a16z Podcast - Are ISAs the Solution to Student Debt?

Episode Date: December 19, 2019

A bold proposal: You go to college for free, then pay back the school after graduation—but only if you get a job in your field of study and make a high enough salary to afford it. It's called an income share agreement, and Austen Allred, the CEO and cofounder of Lambda School, thinks it's the future of education.Student debt currently stands at more than 1.5 trillion dollars, which makes it the second-highest consumer debt category behind mortgage debt. The crisis has saddled much of a generation, with far reaching effects. Income share agreements, or ISAs, have been put forth as an alternative to the current system. Put simply, an ISA is an agreement between a school and a student for the student to pay a defined percentage of income to the school, for a particular period of time, up to a certain cap. It's a seemingly simple conceit with complex design considerations, and it's spurring debate across media and politics.In this episode, Lambda School CEO Austen Allred, a16z general partner D'Arcy Coolican, and a16z editorial partner Lauren Murrow delve into the greater implications ISAs may have for education and the economy. The discussion covers both the promise and the challenges of ISAs—why they've been relatively slow to gain traction, why they've failed in the past, and why some in the political sphere are still skeptical.

The content here is for informational purposes only, should not be taken as legal business tax or investment advice or be used to evaluate any investment or security and is not directed at any investors or potential investors in any A16Z fund. For more details, please see A16Z.com slash disclosures. Hi, and welcome to the A16Z podcast. I'm Lauren Murrow, and I'm here with Austin Elred, CEO and co-founder of the skills-based online school, Lambda School, and Darcy Kuliken, a partner on the consumer tech team. Today, we're discussing an issue that has saddled much of a generation, student debt.

Student debt currently stands at more than $1.5 trillion, which makes it the second highest consumer debt category behind mortgage debt. The national total for student debt is higher than both credit cards and auto loans. One solution that's been proposed is income share agreements, or ISAs. It's a concept currently in the zeitgeist that's spurring debate across media and politics. Here's the idea. Rather than charging students tuition to attend college, often forcing them to then take out loans in the process, they go to school for free. But under an ISA, they are then required to pay back a percentage of their income after graduation, but only if they land a job with a salary that meets a certain amount. In this episode, we delve into some of the greater implications ISAs may have for the future of education, the economy, and more. We also touched on some of the challenges of ISAs. while they've been relatively slow to gain traction, why some have failed in the past,

and why some in the political sphere are still skeptical. As we begin our conversation, Austin gives his explanation for why I say's work. I think of it as a very, very forgiving type of way to pay for education. So if the education doesn't work, basically you don't make payments. So at Lambda School, you don't pay the school anything unless you get a job in the field you trained in, that pays more than 50K. And if you get that kind of a job, you pay a percentage of your income

for a couple of years, and that's it. Put this into context for us, if ISAs are the future, what's the big picture potential? It just doesn't make sense for a lot of people to attend college anymore. Earnings have stayed flat, basically since the 80s,

and tuition has gone up basically 3X, and so that gap is getting bigger and bigger, and it's more and more risky every day for a student to go to college. So there are times when a student will enroll in a school when they're 18 and they'll pick something to study. And when you look at the amount of student debt they're going to have, everybody knows from the school to the person funding the education to the government that that person is unlikely to be successful and ever

pay back their student loans. If you look at most of the student debt, a lot of it is from students who didn't graduate or from students who aren't making enough. And almost all of the default in student loans comes from those categories. In the world of an ISA, that doesn't happen. Almost by definition, a student who is unable to make payments on their ISA is much, much less likely to happen than a student who can't make payments on their loans. But what it really allows us to do over time is better assess who will end up in the right career path, do everything we can to get them there. What do you mean when you say get them in the right career path? So you walk in on day one knowing that if you don't get a good job, you're not paying us anything.

Everything is driven to making you successful in the career path that you may choose. It also diversifies risk, and it allows people to do things that they otherwise might not have been able to do. If you look at education outcomes, especially if it's vocational schools where the ultimate outcome is getting a job, it can be relatively binary. You either get a job or you don't, and when you have those things that have a high variance, it can oftentimes feel really risky for the person to enter. into that. Because if they don't get the job, that risk has then been placed on them. The way that income share agreements work is they're taking that risk from the individual and putting it on the school. Say you have a school with a 75% graduation rate, and then of those graduates, 75% end up being successful in the job that they train for. Those are not bad rates. In fact, most colleges

would take those rates any day. But when you look at the tuition dollars that are flowing into that school, basically 46% of those tuition dollars are coming from people who are unsuccessful. So if those people are saddled with loans, there's no way they're going to be able to pay those off. If you're talking a private student loan, your interest rate could be 10, 11, 12%. If you're only making 30, 40, 50k a year, like, you might be financially ruined forever. The more income has spread that then makes their financing education through debt that much riskier.

