Tidbits

[Marco Van Basten]

One interesting anecdote - a childhood friend is a radiologist in NYC.  He always thought that his job would be replaced by technology, and yet he says that he gets job offers regularly and the market for radiologists is very good in NYC and insanely good outside the city.  

[UK]

3 hours ago, Marco Van Basten said:

If I recall Chubb's 2024 annual report correctly, it gives the underwriting ratio for the company and the industry for the previous decade.  According to memory, industry's underwriting ratio was north of 97 over the previous decade.  So, now you take $20 of revenue of the top, you have $97 of expenses on $80 of revenue, which is 121.25 underwriting ratio.  I don't know how the insurance industry can survive, let alone have a positive return on shareholders' capital under those circumstances.  So clearly the writer of the report can't be bothered with facts, so if I know that he is 100% wrong and flat out lies in something that I know, how can I take anything else that he says seriously?  He might be right on something, but not because he has bothered to do the research like a card counter at a casino but like someone who bets on zero at roulette and gets lucky.  

Well, after reading it all...I thought he actually went quite soft on insurance with this 20 per cent estimate:)). I have no idea if he is right on this or just throwing a bit random numbers here, but why would not insurance industry adjust to a 20 per cent premium loss over a few years or so? Is not this in the ballpark of the normal insurance cycle, at least in reinsurance?

[Spekulatius]

The under wiring ratio includes Sg&A and customer acquisitions  / commission costs, so you could see costs getting squeezed out there. I don’t think AI is necessary, this needs to be thoroughly digitized first. LSO standardization where applicable could do wonders.

[Marco Van Basten]

35 minutes ago, UK said:

Well, after reading it all...I thought he actually went quite soft on insurance with this 20 per cent estimate:)). I have no idea if he is right on this or just throwing a bit random numbers here, but why would not insurance industry adjust to a 20 per cent premium loss over a few years or so? Is not this in the ballpark of the normal insurance cycle, at least in reinsurance?

Well, think about it.  He did not say that the insurance industry can use AI to cut costs by 20%, which by the way is probably way too high.  He said that today, if people would just shop around like AI agents will, the premiums would go down by 15-20%.  Clearly if the next decade is exactly equivalent to the last decade, and in the world of no inflation, how can the insurance industry adjust to a  20% decline in price given 97+ combined ratio?  @Spekulatius, again, he did not say that the insurance companies could cut costs by 20% and hence consumers could see a 20% price cut, he said ceteris paribus, just use an AI agent to shop for insurance and your cost will drop by 20%.  

[dwy000]

3 minutes ago, Marco Van Basten said:

Well, think about it.  He did not say that the insurance industry can use AI to cut costs by 20%, which by the way is probably way too high.  He said that today, if people would just shop around like AI agents will, the premiums would go down by 15-20%.  Clearly if the next decade is exactly equivalent to the last decade, and in the world of no inflation, how can the insurance industry adjust to a  20% decline in price given 97+ combined ratio?  @Spekulatius, again, he did not say that the insurance companies could cut costs by 20% and hence consumers could see a 20% price cut, he said ceteris paribus, just use an AI agent to shop for insurance and your cost will drop by 20%.  

I suspect his assumptions are wrong. If everyone could reduce premiums by 20% the industry must be over-earning by 20% on all those policies. That doesnt show up in the earnings (97 combined ratio).  I would bet that while a couple of people could reduce premiums by that much, they are likely to outliers not the norm.

 

Insurers wont underwrite to losses, they'll raise prices. 

[bizaro86]

2 minutes ago, Marco Van Basten said:

Well, think about it.  He did not say that the insurance industry can use AI to cut costs by 20%, which by the way is probably way too high.  He said that today, if people would just shop around like AI agents will, the premiums would go down by 15-20%.  Clearly if the next decade is exactly equivalent to the last decade, and in the world of no inflation, how can the insurance industry adjust to a  20% decline in price given 97+ combined ratio?  @Spekulatius, again, he did not say that the insurance companies could cut costs by 20% and hence consumers could see a 20% price cut, he said ceteris paribus, just use an AI agent to shop for insurance and your cost will drop by 20%.  

 

And individually if any one customer shops around they can probably save 20%. But you're 100% correct - not everyone can save 20%. If people start switching more frequently than the math of LTV/CAC changes and insurers will stop giving big discounts to new customers - they won't be able to afford it.

 

It isn't a sign of a deep thinker to assume that changing one part of a system won't result in changes anywhere else and nobody else will adjust their actions in reaction to the changes. Sometimes there really are 2nd order effects.

