What would happen if investments in people succeeded?

By Felix Salmon
October 22, 2012
Daniel Friedman has a question:

" data-share-img="" data-share="twitter,facebook,linkedin,reddit,google" data-share-count="true">

Daniel Friedman has a question:

There are now a few companies — Upstart and Lumni to name two — trying to create a market where we can invest in people in exchange for a percentage of future income. If they succeed and reach scale, will student debt problems be alleviated? Will our future incomes be reliably predicted by newly developed algorithms? Will it encourage increased risk-taking?

The first thing to note here its that this is a market which does not exist. Lumni tried to gin up some interest in it and failed; at this point they won’t even return my calls or emails. Upstart has a bunch of high-profile venture backing (Google Ventures, Kleiner Perkins), and has even managed to launch a pilot with a few hand-picked students and investors. But it’s still very, very early days. Equity-in-people is an idea which comes around occasionally, but so far it has never really taken off, and there’s not much reason to believe that this time is different.

But still, it’s an interesting question: what would the consequences be if Upstart, and/or companies like it, became big and successful?

Firstly, Friedman asks, would student-debt problems be alleviated? The answer to this one, I think, is quite clearly no, for two reasons. The main reason is that tuition fees are induced, in much the same way that traffic is. If you build a nice big new road, traffic will appear to fill it up. And if you build a nice new source of funding for students, universities will raise their tuition fees so as to capture as much of that funding as possible. Yes, high student debt is a consequence of high tuition fees — but at the same time, high tuition fees are a consequence of the ready availability of student loans.

It’s worth pointing out here that Upstart’s model gives out funds at the end of a student’s four-year education, rather than at the beginning. If the student has lots of debt, and wants less debt, then she can use the funds to pay down some of that debt. But many students will choose to spend the money in other ways, simply layering the new equity funding on top of their old debt funding. Alternatively, if the expectation is that students will use their Upstart funds to pay down debt, then that only means, at the margin, that lenders will feel free to extend even more credit in the run-up to the early cash-out.

The other reason why student-debt problems won’t be alleviated is that Upstart funds are debt, if you look hard enough. The documentation is subject to change, but at heart would-be Upstarts are always going to find something along the lines of this in any Upstart contract:

YOU UNDERSTAND AND ACKNOWLEDGE THAT THE FUNDING AMOUNT YOU RECEIVE FROM US, IS A DEBT TO US.

The debt is not a typical loan, of course, and it can end up being repaid at less than face value, if the Upstart enters a poorly-remunerated career. But it’s still a debt. And if the student ends up defaulting on that debt, the penalties are generally enormous. When people take out loans, they nearly always have the best of intentions, and don’t think there’s any way they will end up defaulting — especially if the repayment amounts are tied to their income. But defaults happen. And when they do, anybody who has defaulted to Upstart is going to find out the hard way that Upstart loans are the most expensive of them all.

So at the margin, these kind of schemes are likely to exacerbate the student-loan problem: they’re hurting, rather than helping. Remember that ultimately the money for these schemes is coming from investors, who are taking a substantial risk and being promised returns significantly greater than anything seen in the debt markets. Remember too that a significant portion of the students will end up repaying less than they’re originally given — because they go into low-paying occupations, perhaps, or maybe because they just decide they’d rather settle down and have a family instead of a career. As a result, those students who do go into decently-paying careers will end up repaying sometimes double or triple the amount of money they were originally given. That doesn’t seem like an alleviation of student debt problems to me — not when we’re talking about sums in the $30,000 range. It sounds like piling an extra $60,000 of liabilities on top of all the existing student loans you might have.

Friedman’s second question is whether companies like Upstart will start being able to predict students’ future incomes. That’s certainly something that Upstart is spending quite a lot of time on, but I think the effort is premature: first you need to find out whether students are happy repaying the funds at all. And I, for one, don’t think that Upstart’s algorithms are going to be particularly good at predicting incomes.

For one thing, the students with predictable incomes — the doctors and lawyers and the like — are never going to accept Upstart’s offer. For another, there will always be a certain number of students who are looking to game the system. All these schemes have an adverse-selection problem: Upstart funding is going to be much more attractive to students who are pretty sure they’re never going to make much money. If you aspire to being a part-time primary-school substitute teacher, for instance, while spending the rest of your time working on the Great American Novel, then Upstart might be great for you — and it’s unlikely that any algorithm would be able to capture that.

What’s more, the biggest risk, from Upstart’s point of view, is that they’ll end up funding a student who marries someone very successful, gets pregnant pretty early on, and then never returns to the workforce. This still happens frequently enough that an honest algorithm would charge higher interest rates to women than to men. As a result, the algorithm can’t actually be honest.

Finally, there’s the group of students who want to follow the Silicon Valley dream and become entrepreneurs straight out of college. That’s generally a great time to found a company, and Upstart (slogan: “The startup is you“) is positively encouraging such people. But equally, the chances that any given person will have success as an entrepreneur are pretty much random: such things come mostly down to luck, and can’t be reduced to some algorithm. Especially when the really successful entrepreneurs are going to be the ones who can afford lawyers to shield their income from Upstart.

There is no algorithm which can tell you who’s going to be a successful entrepreneur and who isn’t. Some college dropouts are college dropouts; other college dropouts are Steve Jobs, or Bill Gates, or Mark Zuckerberg. Being dyslexic is generally not a great thing when it comes to total lifetime income, and yet dyslexics are massively overrepresented among highly-successful CEOs. If you’re looking for outliers — and successful entrepreneurs, pretty much by definition, are all outliers — then you can’t really start generalizing.

Friedman finishes by asking if this model will encourage increased risk-taking. I think the answer there depends on what you mean by risk-taking. It certainly encourages people to accept low-paying jobs with payoffs which get put into corporate shell vehicles or which don’t end up cashing out for a decade. And at the other end of the spectrum, it encourages people to take low-paying jobs which pay non-financial dividends: ski instructor, say. But in general, I don’t think that a new form of private income tax is a great way to encourage innovation. Generally, if you start taxing something, you get less of it, not more of it. And I see no reason why this model should work out any differently.

More From Felix Salmon
Post Felix
The Piketty pessimist
The most expensive lottery ticket in the world
The problems of HFT, Joe Stiglitz edition
Private equity math, Nuveen edition
Five explanations for Greece’s bond yield
Comments
5 comments so far

“If they succeed and reach scale, will … our future incomes be reliably predicted by newly developed algorithms?”

Yes, at least to some extent; if our future incomes are not reliably (at all) predicted by newly developed algorithms, they will not succeed and reach scale. Which, in fact, is part of the reason they will not succeed and reach scale.

Posted by dWj | Report as abusive

The highly successful outliers will be the first to default.

As happened for instance with German tennis player Tommy Haas, who defaulted in 1999 as he started winning tournaments…

Posted by Panley | Report as abusive

What happens to this debt during bankruptcy? Is it considered a secured loan ;).

Posted by mushr00m | Report as abusive

I think this model could maybe work for a small college. Students would get free tuition, housing & supplies, etc, in exchange for 10% of their pre-tax income, ad vitam eternam. The college could start by offering math & programming, and expand into other STEM fields overtime.

Obviously, it would probably never offer psychology & visual arts…

Posted by JDelage | Report as abusive

Private Equity: the way for smart, rich lazy motherflippers to get even richer.

Public investment in people? No percentage in that. Communism by another name.

Remember: passive income is massive income!

Posted by Eericsonjr | Report as abusive
Post Your Comment

We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see http://blogs.reuters.com/fulldisclosure/2010/09/27/toward-a-more-thoughtful-conversation-on-stories/