Flooded Cars, Blood Testing, and Rookie Running Backs
Why informational asymmetry only sometimes leads to bad outcomes.
The other day, I came across a vexing and unexpected headline. In the wake of Hurricane Ida, it read, consumers should be wary about an influx of flood-damaged cars into the already overheated used car marketplace. The article included all sorts of tips about interpreting vehicle history reports and checking for mildew, lest someone try to con you a wheeled vehicle until recently moonlighting as a boat.
Anyone who has taken an intro-level economics course will recognize used cars as the go-to example of the perils associated with asymmetric information. In his now famous 1970 paper, economist George Akerlof characterized the market for used cars as one containing both “lemons” and “peaches.” Defective vehicles, or “lemons,” he suggested crowd out the higher quality “peaches,” leading to a market failure in which only low-quality cars remain.
When Life Gives You Lemons
The story goes something like this. To start, the market has both “lemons” and “peaches.” However, since buyers lack information about a vehicle’s history (which the sellers have), they have no way of knowing which is which. As such, the going price for a car, regardless of quality, stabilizes somewhere between the true value of the two. This leads to perverse incentives, since the market rate is higher than the true value of a “lemon,” but lower than that of a “peach.” Owners of “lemons” stand to profit, and put their cars up for sale, while owners of “peaches” stand to lose, and do not. Pretty soon, the entire used car selection consists of “lemons” and nobody is happy.
While the abundance of flooded cars for sale paints a bleak picture of human nature, it seems consistent with the model of humans as rational economic actors. People can sell water-damaged vehicles with little consequence, and history shows they are likely to do so. According to CarFax, over 300,000 vehicles flooded during Hurricane Katrina are still on the road today.
Blood, Thicker Than Water?
Learning about the water-logged car issue reminded me of another topic which has flooded the news lately, the highly publicized trial of Elizabeth Holmes, founder of Theranos. Holmes’ company promised quick, inexpensive diagnostic blood tests, and raised a huge amount of capital from this claim. The business seemed revolutionary, before evidence started to mount that the tests worked poorly, whistleblowers came forward, and the whole operation ceased to a halt. Holmes now stands trial as to whether she willfully deceived patients and investors.
Regardless of Holmes’ intent, the Theranos case appears at first glance another example of the “lemon” problem. Those on the outside, VC backers and the users of Theranos products, knew little about the underlying technology. People within the company, on the other hand, likely understood the dubious ground their business stood upon. Due to this disconnect, inaccurate cancer diagnoses, huge financial losses, and other painful outcomes ensued.
The Theranos case seems to prove the idea of a broader trend toward froth in the venture capital world. Under this narrative, investors, eager to replicate the huge profits from tech IPOs in recent years, recklessly invest in startups with little understanding of the actual business. While a tantalizing story for those ticked off by braggadocios mega-investors, portraying Theranos this way may be over-simplistic. Specifically, biotech is a highly specialized field, so outsiders often find it difficult to fully audit companies in a technical capacity. Informational asymmetry is likely more extreme in arcane fields like diagnostic testing than in other industries. Aside from this, VC and used cars markets differ in another fundamental way, which I will get to later.
Moving the Chains
Despite the previous examples, a one-sided lack of information need not always lead to disappointment. I was reminded of this the other day while participating with a group of friends in our yearly fantasy football draft. For those unfamiliar, fantasy football consists of a group of “managers” taking turns picking NFL football players to join their imaginary teams. These players’ real life performances then earn points, and the teams which accrue the most points win.
One of the most perplexing draft day decisions is how to handle first-year players, or rookies, specifically at the running back position. Since novice players have no previous NFL performances, projections as to what they may do in the coming season are often more speculative.
Perhaps un-intuitively though, rookie running backs consistently outperform draft position. There are several reasons for this. Early fantasy draft picks are high leverage, and a failed pick can ruin a season, so many will prefer to choose a sure-fire superstar player rather than an unproven one. Despite this conservatism on the part of fantasy managers, rookie running backs boast a solid historical track record. In the NFL, the position is unique in that players often peak in their first few seasons, unlike tight ends for example which can take years to hit their stride. Of all the players on the field, running backs are most likely to contribute right away, and earn a whole lot of imaginary points in doing so.
Although fantasy football decision makers have little information about how rookie running backs will perform against NFL competition, or the roles their real-world teams intend for them, fantasy drafts are not inundated with poor-performing rookies. This largely has to do with the consistent quality of football players entering the NFL in a given year. Structural factors like the population of kids participating in football, the number of college programs which recruit them, the quality of training athletes receive, and the number of opportunities to go pro, change only slowly over time. Moreover, NFL teams conduct a detailed scouting process for aspiring players — they put on a draft combine, analyze college game tape, and even hold good-old-fashioned job interviews. When it comes to the choice of who gets to enter the league, information is far less asymmetric.
While those playing fantasy football may feel more confident choosing a player with a track record, in reality they pay a premium for the comfort. Here, drafters discount based on a lack of information, when they really should not be. They behave as if buying a used car, when in fact the choice they face is a lot different.
Making a Market
Returning to financing tech startups — is it more like buying a used car or picking a fantasy football team? Let’s consider the following two market characteristics to help us decide:
1) Alternatives — One important aspect of these types of markets is if sellers have other good choices than to participate in them. In the case of used cars, there’s always the option of keeping the car and driving it if the available price feels too low. Since cars are pretty useful — in America at least — this option proves compelling. Yet for talented young football players, the NFL remains the only real top-level professional league. Since they cannot make nearly as much money playing football anywhere else, there is no choice but to sign up for the behemoth which has monopolized America’s most popular sport. Starting a company is often very expensive, and similarly Silicon Valley investors provide often the only choice for procuring the requisite funding.
2) Size — Another point worth considering is if there are limits to how big a market can grow. For cars, there does not appear to be any. If you have a car, you can choose at any time to resell it, and plenty of people tout themselves as more than willing to do so for you. Rookie running backs though face a hard upper limit. Among the 32 NFL teams, each has one, possibly two running back slots in which players can participate consistently in the offense. This restriction keeps the quality of running back talent high on average. Startups more closely follow cars in this example, since no hard limit exists on the number of companies than can be funded. With low barriers to entry and money flying around, unviable businesses can bring down the average.
All this is to say that the market for tech startups lies somewhere in between— monopolistic but also unrestrained. The structural determinants of quality here work in both directions. This no doubt plays into what makes the job of picking winners in early-stage companies so hard. It also makes it hard to write off the field of venture investment altogether as bubble-y or inflated.
So What?
Anyway, none of this comparison really matters unless it applies to real life, which I think it does. Next time making a purchasing decision with less information than a counterpart, consider the dynamics of the market you are participating in. Do sellers have a compelling alternative? If not, quality options will exist. Is the size of the market restricted? If so, the higher the likelihood the option you are considering is a quality one.
By thinking closely about the dynamics the other side faces, you might just be able to work uncertainty to your advantage.
