Another Take on Human Capital Contracts

Fellow Rhymes with Education contributor Andrew Bennett’s decision to ignore title capitalization conventions in his most recent post notwithstanding, he (re-)asks a very interesting question: “Why are the things we study so often so unrelated to the things we end up doing for a living?”

I don’t have the answer, but the folks at Upstart, a SF-based startup which just announced its public launch this week and boasts Dallas Mavericks owner and maverick financier Mark Cuban as just one of its many impressive funders, think that not only have they stumbled on one contributing factor but also purport to have a product which may be able to free future collegians from the dilemma of whether to follow their dreams or take the “safe route.” Their product, which is none other than a standard human capital contract (HCC), will allow those select few collegians who have the right combination of ambition, drive, and focus to escape the soul-sucking campus recruiting cycle, or the even more soul-sucking process of telling your parents you couldn’t find a job and have to move back in with them for a few weeks months years. (Note: Upstart.com is completely separate and different than Upstartbayarea.org.)

The basics of Upstart are these (and they should sound familiar to anyone who has read my earlier posts on HCCs, because they’re nearly identical) :

1. 

2. 

3. 

What separates Upstart from other HCC vendors I’ve read and/or written about before (e.g. Lumni, 13th Avenue Funding) is its choice of target market. As a financial product, human capital contracts are relatively simple and, for all the reasons I’ve listed before, actually better than analogous debt products when it comes to financing education. From a cultural perspective, however, converting to HCCs as a remedy for out-of-control student debt is actually a pretty tough pill to swallow. In nearly all of the conversations I’ve had with people about human capital contracts, the mention of having students sell a portion of their future income, and the corresponding truth that someone else owns that portion of their future income, consistently elicits the following knee-jerk reaction: “That sounds like indentured servitude.”

In the follow-up to his blog post on New York Times’ Fixes last summer, David Bornstein picks up on this objection and writes:

It’s not clear to me why someone who agrees to sell a portion of his future earnings for a given period of time is being enslaved. The essence of servitude is a loss of freedom. What happens today for many college students who take on student debt is that they get locked into high payments that limit their career options. I’ve met many students who say they would love to spend a number of years after they graduate working for a social-purpose organization, or serving in a program like Teach For America, or trying to start a business — but many of them end up going the corporate route because of their loans. That sounds more like servitude to me.

With human capital contracts, students would have wider options. They would know that, regardless of their career choices, their payments would not be unmanageable. For example, doctors who financed their education this way could feel more comfortable going into lower-paying, urgently needed specialties like geriatrics or serving in low-income communities, where they might earn less; young professionals in many fields could trade off some income for the chance to do work that is more meaningful and potentially more fulfilling.

Well said, David.

However, rationally arguing for why people shouldn’t be creeped out by other people owning a share of their income does not mean that we can then dismiss their being creeped out. As my girlfriend and I prove and re-prove time and again, fighting rationality with emotion (and vice-versa) rarely ends well. Best to let things cool off for a while until you can both speak the same language. (Who knew there would be free relationship advice thrown in here?) Anyway, I would imagine that being uncomfortable with people owning a share of other people’s income has its genesis in our country’s complicated and troubled racial history. And just to be clear: I’m talking about slavery.

When one adopts a more culturally holistic perspective, one can understand why the water might feel uncomfortably warm when a company creates a product that “aims to facilitate buying and selling the shares of low-income students’ future income in order to provide financing for their higher education.” Of course that characterization is quite oversimplified, but it’s not technically inaccurate. In fact, as it pertains to the company I hope one day to start or work for, it’s technically accurate, which is to say, it’s dead on.

As I was saying, Upstart chose a different target market and consequently side-stepped the mine field that is America’s racial history and present. (Yes, I did jump from talking about low-income students to talking about students of color. Unfortunately, race and class are still far too inextricably linked when it comes to educational outcomes.) Upstart, on the other hand, focuses on students who see a more entrepreneurial future for themselves. As Founder Dave Girouard wrote in his company’s inaugural blog post:

We have a surplus of bright young people who want to carve their own way – to take a risk, start something new, and make a difference. They have all the energy and passion you’d expect from people in their twenties. In most cases, they’re yet to be weighed down by the obligations that curtail risk-taking later in life – spouses, kids, mortgages, health, etc. And while not generally creditworthy in the traditional sense, there are clear and measurable signals reflecting their accomplishments and hinting at their potential. Yet we collectively tell them to take the job.

Could we imagine a future where talented grads are given a modest window of economic freedom, combined with the help and support to do what they were really meant to do?

With no explicit social agenda other than to expand opportunity broadly, Upstart may just be the company that introduces HCC-like instruments to the masses. In doing so, the risk of HCC’s “experimentation phase” will be borne by those who are best able to bear it: the risk-takers among us. And when it comes down to it, that’s the essence of financial capitalism.