The winners will win bigger, but the losers will lose bigger. And to the extent that that's financed through debt, that just can become a massive overhang on somebody that doesn't win coming out of that education lottery. And one of the important things that economists have realized in the past few years, I mean, Connman and Tversky, they came up with the notion of behavioral economics. It was basically the realization that humans don't always act in an entirely rational way because their downside risk can be so much greater than the upside. So ISA is a way of saying, let's say you have a 90% chance of your education working out and a 10% chance. chance of you being financially ruined for life, a rational person will still not play those odds. So as a school holds the risk, it now becomes 90% chance that somebody is successful and pays back and a 10% chance that you have to eat those losses. The school can actually afford

the risk in the way that an individual can't. And when an individual is de-risk, because they can look for more optimum outcomes and not having to weigh in that downside risk is really, really important. Are there types of students that ISAs make sense for and don't make sense for? Depending on the terms, I think it makes sense for every student. As the ISAs exist today, they're only slightly more expensive than upfront tuition at most schools. So you can take the risk and say, I'll pay $20,000 upfront tuition, or if I'm successful and I get a six-figure job, I'll pay up to $30K. Turns out if you're successful and you have a six-figure job, that 10K Delta doesn't feel as bad, and you would trade that for, and if I don't get a job, I don't pay anything. In my ideal world, you only

actually pay the school if it's less than the increase in your income. The place where maybe the jury's still out, where we're still trying to figure out whether the economics work is places where there's much lower variance. Traditional finance theory would say that's a place where you should take out debt rather than some other form of financing. And so in those instances, which to a certain extent is kind of what college looked like in the 1950s and 60s and that it was much more predictable. And debt was the like natural financial instrument to use then at that point. But now that we're at a place where the risk is so much higher, the outcomes are so much more distributed. Income share agreements are having this moment now because that that's the kind

of population that it really makes sense for. But I do think it does come back to what is the upside risk versus what is the downside risk. This may not be the perfect example, but say an NBA athlete or someone where if you're successful, the upside is millions and millions and millions of dollars. And you could de-risk that by saying, okay, an ISA will pay me 100K for life in exchange for if I'm wildly successful, I'll give up 10% of my income. And you can use the high potential for upside risk to de-risk it for people and cover the downside risk. That's the kind of financial engineering that I find fascinating. The easiest way to think about it is just its insurance. You're saying if something really bad happens, I'm not going to be in really dire straits.

Insurance is a really good way to think of it. And that's the instrument we use in similar scenarios where, you know, okay, I'll pay $100 a month and I don't love paying $100 a month. But if someone wrecks my car and I now owe $20,000 to get a new car, I would like to not have to pay that because I can't afford to. For the most people, when you're talking houses and cars and potential really high downside risk volatility, it makes more sense to de-risk yourself. Let's say you're right in the income-conshare agreements become more common. How might that actually shape education? So what I hope will happen is you can drive the effective price of studying something that is likely to pay well down. What I really think is missing in the economy

is something that will tell you, hey, based on who you are and what you know and your talents and proclivities, here is the best place for you in the economy, and we're going to help you get there. I would imagine that the vast majority of the population could actually end up in a better place than they are today if they only had any idea what opportunities are available to them. Universities don't really serve that function today. When you show up at the front door, they're not handing you a book that says, here are the best majors for you to choose. And if they did that, then professors would lose their minds. So I hope that it becomes more transparent when you're a student. And I think about our economy of human labor being the majority of

GDP, 52, 53% of GDP, and completely unoptimized. And that's insane. We're our biggest asset class that exists. What most of us spend most of our waking hours do is not optimized at all. Education as it is today is this kind of bundled value proposition, where it is skills training, it's career development. It's also networking. What ISAs are really good at doing is they will kind of like debundle that and they will atomize the different parts of education and take this part around kind of skills development, training, the ability to focus on optimizing and maximizing your long-term income. And it will make really clear which programs or institutions are optimizing for that. I think it's insane that if, for example, there's a factory that goes out of business in Detroit,

