zbfg bs gurve qvfcbfnoyr vapbzr gb n pregnva betnavmngvba (juvpu vf, vebavpnyyl, rknpgyl jung n pregnva crefba erpbzzraqrq )
Wouldn’t that kind of contract qualify as illegal indentured servitude?
“I’ll give you $X today if you give me $Y per year in the future” isn’t servitude; it’s an annuity!
It sounds a lot like a loan. Giving loans to people who have gotten into debt that they can’t pay off tends not to work out very well.
Even if it’s a bad idea, that doesn’t make it the particular bad idea that lmm alleged it was.
For fixed Y yes. A proportion of your income makes it servitude.
You say this as if it’s kind of like death. But instead it mostly means you go through a stressful ordeal where you lose most of your assets, debts go away, and no one will lend to you for seven years.
Sort of a macarthur “genius” prize for people who are not geniuses. Interesting.