Upstart is beginning this fall with recruitment efforts at 5 universities: Arizona State, Dartmouth College, Rhode Island School of Design, University of Michigan and University of Washington. A quick review of Upstart’s founding team reveals notable diversity (except with respect to age, but nobody’s perfect!) for what is essential a financial/technology start up (traditionally white and male sectors). Most notable, in my opinion, is the inclusion of Damon Whitsitt as a principal. While nearly all companies’ first priority is sales, Upstart has made the impressive realization that its product is people first, and its platform second. Damon’s background is not in sales and marketing, but rather in staffing (albeit most recently nearly six years in sales and marketing divisions at Google.)

As Upstart gains traction in the marketplace, I expect they will quickly start running into SoFi. I’ve written about SoFi earlier here. Needless to say, the financing options available to students are going to get more complicated before (if ever) they get simpler.

But wait! There’s more! Upstart isn’t the only one trying to get in on the action of helping America’s 20-somethings get a jump start on changing the world. Check out Thrust Fund.  With much the same idea and target profile as Upstart, Thrust Fund, which currently lists only two entrepreneurs seeking funding. Though it should be said that if Jon’s and Saul’s profiles are any indication, Thrust seems to be going for quality over quantity.

Anyway, as always, I’m excited to see if Upstart and Thrust can make strides towards popularizing HCC-like instruments.

Big Data and Education: A Natural Marriage (and an Unnaturally Hard Topic for Me To Comprehend Fully)

Let me start by acknowledging my own limitations here. I am neither a data scientist, a computer whiz, nor an accomplished researcher. Nevertheless, it seems clear to me that the future of education, particularly when it comes to innovations aimed at increasing overall achievement and access, is going to be driven by those who can most intelligently mobilize the gobs (I’d try for a more technically appropriate word here but it would only serve to underscore further my ignorance) of information that exists out “there” on what works and doesn’t work for individual and groups and students.

[Insert clever lead about Big Data here.]

New to big data? So am I. Let’s start here: right now you’re reading this on your computer or your phone, or some other digital device that has access to the world wide web. Hopefully, you’re also familiar with terms such as megabyte or the even larger gigabyte. In computer speak, a bit is a 0 or a 1, the elemental unit of computer memory or capacity. A byte is a string of 8 bits (e.g. 00001111 or 01010101 or 00000001 or 11101110). From Wikipedia: “Historically, a byte was the number of bits required to encode a single character of text  in a computer, and for this reason it is the basic addressable element in many computer architectures.” A kilobyte is approximately one thousand bytes (actually, it’s 2^10=1024). A megabyte is million bytes or a thousand kilobytes: 2^20. A gigabyte: 2^30.

For reference, consider that all the text on this entire blog is somewhere on the order of a few kilobytes. A typical song in .mp3 format is usually a couple megabytes. And until very recently, even the nerdiest of computer users would rarely speak about volumes of information larger than gigabytes. Now, open your mind to the nearly incomprehensible fact that innovations in digital storage and analysis have allowed us not only to collect and keep but also to search for patterns and hidden insights within terabytes (2^40), petabytes (2^50), exabytes (2^60), zettabytes (2^70) and yottabytes (2^80) of data, the latter few of which sound as if they were made up by Johnny Depp as Keith Richards as Jack Sparrow mocking Harrison Ford as Hans Solo in a Home Star Runner mash up of Pirates of the Caribbean and Star Trek. At least that’s what they sound like to me.

If the math or my pseudo-pop-cultural references don’t assist in your understanding, check out Wikipedia’s Big Data page, this story from February about how Target may know you’re pregnant before you do and Quentin Hardy’s recent NYT article entitled How Big Data Gets Real.

Two recent stories have caught my eye and provoked this post.

1) About two months ago, Civitas Learning publicly announced its launch. Founded by Charles Thornburgh, and funded by Austin Ventures, First Round Capital and Floodgate to the tune of $4.1 million, Civitas has set out to solve this problem:

As colleges and universities strain their resources to accommodate an ever-growing, diverse student body, it becomes increasingly difficult to track the progress of individual students and figure out what’s working in their curricula, coursework, testing, and teaching styles. An appalling number of students in the U.S. are still dropping out before they earn a degree, and despite Mark Zuckerbergian exceptions, attrition means sunk costs for students, difficulties finding jobs, and loan defaults. [from TechCrunch article]

How Civitas aims to help is by employing the same strategies Target, Amazon, and others are already using to try and attract our dollars. Again, per Rip Emson at TechCrunch:

Big data, predictive analytics, machine learning and recommendation engines are transforming the way we buy products, play games and watch movies, and Civitas Learning believes that the same should be true for education.

I’m looking forward to hearing more about CIvitas in the future.

[Side bar: If anyone from Civitas is reading this and you want an MBA intern for next summer, call me maybe?]

2) This NYT article about big data on college campuses. It’s a great article and deserves much deeper treatment, but for now, allow me to skim the surface. Arizona State is the country’s largest university. Not coincidentally, it has also become the country’s (if not the world’s) largest laboratory for applying big data analytics to education:

The new breed of software can predict how well students will do before they even set foot in the classroom. It recommends courses, Netflix-style, based on students’ academic records.

From a human capital contract perspective, these developments are encouraging and exciting. One of the biggest challenges in implementing human capital contracts in any sort of broad manner is appropriately pricing them on a student-by-student basis. Algorithms that can help us effectively and efficiently compare students are a huge step in the right direction.

