Yes, the title of this post is about major league baseball free agency. Yes, this blog is supposed to be about insurance and/or investing. Am I suffering from mission drift already? Well, maybe not. At times, I will explore things that are off topic where the situation reminds me of something we might see in the insurance world or in capitals market.
OK, so that being said, I’m writing about baseball because I think it will get more clicks than some boring insurance topic, right? Actually, no! It really is because I think there is an interesting parallel between how cat treaties get priced and how baseball free agents get priced.
The Problem
For the last two offseasons, free agency has been a dud for players. Contract offers have been for far less than players are used to. Rather than the free agent frenzy we see in football or basketball, where players are signed within minutes of the start of free agency, baseball players are now languishing until after the start of spring training. Even with the recent signing of Manny Machado, there are still nearly 100 free agents, including Bryce Harper, waiting for better offers.
Collusion?
Why has this happened? Players want to blame things like “tanking” and “collusion”. In other words, teams are prioritizing profits over winning. While there may be some of that, I think there is a bigger villain: analytics.
Rather than spending like drunken sailors on free agents with recognizable names – but who may be past their prime – teams are using models to predict a player’s future statistics and assigning a monetary value to the added wins that a player can expect to bring.
If this sounds an awful lot like predictive modeling in say worker’s comp, well, that’s because it is! Instead of leaving the decision to the underwriter’s “feel” for the account, we are letting algorithms decide the price. This leads to fewer underpriced workers comp accounts – and fewer overpaid baseball players.
OK, so I’ve filled a bunch of white space telling you analytics has spread beyond insurance. Fascinating insights! Is this what I mean by “nominal returns”? As in nominal return on your time wasted reading this tripe? Well, as they say on the late night infomercials…
The Allegation
After signing with the Chicago Cubs, pitcher Brad Brach discussed the negotiating process.
NBCSports Chicago
We talked to certain teams and they told us, ‘We have an algorithm and here’s where you fall in that scale.’ It’s just kinda weird that all offers are the same that come around the same time and everybody tells you there’s an algorithm, but you figure teams have different ones, but I don’t know.
Now, that’s interesting. Everyone is using similar models and coming up with a similar “technical price”. Where have I heard that before? Boy, that sounds familiar? Anyone hazard a guess?
No, not comp. Not auto. Those models have enough moving parts and are all “proprietary”. After all, every auto commercial tells you they can save you $500. Those are not similar algorithms.
The Reality
So what’s the answer? That’s right, cat XOL pricing! Remember back to the pre-KRW days when every underwriter had their own model (or none at all!)? That was the “old” baseball free agency. Markets didn’t always clear at rational prices. There was always an “undisciplined” carrier who would “ruin” the market.
Over time, nearly everyone started pricing off the two major vendor models. Maybe they had their own overlay or maybe not, but the market started to clear around the IRR hurdle that was implied by the AIR or RMS loss function for the given zone. Often, you would hear about companies being disappointed they couldn’t raise pricing post a loss but that things were still “disciplined” because once you got below a certain price, all the capacity would dry up. Basically, there was infinite capacity above the clearing price and no capacity below it.
This sounds an awful lot like Brach’s comment above, doesn’t it? “All offers are the same…around the same time…there’s an algorithm”. He is describing the cat market!
This explains why the pricing response to the cats of the last two years (as well as 2011) was so disappointing to many. Long time participants were counting on psychology to drive pricing upwards. However, fear & greed don’t work the same way in the age of algorithms. With the exception of wildfire, most of the large cat events were “within the model”. In other words, if the models aren’t going to change, pricing isn’t going to change materially. The models temper the emotional response to loss.
Now, one might say, well then why has cat pricing been on a relatively steady downward trajectory for the past decade? There is an easy answer to that: supply!
Unintended consequences
Ironically, the “certainty” implied by robust models opened the path for the capital markets to enter the cat space through ILS. So, one important observation is that models that create consensus around pricing will also attract competition and lower overall pricing.
In other words, good algorithms are bad for insurance returns because they commoditize the business, which overwhelms the benefit from providing a firm pricing floor that prevents catastrophic mistakes (well, unless you used to be one of those fools who priced way below the model’s implied floor). It wasn’t ILS that “ruined” cat for traditional players. It was the uniform models that attracted ILS that ruined cat.
A Difference
Back to baseball, what prevents the increase in supply that model certainty brought in cat? Major League Baseball! MLB has only 30 teams. There is no way to create more unless they choose to do so. In a fully open market, we would see new bidders emerge that would probably be willing to overpay what the model says a win is worth in a vain pursuit of on the field glories. With competition limited, the value of a win is likely to remain fairly constant (especially with teams afraid of the luxury tax) and player contracts are likely to remain in the new equilibrium.
This leads to an interesting conclusion: while MLB is not engaged in active collusion to hold down prices, the end result does indeed look like old school collusion. Every team knows that the other teams are using algorithms and, because all price decisions are public, teams that are consistently overpaying will realize it in the market and adjust their behavior. There is no need to get in a room and participate in behavior that is going to get you sued. So this is the new reality for baseball players. They are not likely to ever see a “hard market” again (without a change in the collective bargaining agreement).
As for insurers, be careful of innovation. Developing more robust, transparent models may sound good in theory. Who couldn’t benefit from eliminating the worst segments of your book? In practice though, that transparency invites new competitors who will drive down your returns.
Want more great insights from Ian’s Blog: Nominal Returns? Look to your right and enter your email in the subscribe box!
Thoughts to share? Drop a comment below…
Good stuff Ian. Texeira was talking analystics on MLB Tonight the other night. Hard to think it’s collusion when everyone using the same model.