Berkshire Hathaway - Insurance Operations

[Guest longinvestor]

After all Insurance deserves a separate topic here. John Hjorth suggested that also. The general news section has become unwieldy. Besides, what's BRK without Insurance?

 

There have been some major shifts over the recent past. Ajit Jain now runs all reinsurance, Geico motors on towards the #1 spot, Float generation is growing at a healthy clip, despite our Chairman's guidance that it could flatten, float is at $110 B or so, insurance earnings is steady enough to earn a regular place in the earnings table, lots of idiot capital coming into insurance, it has been a benign claims period for sometime now, and ...

 

What does the future hold? Insurance is the one segment that I believe runs on it's own, with Buffett playing little more than cheerleader.

 

Chime in. If you are an insurance industry insider, even more welcome!

 

 

 

 

 

 

[Guest longinvestor]

So, what is the Insurance business worth? What proportion of BRK's IV is insurance?

[John Hjorth]

Thank you for doing this split, longinvestor, - it's a great idea & initiative - for future ongoing discussion and information sharing about the insurance parts of Berkshire here on CoBF.

 

- - - o 0 o - - -

 

I have attached a screen shot from the latest Berkshire Q-10, p. 24, lower part - please focus on the "Berkshire Hathaway Reinsurance Group" line.

 

What do you see?, and what do you think about what you see?

[Guest longinvestor]

Thank you for doing this split, longinvestor, - it's a great idea & initiative - for future ongoing discussion and information sharing about the insurance parts of Berkshire here on CoBF.

 

- - - o 0 o - - -

 

I have attached a screen shot from the latest Berkshire Q-10, p. 24, lower part - please focus on the "Berkshire Hathaway Reinsurance Group" line.

 

What do you see?, and what do you think about what you see?

 

Okay, I will do the  financial media headline first. Berkshire Hathaway earnings disappoint analysts due to weakness in Insurance division's  earnings. 'm not making this up. Those were actual headlines.

[Dynamic]

I'm interested in the potential impact of autonomous vehicles and prior to that many collision avoidance systems built into most Teslas and increasing numbers of other cars, working the way down from the premium end of the market to the mainstream.

 

I welcome this as a way of saving lives, but if self-driving cars were to reach level 3 to level 5 autonomy in the next 5 years and then perhaps the new vehicle fleet (cars, buses and heavy goods transportation) were to be almost all autonomous about 10 years from now, gradually bringing the majority of the US vehicle fleet to full autonomy over it's 15-20 year life cycle we could see insurance premiums and float from GEICO gradually decline to perhaps a third or less of their current levels by about 2040-2050 (inflation adjusted).

 

Such changes might also reduce car ownership if the cars can perform Transport as an Application so much cheaper than Taxis and human-driven Uber/Lyft cars. I suspect the total miles driven over the fleet would be similar in such a scenario, but with fewer, newer cars having much higher utilization (perhaps up 10-fold or 20-fold from the 4% utilization nowadays).

 

I'd also imagine that there could be knock-on effects in other industries - e.g. sale of alcohol, if people can get around safely and legally when they're wasted! And as autonomy is likely to coincide with electric vehicle adoption as it becomes cheaper than gasoline/diesel, the oil industry as a fuel source is likely to decline gradually over the next 2-3 decades, and with falling demand, so might the oil price and the use of pipelines. Use of oil in aviation and sea shipping may be more persistent and also the use as chemical feedstock.

 

But, this is an insurance thread, so I'm thinking we might need to expect that the GEICO portion will continue to grow float for a decade or so, but could be heading for a gradual decline at perhaps 2-5% per year. GEICO's superior cost structure and expense ratios may see it able to thrive and take market share from competitors with higher costs in a market where insured losses and thus premiums are expected to fall as increasing autonomy reduces the accident rate and severity, so expenses in running an insurer could become more significant to the premium. So I don't see float vanishing overnight, and have ground for optimism that GEICO could compensate with market share gains.

 

Do you agree or disagree with anything I've written?

[Guest longinvestor]

The hypothesis that needs to be verified,

 

Null; Myu=0

Alternate; Myu>0

 

Myu is the average# of accidents experienced by drivers driving autonomous vehicles.

 

If Null hypothesis is validated, auto insurance as we know it today will be different. Who knows, it's gone. In a sense, automation will have overcome human behavior (finally, whew!). Won't that be the headline?

 

If the Null cannot be verified, the alternate hypothesis holds, little will change. Human behavior will continue to be the reason for things like insurance. I suspect that I'll be dead before this. But my kids may, so I look for null to be verified in their lifetime.

