Back with another Brief since too many things to talk about and not enough time to write them all up separately…
That Didn’t Take Long!
It was only a few weeks ago I outlined a new business model for life insurers. I suggested established insurers should focus their energy on distribution and use other markets like private equity to bear balance sheet risk.
Only days later, we saw MassMutual essentially implement this approach with their offer to buy American Equity. MassMutual is partnering with Athene on the deal. Mass would only keep 20% of the book and reinsure 80% to Athene.
However, Mass would get full operational control including distribution. This allows Mass to reduce its capital commitment and concentrate on product and distribution and increasing operational efficiency. This is where the real value is.
The one problem is the deal is being structured as reinsurance so Mass now has a large recoverable from Athene which will only grow over time. It would be better to follow the rest of my advice which is to let Athene write 80% of the new business on its own paper. This would allow Mass to act more as an MGA and reduce its financial risk.
ESG: Theory vs. Practice
On September 29, 2016, Allstate CEO Tom Wilson wrote an editorial for the Washington Post criticizing the great Milton Friedman for his belief that the purpose of a corporation is to maximize profit. This was a forerunner to the Business Roundtable endorsing a new corporate purpose last year focused on placing employees, the community, and other interests above profits.
In the editorial, Wilson wrote the following:
Today, corporate leaders are graded on stock price, not on the amount of good their companies do. We must broaden our evaluation of corporations beyond share prices to provide space, light and water for their role to grow.
Shareholders should be asking how corporations are building intangible assets such as customer relationships, their employee bases and their reputations, not just pushing for share buybacks.Washington Post, 9/29/2016 (emphasis added)
Nearly four years to the day later, Allstate laid off 3800 employees to save costs and “maximize profit”. Note, just a week earlier, Allstate announced a new $750M stock buyback. Oh, irony, you are a worthy adversary!
Do What I Say…
So Tom, I’m asking, what are you doing to prioritize employees over share buybacks?
Now, I don’t mean to single out Tom. I know him well and respect what he has done for Allstate. He wasn’t the only one who endorsed the Roundtable statement (he didn’t sign it as Allstate is not a member). He isn’t the only one who has acted against what he espouses.
On the same day, Goldman Sachs announced its own layoffs and JP Morgan did earlier in the week. Disney, the happiest place on earth, laid off 28,000 the day before. All endorsed the Roundtable statement.
Being a CEO means making tough decisions. Sometimes, those decisions require letting employees go. That’s part of the territory and they shouldn’t be criticized for that. With few exceptions, most CEOs don’t enjoy firing their people.
However, don’t then tell me how you are prioritizing employees and shareholders need to take a back seat. If you really hold dear the values the Roundtable promotes, then keep the employees on board, let earnings shrink, and let the stock price drift lower.
Be consistent with your principles! I can respect that. I wouldn’t own your stock, but I’d respect you standing up for what you believe in.
Otherwise, don’t say things you’re not prepared to back up because people like me have long memories and will remind you of your empty words!
There is one other solution, by the way, that appeases both employees and shareholders! Most of the aforementioned CEOs make in the neighborhood of $20M/yr.
If the CEO believes so strongly in doing what’s right for the employees, he or she could easily re-allocate half of their compensation to give raises to employees. 10,000 employees could get a $1000 raise each if these CEOs were willing to cut their pay by half.
In fact, I would even support shareholders making a matching contribution (say in a lower dividend payout) so we can give those employees $2000 each. If you can find me CEOs willing to take the pledge, I will submit the ballot proposals to reduce that company’s dividend $10M to match!
It’s a lot easier though to ask the shareholders to take less than for the CEO to take something out of his or her own pocket.
Does AI Actually Work?
You know those people who throw out fancy words to try to make you think they know what they’re talking about it, so you assume they do because you’re too nervous to call them out on it?
I’m starting to think that’s what happens a lot when companies say they use artificial intelligence or machine learning. Certainly there are times when these approaches add a lot of value, but in the wrong hands, they can be very dangerous. Like the discussion of models earlier this week, AI seems to suffer heavily from garbage in, garbage out.
For example, ESPN has touted how they’ve used the venerable Watson – of Jeopardy fame – to suggest fantasy football trades to improve your team. Curious, I took a look. I mean if it can beat Ken Jennings, how can it not help my team?
It was laughably bad. It suggested either trades no other team would ever agree to or trades that would get accepted immediately (because I was obviously getting ripped off).
I’ve noticed this in other areas where AI is touted. The results are surprisingly naive and disappointing.
This isn’t to say AI can’t be done well. It certainly can, but the idea that because someone says they’ve used AI to test something makes the result infallible should tell you something about the person making that claim…and not something good.
Compounding of Errors
Like any other model, it needs to be tested and verified. Unfortunately, too many people are intimidated by the language and accept that everything works as intended. This brings us to Root, the auto insurer that filed its IPO this week.
I’ll have more to say on them in the near future, but it’s worth reading Coverager’s work on the issues with their AI. Namely, the AI has a tough time telling when someone is driving vs. doing another activity or being a passenger.
Now, maybe they’ve improved this over time, but if your business model is touting your ability to measure a driver’s behavior and price it accurately, you better be sure you can actually measure the driver’s behavior accurately before building a pricing model off of that data.
For any insurance business based on an AI driven pricing algorithm, you have two obstacles. First, you have to verify that the AI is actually doing what you think it’s doing. Then, if it is, you need to demonstrate that it knows the appropriate price for those risks.
This adds a degree of difficulty that investors should be asking questions about before investing.
Motivated To Moderate
OK, quick last thought. I have a bit of a reputation for asking hard questions in a way that people will actually answer them. After watching these awful debates, where questions are largely ignored and the candidate speaks about whatever is on their mind, I’d like to volunteer to moderate the last debate.
I’m pretty sure I can keep things reasonably on topic. If someone isn’t answering the question, you either jump in when there’s a pause and say “sorry, I don’t think I made my question clear, what I was trying to ask was…” or you let them finish and then say “that’s very helpful, but I don’t understand how that explains (back to the topic)?”.
These moderators just let the candidates off the hook when they riff on the Supreme Court when asked about healthcare or Covid when asked about abortion. It’s not just about coming up with the topics. You need to get the candidates to speak on the desired topics.
If the moderator isn’t competent enough to ask questions in a way that generates appropriate answers, then get out of the way and let someone else step in. I’m available.