Carfax for Students? How to Overcome Adverse Selection Problem for HCCs

One of my earliest posts on this blog entitled, “A Peak Under the Hood of Human Capital Contracts,” elicited two comments from faithful followers. The first comment–and the one I want to focus on here–was from “Thomas Colfax.” He wrote:

I understand the value of these deals to the contract provider, and I also understand the value to the risk-averse, low-income student who fears being laden with debt in an uncertain economy. What I don’t understand is how this would work for a highly-driven, high-achieving, middle-income student who need money to attend a top college. For them, it seems that a human capital contract might provide a disincentive for entering certain careers. Why would I go into investment banking if this company is going to take a huge percentage of my salary? Why would I go to law school if I am penalized for landing a more prestigious, higher paying job? In the case of students who plan to enter high-paying professions, it seems more sense to take on loans at a steady interest that is unrelated to salary than to agree to a contract that ensures that a higher-paying job will require substantially larger loan payments.

Basically, what Thomas was asking about was the problem economists refer to as adverse selection, which is simply the negative or “bad” result of a market in which buyers and sellers have asymmetric (read: different) information. The classic example of adverse selection most often cited in introductor economic texts is the used-car market, or what Geroge Akerloff called, “The Market for Lemons” in his 1970 paper which bears the same name.

Consider briefly the experience of buying and/or selling a used car. The seller presumably knows much more about the car than the buyer. Even if the seller wishes to be honest in the deal, she is likely to leave out details about the car’s history either because she forgets or because the knowledge is so esoteric and wrapped up in the minutiae of her driving style as to be almost incommunicable. This is what economists mean when they talk about information asymmetry. Given this dynamic, buyers are understandably wary and consequently reduce the amount they are willing to pay for a used car in order to compensate themselves for the risk of taking on a car of unknown quality. Moreover, sellers, knowing that this dynamic will cause buyers to be wary and undervalue their cars, reexamine their decision to sell their car in the first place. If you’ve ever had a parent (Dad!) tell you not to sell something because its worth far more to you than it is to anyone else, this is exactly what they are talking about. When this cycle is taken to its logical conclusion, the only cars left available for sale on the used car market are those with serious defects, that is, the lemons!

So what would adverse selection look in a hypothetical market fro human capital contracts? Put simply, students with potentially higher earnings, facing a decision between financing their education through debt or equity, would be more likely to choose debt, at least when compared to students with lower potential earnings making the same decision. As a result, students with lower earning potential would be overrepresented in the population of students interested in human capital contracts.

Additionally, as a 2002 UPI article on HCCs noted, “statistically speaking, the lower wage earners would likely outnumber high wage earners because there are simply more people on the lower end of the wage scale in the United States.” However, this a reality that both HCCs and student loan originators face, so it would not necessarily disadvantage HCC originators any more than it would lenders.

Eric Hanushek, then a senior fellow at the Hoover Institution, commented in the same article:

You would tend to get more people signing up for (human capital contracts) who are going to be low earners, and what they do is push up the repayment rate, essentially. I think there are attractive aspects of this plan, but its success rests on the willingness of some successful people to subsidize those who are not economically successful.

Making my job significantly easier, though, I needed look no further than the same article for the perfect response. Naturally, it comes from Miguel Palacios who says actuarial models, like those used by insurance companies, are the key to overcoming adverse selection. His quote from the article: “The way to face the challenge of adverse selection is to price every contract according to the different abilities of every student.”

But seriously, this isn’t an insane or unattainable idea. Listen, I have no idea what the used car market was like in the 1970s and I don’t have any interest in taking the time to research it. But, Akerloff was clearly a pretty smart guy cuz eventually they–and by they I mean the people who decide who gets the Nobel Prize–decided in 2001 to give him a prize–and by prize I mean the Nobel Prize. His co-recipients that year were Joseph Stiglitz and A. Michael Spence. Spencer I’ve never heard of, but Stiglitz is the man. But I digress…

So Akerloff made a very astute observation about the nature of asymmetric information and the effect it should have on the used car market. And yet, we all know people who have bought or sold cars on the secondary market. Perhaps we have even participated individually. What changed? I’m sure lots, but one big innovation stands out: Carfax.

Founded initially to combat odometer fraud (ingeniously demonstrated by Danny DeVito in his role as Mr. Wormwood in Matilda), Carfax has grown to possess a database of over 8 billion records from over 34,000 sources in all 50 states and has become a must-have report for any secondary market participant. What does this have to do with HCCs? Maybe not a lot, but definitely some. Students aren’t cars and fax machines are out of date, but here’s a quote from chapter 3 of Saving Capitalism from the Capitalists (I’m reading it on my Kindle so I have no idea what page it’s on):

Some financial instruments may appear very risky a priori, and apprehensions about their risks may indeed be justified. But if sufficient investors become interested in them, liquidity in the instrument increases, and more resources are devoted to understanding and laying off its risks.

Just as Carfax emerged to fill the markets need for a solution to asymmetric information in the used car market, I am confident that another service will evolve to fill the same need in any hypothetical market for human capital contracts. In fact, that company may already exist: PayScale.