[UK]

2 hours ago, Marco Van Basten said:

Well, think about it.  He did not say that the insurance industry can use AI to cut costs by 20%, which by the way is probably way too high.  He said that today, if people would just shop around like AI agents will, the premiums would go down by 15-20%.  Clearly if the next decade is exactly equivalent to the last decade, and in the world of no inflation, how can the insurance industry adjust to a  20% decline in price given 97+ combined ratio?  @Spekulatius, again, he did not say that the insurance companies could cut costs by 20% and hence consumers could see a 20% price cut, he said ceteris paribus, just use an AI agent to shop for insurance and your cost will drop by 20%.  

 

Probably the dumbest answer, but I think in you scenario other customers eventually end up paying more, because the industry will not be able to operate at 117 CR very long? Capital would leave and eventually prices will adjust to some more happy equilibrium? I am not saying I would celebrate such scenario, but it is not that different or maybe even not so severe, comparing to any soft market in reinsurance or some specific insurance lines. Anyone who knows insurance better feel free to correct me.

 

It is funny we a debating this because of some fairytale, but I think these are very good questions to debate, in the light of what is going on in the marker and the fact that perhaps majority of the board is overweigh in insurance:). But I think if I wanted to really scare insurance investors, I would write about some dystopian disinflation because of AI, which would lead to a severe TAM shrinkage. Now if this could happen in western political system is another question. I would bet that some kind of universal based income happens first and it will take care of this and many other things. 

 

Edited February 24 by UK

[UK]

2 hours ago, bizaro86 said:

 

And individually if any one customer shops around they can probably save 20%. But you're 100% correct - not everyone can save 20%. If people start switching more frequently than the math of LTV/CAC changes and insurers will stop giving big discounts to new customers - they won't be able to afford it.

 

It isn't a sign of a deep thinker to assume that changing one part of a system won't result in changes anywhere else and nobody else will adjust their actions in reaction to the changes. Sometimes there really are 2nd order effects.

+1

[UK]

2 hours ago, dwy000 said:

I suspect his assumptions are wrong. If everyone could reduce premiums by 20% the industry must be over-earning by 20% on all those policies. That doesnt show up in the earnings (97 combined ratio).  I would bet that while a couple of people could reduce premiums by that much, they are likely to outliers not the norm.

 

Insurers wont underwrite to losses, they'll raise prices. 

+1

[Saluki]

US defense firm Anduril set to double its valuation with $4 billion funding, source says

 

There is no direct way to play this, other than waiting for the IPO. But despite my periodically trimming Kraken Robotics shares, it continues to do well based on the indirect exposure to Anduril through the Ghost Shark sea drone program. 

[DooDiligence]

Florida man just tryna' keep the lights on.

 

https://ricksblog.biz/florida-ranks-3-electricity-disconnects/

[Spekulatius]

On 4/28/2026 at 3:58 PM, DooDiligence said:

Florida man just tryna' keep the lights on.

 

https://ricksblog.biz/florida-ranks-3-electricity-disconnects/

I am not sure what these numbers mean.  Florida and Alabama‘s electricity disconnect rates are between ~1.5% and 2% per month!  Could it be that this is counted differently in states and some of it is due to people moving?

 

[ratiman]

This is a very small tidbit but in the last two weeks I've run into two people who could not correctly calculate change and I don't remember that ever happening before. The person behind the counter could not subtract 12 from 20 or turn 50c into coins. 

Edited May 5 by ratiman

[Fly]

1 hour ago, ratiman said:

This is a very small tidbit but in the last two weeks I've run into two people who could not correctly calculate change and I don't remember that ever happening before. The person behind the counter could not subtract 12 from 20 or turn 50c into coins. 

 

The Flynn effect peaked in the 1980s, it's all downhill now. 

[SharperDingaan]

2 hours ago, ratiman said:

This is a very small tidbit but in the last two weeks I've run into two people who could not correctly calculate change and I don't remember that ever happening before. The person behind the counter could not subtract 12 from 20 or turn 50c into coins. 

 

In a lot of US/Canada border states; for cash purchases, it is common to pay for USD 60 of items with CAD 60. Store owners letting it go as a promotion expense, as it is cheaper/more-effective than trying to change employee behaviour. 40% FX gain for free.

 

SD

[Spekulatius]

Pretty good episode from Patrick Boyle about fiscal deficits and trade:

 

Edited May 19 by Spekulatius

[Spekulatius]

Ouch.
 

[Sweet]

7 hours ago, Spekulatius said:

Ouch.
 

Presumably inflation pricing?  
 

Wonder what the Fed will try now Trump has his pick in.

 

Seems like it will be hard to lower the rate.

 

Edited May 19 by Sweet

[Spekulatius]

1 hour ago, Sweet said:

Presumably inflation pricing?  
 

Wonder what the Fed will try now Trump has his pick in.

 

Seems like it will be hard to lower the rate.

 

Yes, it’s due to inflation and higher deficits coming (war financing). I don’t think the new Fed chairman can do much.