 

 

 

 

 

[Dynamic]

The NHTSA in reviewing the headline-grabbing fatal Tesla Model S Autopilot crash, acknowledged that the collective crash rate of drivers when using Autopilot was 40% better than the normal rate for human drivers.

https://techcrunch.com/2017/01/19/nhtsas-full-final-investigation-into-teslas-autopilot-shows-40-crash-rate-reduction/

 

It seems that we've reached the point that when Tesla Autopilot is able to remain engaged (which doesn't mean in any environmental conditions on any road yet - that would be level 5 autonomy - and we're not quite at level 3), it already has a lower crash rate than humans.

 

Myu is >0 for sure, but I suspect it's much better than the status quo, so will very gradually lower crash rates as it slowly becomes more widespread and begins to get better, eventually reaching levels 4 and 5, which would mean that I would anticipate a gradually lowering but still real crash rate across the fleet, even if autonomy spreads as rapidly as anti-lock brakes or airbags did. But perhaps somewhere in the decade 2020-2029, insurers and maybe even emergency rooms will begin to see the effects of a declining accident rate, albeit gradually, though the decline might accelerate sharply once the level of human drivers declines below a certain point.

 

Although not fully relevant to insurance, I could also see autonomous vehicle making a major difference to the trucking industry, where the economic effects of a level 5 fully autonomous vehicle being able to drive 24/7 could be enormous, and likewise to the railroads if the cost of trucking declines dramatically and reduces the railroad's inherent advantages.

[DooDiligence]

https://www.forbes.com/sites/patricklin/2016/04/25/self-driving-cars-wont-kill-insurance-industry/

 

*** Quote from article ***

 

Imagine that you step into a self-driving car of the future. The car’s heads-up monitor issues an alert: “I sense that you’ve been drinking. I probably don’t need your driving assistance, but there’s always a slim chance I may, in the event of an unexpected emergency. If you want to operate this vehicle under the influence of alcohol, your insurance rate for the next 12 hours, or until you arrive at your destination, will increase to $2 per hour. Press here to accept this increase in your insurance.”

 

You might get a similar alert if you’re driving late on a weekend when other drivers may be drunk, or in heavy traffic, or at high speeds, or if you’re parked in a neighborhood known for break-ins, or at any other time that increases risk.

 

Thus, one alternative business model in insurance could be to charge for these micro-risks, and even with micro-payments of pennies per hour (or smaller increments of time) or per mile.

 

Of course, this scenario would be a privacy nightmare. And, as insurers get closer to a state of perfect information, it may force a question about the ethics of their business: is it problematic to offer insurance bets when the outcome is known?

 

*** End quote ***

 

Why not, casinos do this every day?

(albeit, with minimal liability re: bodily injury.)

 

---

 

As an aside, people will want to be entertained while riding.

 

Game developers could work micro-premiums into your fare?

 

A rider could actually choose to go into an area known for robbery & vandalism & place bets on their ability to navigate thru with minimal damage.

 

---

 

Just throwing crap against a wall here.

 

---

 

*** Quote from article  ***

 

To help insurers and policymakers better understand capabilities and risk, we might need to consider inventive proposals, such as creating new positions in government and even a new agency to concentrate technical expertise...

 

*** End of quote ***

 

(inventive proposals? uhhh, you lost me there buddy)

[DooDiligence]

on a more serious note...

id-market-place-of-change-automobile-insurance-in-the-era-of-autonomous-vehicles.pdf

[Dynamic]

So, what is the Insurance business worth? What proportion of BRK's IV is insurance?

 

Hmm, I'll give the autonomous vehicle thing a rest. This is really meaty.

 

To get a ballpark, we could consider a few things and sanity check them.

 

Please treat this as thinking aloud by a non expert, not a fully reasoned analysis. I'd like to see how others might go about it and see if there's a better model to valuing the insurance business.

 

I'll start with just one part of the picture, and see what you think:

 

The value of the underwriting profit

The 2016 underwriting income was $2.131 bn.

 

For a sanity check, over the 14 years of consecutive underwriting profit, a time during which float and underwriting have grown, $28 bn has been earned. The average over 14 years is $2.0 bn per year.

 

And Berkshire tends to underwrite conservatively, with 'loss development' for long-term risks tending slightly to work out in the shareholder's favour, not the norm in the industry. In fact, some of the 'losses' in the insurance division in the last 6 months may be as a result of such conservatism. Berkshire does not need to present the biggest underwriting profit it can, nor to pay taxes now on profits presumed by rosy projections of future insured losses that might not materialise.

 

Yes, only picking the 14 'up years' is slightly cherry picking to avoid the down year that preceded it, but I think GenRe's problems at purchase are now fixed and the trend is upward.

 

So if we are to capitalise some sort of underwriting profit, and still anticipate growth in the coming decade and perhaps the odd down year, despite the reduced megacat exposure, I think we should anticipate an average Underwriting Profit of $2.0 bn per year, the lower of the 2016 and the 14 year average is probably neither too conservative nor too optimistic if we capitalise at a reasonable yield.

