Hail Jeffrey Wernick

I started to write about prediction markets in 1988, and not until 1999 about using conditional markets to make org decisions. Yesterday I recorded a video discussion with Jeffrey Wernick, wherein he shocked me by saying that, for ~15-20 years starting in 1981, he founded and managed two firms using prediction markets!

First, he founded AVI Portfolio Services Company to make and sell a financial risk management model, and sold it in 1984 for a big profit. During that time,

Basically all decisions were made by people making bets. Almost every decision that had to be made to govern the company, I built a market about that.

For example, his firm grew from 8 to 20 employees, and after his partners helped him hire the first few folks, all hiring decisions were then made by markets. These markets traded assets that paid off zero if the person didn’t last six months, and if they lasted longer paid off in proportion to outside-the-firm oracle evaluations of employee performance. There were also markets like this for every employee re their next six months.

As another example, they added features to their model, and had markets in profit metrics for each feature, if it were added. They also had pricing decisions to make. They considered ~20-30 features, and actually implemented five. For all these choices they had conditional markets, called off if a condition wasn’t met, in metrics given the implementation of a feature, or the setting a price, or the choice to hire someone.

Any employee, and many firm clients, could trade anonymously (or not) in these markets, by calling or faxing a secretary who recorded trades and managed an order book. All the company’s key financial and performance info was made visible to all traders. Their lawyer designed some way to reward those best at betting while avoiding various legal problems.

At peak there were about ~15-20 markets open at a time, with open interest typically in the range of ~$75K-$250K per market, though a few markets reached a few million. Wernick says no one reported concerns about folks betting against people/features and then trying to sabotage them, though he admits he discouraged folks talking to him about anything but complaints about him or advice from him.

After those three AVI years, Wernick started a real estate venture, and let his 60-80 employees and tenants trade in those markets, which continued for ~10-15 years.

He also moved into venture capital, where he tried offering to founders that he’d take ~25% less if they’d use what he saw as good governance practices, including prediction markets. They’d sometimes accept other conditions, but never prediction markets, though ~1/3-1/2 of his added incentives were tied to that.

They reason they gave was that they saw their own personal judgments as superior to markets, and didn’t want bad decisions to result. Wernick says he never thought to have managers give themselves a big enough budget in such markets to ensure their preferred decisions always happened. I think in that case they’d initially get their way, but learn to give way as they lost on average to other traders.

Sadly and strangely, none of the many associates who saw all these markets ever tried to copy them elsewhere.

Wernick says he was inspired by meeting Ronald Coase at U. Chicago, learning his theory of the firm, and wondering if markets could go further to displace hierarchy.

Wernick blames the ego of managers who think their judgment best, hire sycophants, and keep key org info close to their chests. But he liked the idea I explained to him of using prediction markets to manage decisions that managers do not now control, such as how much funding a firm is allowed to raise when from investors, or mutual decisions made with key suppliers and customers. Thus there’s hope.

Wernick stopped all this when he got divorced, moved to Mexico, and became a “hermit” for a long while.

So, wow, my dream was actually tried successfully years before I even thought of it!

Added 10Dec: I have many questions I will ask when I talk to him again.

Added 12 Dec: I just talked to him again, and added resulting info to the above.