There are all these people that are unemployed, and even if there is huge demand in other parts of the economy that those people could do, there's nothing that matches those two. There's nothing that will take an unemployed factory worker in Detroit and get them, even if there's a company next door that desperately needs something, there's no bridge between those two. So at a higher scale, I hope ISAs will enable a broader economic clearinghouse where we can move people to where they're most valuable in the economy, where they're happier, where they're paid more. eliminate the fraction of doing that. I think it's probably the biggest barrier to economic progress that exists today. Assuming you have this ecosystem that's built on top of ISAs, then you will have this much more efficient system that can move people in and out of the workforce and retrain and reskill people much more quickly. A critique of ISAs is people say graduates with high salaries might end up paying more under an ISA than say a traditional student loan program. Yeah, it's a

trade-off that you have to make for sure. But we offer both up from tuition and an ISA, and 98, 99% of our students pick the ISA. Because in the scenario where you are paying a little bit more, you're fine, and you're making enough money that it's not a big deal. And frankly, for us, you only pay it for a couple of years. So it's actually not a big delta relative to if you had $20,000 on the line and you have no idea whether it's going to work out and you owe interest on it no matter what. Yeah, you might end up having to pay a little bit more than you would if you just

paid straight up, but you're getting the psychological use of not potentially having that debt hanging over you. Most ISAs are capped. So no matter if you're making gazillions of dollars, there is a maximum amount. You will pay, whether it's like $20,000 or $30,000 or $40,000 over the lifetime of an ISA, you'll only pay X amount of money. You're paying for risk protection on the downside, which can have a bunch of knock-on effects in terms of how you think about your lifetime earnings. If you're able to not only take the risk of going to school or going back to school or taking time away from the workforce, but also thinking about your career over a longer time horizon because you don't necessarily have to service debt immediately upon graduation.

That's one thing that can actually change the dynamics of a career. Yeah, one of the instances that's been most fascinating to watch is if, for example, you took out a loan to pay Lambda school, the loan payments start becoming due immediately upon graduation. So we've seen circumstances a couple weeks ago where a student got a job offer for 70K, and he was saying, okay, like, that's not a bad job offer. I can take that. And we thought he could do a little bit better. So, like, you know, you might want to turn that down and keep looking at the market because based on what we're seeing, we think you could get more. If you have a debt payment that's about to come in, you take that 70K offer because you have to make those payments.

In this circumstance, he waited a couple months. We didn't, we waited longer to get paid, but this particular student ended up getting a job for 240K. So a massive, massive difference in income just because he had the psychological ability to wait a little bit and to pick the actual optimum outcome for him as opposed to optimizing for, I've got these payments that I need to make that are going to be due no matter what my life looks like. The financial ability to do that too. It's like psychological and financial. It's freed him or her to kind of make that better decision.

And we all know these people that graduate from university. They take the job in like investment banking or something that's going to give them a high cash component for like three years because they want to pay off their student debt. And then at that point, they're going to go do what they really want to do. And that's like this common thing that you see amongst graduates all over the U.S. And income share agreements can mitigate that. And it can change the immediate incentives you have upon graduation. And we're seeing those incentives play out at a broad scale. People are starting families later. They're not buying houses. They're not starting businesses. White Combinators trying to do a study

right now that determines how many people would be starting businesses if they didn't have student debt. If you have $1,000 a month in debt payments, it becomes really difficult to start a company. So you're saying the impact of ISAs may extend beyond the issue of student debt? I think a lot of the reason that people get excited about ISAs is because of the second, third level effects of it. The ISAs are kind of like the building blocks, but then you do need this robust ecosystem built on top of it. And I think we're in the early days of building that ecosystem right now. What else is in the ecosystem? If you think about ISAs broadly, there's places where you can use ISAs beyond skills training. You can think about it as I can move you from place

A to place B and pay for that and take those upfront costs and then take an income share agreement on the back end and that's the way to finance that. Right now, mentorship exists in this informal, nice people do it type thing, but you can think of a way where that gets much more formalized. There's tons of value propositions that can be layered on top of income share agreements that I think people get very, very excited about. I've seen things go as far as an ISA for immigration, where we'll pay for all of your legal fees, we'll do everything that we can, and then if you end up in a certain country where you're making way more money, you pay a percentage of income for a few years, but net net is

going to be more than you were making wherever you were and you can create a new life. The design of it is so important. You really have to consider the details when structuring it. You're 100% right that the devil's in the detail. So in theory, you could make an ISA that makes a lot of sense, or you can make an ISA that makes no sense whatsoever. And my hope is that as there is more competition in the ISA space, and as there becomes better regulation, more regulatory guidelines, more repayment history on these things, we can start to figure out what the optimum rate is. Right now, it's a little bit up in the air. When we started, we were basically guessing about what repayment rates would be, and we created a terms

that made sense in this imaginary model. A loan has a par value, it has an interest rate, you know what payments are going to be coming in every month, and then if they don't come in, then you call that a default, and you report it to the credit bureau. With an ISA, there are times when a payment is not coming in, and that's okay, because it's built into the agreement itself. In an ISA world, that's actually not a default. That may be exactly what the instrument intended for. The entire ISA market today is maybe a couple hundred million dollars a year. It's not big relative to basically any other asset class, whereas student loans are in the trillions. You bring up a good point in that ISAs, though they're getting a lot of buzz, in politics, and in the media, are still a relatively small percentage of schools that are actually pursuing this.