 

Perhaps capitalise that at a 10% yield pre-tax or about 7% post-tax (P/E = 14) - what do you think? The underwriting earnings power might be worth $2,000 mn / 0.10 = $20 bn.

 

If there were to be a mega-cat, I anticipate that Berkshire would make up in the future what it lost in that year through increased business (as a good payer, a company that survives and an opportunist value investor) and in increased premiums during a 'hard market' following major losses, so I'm not too concerned to make allowances for possible down years, especially as I'd anticipate decent growth in underwriting earnings to add to the total return.

 

The next step would be to value the investments and/or account for float, but that could be another post.

[Jurgis]

I think people are making mistake thinking that you need level 4-5 autonomous driving for the number of accidents to drop significantly and impact insurance companies. IMO, active collision avoidance systems are likely to drop the number of accidents almost as much as the level 4-5 and they will come much sooner.

 

Of course, you still need 10 years+ for the whole auto inventory to get collision avoidance, so I'd say there are ~10 years+ for insurance companies. Also, there will be time IMO when the insurers still charge higher premiums and get extra profits even while the accidents are declining. (The corollary though is last year when Buffett said that Geico had lower margins due to more distracted driving (presumably)).

[gfp]

Big factor is that repair costs (all those sensors in the bumpers, fenders, mirrors, etc) have gone way up.  Lawsuit costs continue to be a very big portion of rates.  Accident frequency and severity has only increased since all these safety features have been introduced.  Likely because of the large screen smart phone.  Auto insurance will always be required because of the liability aspect.

 

GEICO has been making a calculated bet to sacrifice near term profits in order to grow their business through this period - customer growth at GEICO has been outstanding recently and new customers are not profitable their first year with GEICO.  I think it's a good strategy as GEICO will retain those customers if history is a guide.

 

GEICO quoted me a policy at a significant discount to USAA, which is anecdotal as far as how aggressive they are being on price right now.  USAA isn't known as a pricey insurer.

 

I think people are making mistake thinking that you need level 4-5 autonomous driving for the number of accidents to drop significantly and impact insurance companies. IMO, active collision avoidance systems are likely to drop the number of accidents almost as much as the level 4-5 and they will come much sooner.

 

Of course, you still need 10 years+ for the whole auto inventory to get collision avoidance, so I'd say there are ~10 years+ for insurance companies. Also, there will be time IMO when the insurers still charge higher premiums and get extra profits even while the accidents are declining. (The corollary though is last year when Buffett said that Geico had lower margins due to more distracted driving (presumably)).

[Dynamic]

I think active collision avoidance carries the risk that people will engage in risk-compensation behaviours and start driving with a small gap to the car in front etc. or other risky behaviour that they believe their car will bail them out of.

[Jurgis]

I think active collision avoidance carries the risk that people will engage in risk-compensation behaviours and start driving with a small gap to the car in front etc. or other risky behaviour that they believe their car will bail them out of.

 

As if they don't do that now without collision avoidance.  ::)

Anyway, you might be right somewhat.

 

At this point I would not change my investment behavior even if I was invested in a pure auto insurer. They definitely have 10+ years of mostly-not-much-change business conditions. And anyone who tries to predict what happens after 10+ years is likely deluding themselves.

[Dynamic]

I wouldn't want to predict too much or try to create a precise valuation, but I'd be wary enough of the potential disruption to ensure I bought with sufficient margin of safety to limit my losses or make back my investment in a relatively few years.

 

I was late recognising that local newspaper monopolies had been comprehensively supplanted by the internet and smartphones for a huge chunk of their earnings and it wasn't just the credit crunch that was hitting their ad sales and portion of the older generation who preferred print ads wouldn't be enough to leave very much value in their business model. That was probably my last big investing mistake that I'm so far aware of.

 

With regard to Berkshire, and even GEICO specifically, I think there's at least 15 years of reasonably good business ahead even if these technologies cause a fairly major disruption in new car sales pretty fast. Even if autonomous Transport As A Service becomes dominant over individual car ownership, the sheer number of miles driven by each TAAS car every year should allow for a fairly good insurance earnings in aggregate, even if the fleet of vehicles declines fairly rapidly and people with older cars in the fleet move over to TAAS because it's so much cheaper. A time of disruption will see some who fail to adapt go bust, but I think GEICO will adapt and take advantage of offering insurance to TAAS vehicles.

[Guest longinvestor]

I think active collision avoidance carries the risk that people will engage in risk-compensation behaviours and start driving with a small gap to the car in front etc. or other risky behaviour that they believe their car will bail them out of.

