The reverse Lemon Paradox

Or how to Beat Sellers in Secondary Transactions with the Power of Data & AI

Or how to Beat Sellers in Secondary Transactions with the Power of Data & AI

[This article was initially published as a guest post on Data Driven VC Newsletter]

I used to live in Washington, DC. The city is manageable without a car, but my employer at the time, Orange Business Services was based out of Herndon, VA and a car was the only way to commute there.

So fresh from Paris, I told Dan, Brian and Steve, my new American colleagues that I was after some tips to buy a decent and affordable second-hand car as I couldn’t afford a new one. The first sentence I got back was “so are you one of those “tyre-kickers” who will end up buying a Lemon ?”.

They explained that a “tyre kicker” is someone who is content with kicking the tires with an expert air, pretending to assess them even if they don’t know much; and that a lemon is a car with a significant defect or malfunction that makes it unsafe to drive!

Obviously, it’s not always easy to determine if a second-hand vehicle is in as good a condition as the seller claims. So, there is a risk that you buy a car which breaks down shortly after being purchased.

This concept goes beyond my personal anecdote: in economics, a lemon paradox describes a transaction where the seller has more information than the buyer. Lemon markets were studied in detail in the 1970s by American Economist George Akerlof (who also happened to be professor at Georgetown, Washington, DC… I wonder if it’s a coincidence).

Most purchasing transactions of our day to day lives display symmetrical information: whether you’re buying a pair of jeans, a piece of furniture or a smartphone, you know exactly what you’re buying and if the product proves faulty there are laws protecting the consumer who can get an exchange or refund.

Asymmetry of information is more typical in a second-hand transaction as there is less consumer protection. But for most items (a piece of cloth, of furniture etc..) the buyer can inspect the product and detect potential defects relatively easily.

A car is a complex product, and it’s sometimes possible to hide potential issues, especially to a buyer with limited knowledge in mechanics.

Now, why am I mentioning this ?

In the world of tech investment, any transaction can be seen as a typical case of asymmetry of information: i.e. the investor is buying shares of a company without knowing everything and has to rely on the word of the seller (the founders).

But in practice, this is relatively controlled: the founders stay onboard and continue working with the investor many years after the transaction, so they have limited incentive to hide anything, and the remaining uncertainty is part of the business of investing. Finally, there are usually thorough Due Diligence done by professionals before the transaction is closed.

Secondary transactions, i.e. where an existing investor in a tech company is looking to sell its shares to another (and new) investor, represent a more interesting case from that perspective:

Let’s look at the typical sellers’ profiles. They can be of different types, but the most commons are

- Early employees who may or may not be with the company anymore

- Business Angels who invested very early on and have never been particularly involved in the governance.

- Other early investors, seed funds or family offices who often used to be on the board but lost the privilege as they got diluted with subsequent investment rounds.

We can notice many similarities with the second-hand car examples:

- most of the time, the seller is fully exiting the company. They may regret having sold too early if the company performs well, but they won’t be involved with the company after the transaction is concluded.

- They usually don’t providing extensive Representations and Warranties that could somewhat protect the buyer

- The management of the company often doesn’t have time to get involved in the transaction. After all, the transaction will not bring additional cash into the company so except if the buyer is a well-recognised added-value investor, their time is better spent on other matters.

What about the potential buyer?

In most cases, the buyer has even less information than the seller, because they haven’t previously been involved with the company and because most secondary transactions are not intermediated by a banker or advisor.

That’s where the power of data and AI comes in.

In previous articles, I’ve highlighted how Red River West is a very data-driven VC fund, collects millions of data points from public or private sources on any given startup, structure quantitative or qualitative data, and use Natural Language Processing and smart algorithms to calculate “scores. With these modern techniques, we can assess the momentum of a company, rate its culture, the quality of its recruits, identify new customers or customers which churned, potential customer dissatisfactions etc…

Such a data platform doesn’t require the involvement of the company, nor the seller. It provides critical knowledge to the prospective buyer. A data-driven secondary investor can be in a situation where they do know more than a seller who often receives nothing but a yearly high-level report from the company.

This is a rare case of reverse lemon paradox: the buyer has more information than the seller. This is especially true for recent information. In a world that changes fast, a yearly investor report is quickly outdated whereas the digital footprint of the company can tell a different story

Here are a few examples of negative elements a data platform could identify

- Departure of key employees

- Emergence of a well-funded competitor

- Slowdown of revenue growth (or decline in revenues)

- Customers dissatisfactions

- Etc..

In his 1970’s theory on lemon markets, George Akerlof explained that the asymmetry of information on a second-hand car, leads the buyer to estimate the car to be of average quality; and therefore, only willing to pay the price of a car of known average quality.

With the same, but symmetrical reasoning, for secondary transactions in the tech ecosystem displaying asymmetry of information towards the buyer, the seller has to assume that the asset is of average quality; and consequently is likely to use basic valuation methods from average companies in the same sector (typically revenue multiples), whereas the data-driven buyer is able to leverage her knowledge and convert it into an ability to price the transaction in an optimum way (for example, factoring a better view on growth or market conditions) and have a better chance to maximise return on investment.

I did buy a second-hand car in Washington, DC at the time. A couple of months afterwards, two tires blew up at the same time on the highway (no harm)… The rest of the car and the engine in particular were perfectly fine but I guess I should have inspected the tires carefully rather than merely kicking them ;-)

OH.