Why aren't they gaining more traction? Well, as a school, today you will make less money with an ISA than you will with a loan. The way the ISAs are structured today, if my upfront tuition is $20,000, I'm actually not planning to make $20,000 on average from an ISA, and I might have to wait a couple of years for that to happen. So it's actually a worse deal for the school in many cases. But the important flip side of that is it opens up access to students who wouldn't otherwise be your students. So Lambda School is based on saying, what if we increase accessibility to the point where basically anybody with a laptop can attend? So now there are millions and millions of people who wouldn't be willing to take the risk upon their own financial future to go to a code school. And we say, try it out, see if it works for you.

And if it works for you, you're going to pay us back. But you have a great job now. So it all works out. But the cost of capital is high enough. and they're different enough from loans that you, on average, expect to make less. So the reason that so few schools are doing it is because you really have to design the school around making that work. So do you think then that ISA's work well for vocational schools and coding schools, but the jury still out on more traditional higher education?

One of the ways they're being used in traditional universities, which is interesting, is as a, a tool to help people stay in school. The biggest problem that all the universities deal with is retention, retention, retention. A university looks at how much financial aid they have. Some is from the government, some is internally, and you can pretty well identify the students that are going to need that student aid or they're going to drop out. That student aid is not enough to cover everybody. So there are people who, it's basically either they're going to drop out of a university or you're going to give them an ISA. So even if you don't get full tuition value dollar on the dollar,

it's better to have those students pay you, maybe it's 90 cents on the dollar, maybe it's 75 cents on the dollar to allow them to not drop out. Obviously, that approach is different than a vocational school like ours that is built in such a way that you don't pay unless you get a job. There's a bunch of things probably throttling the ISA market right now. I think regulatory uncertainty is probably one of those. the investor side likes to invest based off of data, and they like to invest into things that are very predictable.

The nature of ISAs is that they just take years and years and years to really build out that data set. So I think the nature of ISAs is just something where it's destined to grow relatively slowly just because the capital markets don't unlock without predictability. Not to mention the fact that if you're a university, your cogs are so high that you can't create an ISA with a two-year repayment window. They're all eight years, 15 years. So that's going to take a very long time to know what the repayment predictability is like. So the number one thing holding us back from creating a bigger data set is there just fundamentally aren't enough ISAs to get the capital in so you can watch the data roll through. If you look at the history of lending, it was very similar.

Nobody knew what their right interest rate would be. Nobody knew how to securitize loans and turn them into a bond and start selling them. Today, there are trillions of dollars that are looking for returns, but they have to be very, very sure returns before you'll put those trillions of dollars in. So our hope is that we can make that data happen faster and bring down the risk faster. You bring up an interesting point in that ISAs, they're very buzzy right now, are not a new concept. People have tried this before, maybe failed. What are you doing differently? The most popular experience with ISAs was done by Yale.

in the 70s. And they did it in a very interesting way, which was they said, okay, this class will cost us end dollars in tuition, and everybody's going to pay a percentage of their income until we've reached that amount plus a little bit of interest. But it was as a pool, not as an individual. And it was very explicitly, once we've hit that amount, then everybody's done. Not an individual ISA with a cap. It was the cap for a class, which feels much worse if you're a high earner. And so that actually didn't work. Importantly, they also gave people the opportunity to buy out early. So not only did they structure it was a class and there was this rate of return that the class had to hit as a cohort, but also they gave people the ability to pay, I think it was

like 150% of the cost of their tuition. So the people who would have been the really big earners later on in life had the ability to buy it early. And then obviously people who weren't doing well were paying a smaller percentage of their income. So it really got burdened on this middle class experiment. And so I think the lesson from the Yale experiment, which I think is the lesson of all ISAs in general, is that design is really, really, really important. All financial instruments in general. Exactly. It's like, are you giving people the option to like opt in or opt out? And when is that happening? And how is that creating adverse selection in your pool? How is that changing everybody else's burden? These things matter a lot. And so I think ISAs are extremely exciting

at the conceptual level. But I think also whether it's the Yale experiment or there was a series of companies in the 2000s and the early 2010s that came and went and they were using ISAs, but they didn't get those details right. I think another big difference between kind of ISAs today versus ISAs in the early 2010s and in the past is most of the ISAs today are used in vehicles that are meant to shift your income. I sometimes refer to ISAs as like selling an out-of-the-money call option against yourself, right? So you'll pay, but only in a circumstance where you're, you'll pay.