 

+1

 

Even without sensors and such, in my family, we've had two rear end collisions within the last 10 years. It's not the car you have nor your driving behavior. It's also the behavior of the others on the road. In my case no tech on my car would have helped. I will admit, I do not drive at speeds that are clearly way over the limit. I live in Chicagoland and based on anger I can visibly see behind me in my rearview mirror, autonomous vehicles won't work for me. There are school bus drivers around here who drive recklessly and will absolutely tail you. I refuse to drive faster than what I want to, no matter what the idiot behind me wants me to do.

[james22]

Here’s one more example of how central banks’ global coordinated monetary stimulus in the wake of the financial crisis has increased systemic risk in the US: According to an analysis conducted by BlackRock, insurers are more vulnerable to a market downturn now than they were ten years ago.

 

The reason? Ultralow interest rates have forced insurers to venture into markets with higher yielding assets, forcing them to stomach more risk along the way. Whereas insurers once tended to adhere to only the safest types of fixed-income products – typically highly rated government and corporate debt – they’re increasingly buying exposure to risky high yield and EM products, along with illiquid private equity funds, to try and boost their earnings back to pre-crisis levels.

 

These products carry a potentially higher reward for insurers, but heightened risks are also omnipresent. In a downturn similar to the 2008 crisis, BlackRock estimates that US insurers' holdings would drop by 11% - even more than they did during the crisis. Such a drop would be tantamount to $500 billion in losses.

 

http://www.zerohedge.com/news/2017-08-29/insurance-companies-could-face-staggering-500b-loss-during-next-downturn

 

Might be an opportunity to buy up come any downturn.

[Cigarbutt]

It also depends where you are in the underwriting cycle.

 

https://www.canadianunderwriter.ca/insurance/m-best-projects-combined-ratio-100-3-u-s-industry-2017-1004108128/

http://www.insurancejournal.com/news/national/2017/06/29/456156.htm

 

But reaching for yield seems to be prevalent these days and, given the unavoidable volatility sometimes down the road, industry players with a sufficient margin of safety on the asset side are likely to benefit.

 

Time to work on the watchlist.

 

[John Hjorth]

I'm not trying to clog up this topic with a Buffett succession speculation post, but have to bring up a cut from a Bloomberg article today: After Decades of Hints, Buffett's heir May Now Be More Apparant.

 

... A key distinction between the wo executives is age: Jain is 66, Abel is 55. Buffett is proof that the CEO can do well by shareholders long past typical retirement age. Even so, Jain has been facing some health challenges that could eventually make working more difficult, according to people who've recently spent time with him. ...

 

Say it isen't so.

[Guest longinvestor]

I'm not trying to clog up this topic with a Buffett succession speculation post, but have to bring up a cut from a Bloomberg article today: After Decades of Hints, Buffett's heir May Now Be More Apparant.

 

... A key distinction between the wo executives is age: Jain is 66, Abel is 55. Buffett is proof that the CEO can do well by shareholders long past typical retirement age. Even so, Jain has been facing some health challenges that could eventually make working more difficult, according to people who've recently spent time with him. ...

 

Say it isen't so.

Same sentiment. Ain't no known substitute for him. Carole Loomis called this type of scenario a few years ago. Stop fussing over Buffett 's age. Worry about the next guy. Especially as Buffett keeps motoring on, like the energy bunny.

[Mephistopheles]

Successor is Ted Weschler, I just found out through my psychic reading abilities.

 

Jk. This focus on Jain/Abel is silly. There are like 80 execs at Berkshire. And he won't pick someone in their 60s to be CEO; it will be someone much younger.

[John Hjorth]

I think we also here have to reply on the Berkshire system built up over many years by Mr. Buffett and Mr. Munger, in this particular case the input to Mr. Buffett from Mr. Jain related to the memo every second year from Mr. Buffett to his superstars at the Berkshire subs about succession in the subs.

[John Hjorth]

I'm not trying to clog up this topic with a Buffett succession speculation post, but have to bring up a cut from a Bloomberg article today: After Decades of Hints, Buffett's heir May Now Be More Apparant.

 

... A key distinction between the wo executives is age: Jain is 66, Abel is 55. Buffett is proof that the CEO can do well by shareholders long past typical retirement age. Even so, Jain has been facing some health challenges that could eventually make working more difficult, according to people who've recently spent time with him. ...

 

Say it isen't so.

 

Posted by Munger_Disciple in a separate topic here on CoBF: Here.

 

He established the Jain Foundation, the mission of which is to cure limb-girdle muscular dystrophies caused by dysferlin protein deficiency.

 

I'm speculating about if there is a connection here.

[Voodooking]

It is Ajit's son who unfortunately suffers from these health issues... I heard him discuss it in an interview.

[John Hjorth]

It is Ajit's son who unfortunately suffers from these health issues... I heard him discuss it in an interview.

 

Voodooking,

 

Source?