your income changes so much that you don't care. In the past, when it's just been, hey, anybody can get an ISA, there's the selection bias, right, of why would I take out an ISA if I can take out a loan? But when tied to an educational institution, the reason you take out an ISA as opposed to a loan is because the institution is promising to shift your income or you don't pay, which is very different than just a raw ISA to pay for a car repair or something like that. They're not lending you dollars. They're lending dollars to a school that's going to shift your income. Which is one of the things that was new about this generation of income share agreements that didn't exist in the previous generations.

So I want to get into the political spectrum a little bit. So the student debt crisis is something that has gotten a lot of attention in the media, particularly leading up to the upcoming election. In June, members of Congress sent this letter to Secretary of Education that read, in part, ISAs carry many common pitfalls of traditional private student loans with the added danger of deceptive rhetoric and marketing that obscure their true nature. What would you say to skeptics? A world of ISAs, I think, would be demonstrably better net net than a world of the existing student loan arrangement. I've spent a lot of time on Capitol Hill talking to Congress people and senators on both sides of the aisle.

And what I've generally found is just a lack of understanding and a curiosity and kind of just a fear and skepticism that's probably healthy of, hey, there's this new financial instrument. And one of the reasons we're pushing so hard for regulation in the space is because it's not hard to imagine an ISA that would be predatory, right? The same way, if there were no lending regulations, there are a million different ways that you could abuse loans. You could say, this loan has a 9,000% interest rate, but you get your car today. That's not a win for anybody. So I haven't seen anybody abusing ISAs, but I think politicians are skeptical. What the Department of Education is saying is, let's look at this deeper. So similarly, we still at the federal level haven't seen that data yet.

We haven't seen what the returns are like. We haven't seen what will happen to student repayment. We haven't seen what the burden is like. we haven't even created rates as the federal government for what ISA repayment ought to be. So I think it makes sense that there's skepticism. Let's say someone completely wipes out existing student debt. Does that affect the outlook for ISAs at all? Well, if you wipe out the existing student debt, based on the way we're pacing,

will be basically where we are today in about 10 years. That's depressing. So you can wipe out the student debt, but the bigger question is who pays for the education, and how? And is it just the federal government will now make all university tuition free? Which I suppose is an option. I'd be shocked, but that's an option. Even in that world, there's still probably a place for ISAs. I'm from Canada, and health care is free, but there's still for fee medical service. One of the thing that's really interesting is one of our bigger markets right now is in the EU. And in a lot of countries, there's free

education, but it takes a long time, it's low quality. And a lot of our students are saying, Even though I have the opportunity to go to college for free, Lambda School is still providing a better return because it gets me the things that I need, it gets me into the career faster. And our ISA in the EU is 10% for four years. And so we can make everybody's income, I think, 10% higher than it would be at a university over the same lifetime. It makes sense that even if there's a free option, there would be a premium option that would exist that would attract a certain segment of the population. The fact that you're offering an ISA is a signal that we believe we can fundamentally change your outcome in a way that's a delta to the free system, right? Could there then be unintended consequences? To the extent that ISAs become dominant, majority, exclusive way to finance education, the entire education system then will reorient itself around this idea of maximizing future income, which for a lot of students, that is what they want, for some students, that is not what they want.

And so you would have these knock on secondary effects of income is the thing we can measure. It's the thing we can optimize against with this financial instrument. There's all kinds of other value being created within the current educational system. And where and how those things would slot into this new different world of exclusively ISAs is an open question. So I think when we talk about kind of secondary effects of this world, I think that is probably one that is worth paying attention to. Agreed. We talked about immigration earlier. Is there potential for this model to play out in fields beyond education?

I think about the excitement around ISAs, and education is one big piece of it. The idea of extending it out into other layers of the educational sphere is exciting, but more unproven. And then you trace it out farther and farther and farther, and you have this version where you can be using ISAs for all these other things. So whether it's immigration, whether it's diversifying risk amongst athletes and artists, which is another use case that people, are talking about. You can also think about how that plays out in larger and larger context. We're IPOing cars, we're IPOing pieces of art. There is this kind of trend towards more and more securitization of different types of assets. And so it makes sense that the ability to finance advancements in human capital would naturally tend towards some version of that.

Well, thank you so much for joining us. Yeah, thanks for having me. Thanks.