Post New

FWIW

Definitely a lot more traditional variance presentation of business units from Greg.

Lots of moving parts.

YOY performance a little weaker and may continue in 2026.

However, when you add it all up with the model. My overall estimate of IV is slightly higher. Mainly because even though underwriting for the insurance unites declined, it is still well above historical average and my future value for IV is based on historical average and on the insurance. Operating divisions seem to be growing earnings at a mid single digit level.

Bottom ine is my overall IV goes from approximately $530 to $539. Of course a precise measurement like this is unrealistic. So lets say $525 to $550. Likely to be at least $575 by year end.

However, my guess is with the headlines and lack of buybacks and dividend prospects the stock price is likely to move back under $500 next week.

One additional item of note.

Pilot, while reporting zilch for earnings threw off 1.7 BB in net cash flow.

Not too shabby?

Definitely a lot more traditional variance presentation of business units from Greg.

Yeah. He definitely seems to have taken Charlie's comment "Greg will keep the culture" to heart. That's good.

Good letter. A bit buzz-wordy. Nothing exciting.

YOY performance a little weaker and may continue in 2026.

Yup. Op earnings down 6%.

However, when you add it all up with the model. My overall estimate of IV is slightly higher. Mainly because even though underwriting for the insurance unites declined, it is still well above historical average and my future value for IV is based on historical average and on the insurance. Operating divisions seem to be growing earnings at a mid single digit level.

Bottom line is my overall IV goes from approximately $530 to $539. Of course a precise measurement like this is unrealistic. So lets say $525 to $550. Likely to be at least $575 by year end.

As usual, my 5-groves spits out a lower number: $460-$480
Look through: $470
Shareholder Equity + Float + 1/2 Deferred Taxes: $424
Book value: $332.55

I had trouble with Partially Owned Businesses Earnings Value. It's getting to be quite trivial though (1% of value).

However, my guess is with the headlines and lack of buybacks and dividend prospects the stock price is likely to move back under $500 next week.

Agreed. Can't predict the market and don't try, but I'm not a buyer or seller.

Appreciate the comments. Can we take another swipe at comparing 5 groves.

For me.

Grove 1: $23.10 AT operating earnings for 2028 FY non manufacturing, service, and retail. Assume P/E of 20.1. Slightly higher due to recent decline in 10 YR treasury, but lower than S&P 500 (non mag 7). Keep in mind Treasury rate of 3.99% implies P/E of 25X for no growth.
Gets per share value of $222.24

Grove 2: Equities less OXY & KHC stock. Included OXY preferred. $297.7 BB less anticipated taxes.
Gets per share of $117.12

Grove 3:
KHC, OXY & Berkadia.
Based on current prices.
Gets per share value of $10.23

Grove 4:
Cash, etc.
I include railroad cash $4 BB
I include YTD estimated retained earnings of $7.1 BB
I exclude reserve of $26.4 BB
Net is $372 BB
Gets per share value of $172.45

Grove 5
Earned premiums times 3.3% average annual after tax underwriting profit performance per 2024 BRK Buffet annual letter. 2024 was 8% ish and 2025 was 6.5% ish.
So $88 BB earned premiums times 3.3% times P/E of 12.5 (estimated insurance average for peers, much lower than the operating units).
Calculates to $36.38 BB
Gets per share value of $13.43

Total share value of $539.03

Can you pinpoint the hopefully few areas we differ?

"As usual, my 5-groves spits out a lower number: $460-$480
Look through: $470
Shareholder Equity + Float + 1/2 Deferred Taxes: $424
Book value: $332.55"

Just to add another valuation model, the weighing machine model gives a year-end IV of $770K/A-share, or $513/B-share, +/- a few percent uncertainty like all models.

I don't know anyone else who uses this model, so do with it what you will. Personally, I think it makes perfect sense that price and IV should track metrics such as sales, earnings and BV. Berkshire's market cap has tracked BV almost perfectly over the last 60 years (r^2 = 0.98). In my opinion the trendline of price versus BV provides a reasonable estimate of IV at any point in time over that period.

Share repurchases are another important capital allocation option. We will buy back Berkshire
shares when they trade below our estimate of intrinsic value, conservatively determined, ensuring
that repurchases enhance per-share value for continuing owners. We may also purchase large
blocks of shares directly from major holders when the opportunity presents itself. These purchases
allow shareholders to own an incrementally larger piece of Berkshire’s businesses, without
deploying any additional capital of their own.

This may be interesting wording. Maybe new maybe not.

1) Will buy up to our estimate of IV, conservatory determined. Not 10% below or 20% below. Is this characterized differently?

2) May purchase large blocks of shares directly from major holders. Does this imply willing to pay higher than IV referenced above?

Any thoughts?

<<We will buy back Berkshire
shares when they trade below our estimate of intrinsic value..<<<

I caught this too and I thought it was quite noteworthy. Buffett has indicated he wanted to do purchases at substantial discounts to intrinsic value. I think this opens the door to quite substantial open market purchases of the stock,I would be surprised if Berkshire has not been a big buyer of its shares this quarter.

Abel made it very clear he is not interested in issuing a dividend in the near future, but he clearly opened the door to buybacks. Perhaps a lot more aggressively so than Buffett. We learned a lot.

I caught this too and I thought it was quite noteworthy. Buffett has indicated he wanted to do purchases at substantial discounts to intrinsic value. I think this opens the door to quite substantial open market purchases of the stock,I would be surprised if Berkshire has not been a big buyer of its shares this quarter.

Abel made it very clear he is not interested in issuing a dividend in the near future, but he clearly opened the door to buybacks. Perhaps a lot more aggressively so than Buffett. We learned a lot.

You guys are wrong about this. The repurchase criterion didn't change one bit from one espoused by Warren. From 2024 10-K, page K-104:

Berkshire’s common stock repurchase program, as amended, permits Berkshire to repurchase shares any time that
Warren Buffett, Berkshire’s Chairman of the Board and Chief Executive Officer, believes that the repurchase price is below
Berkshire’s intrinsic value, conservatively determined.

Nonetheless, I still think Greg would be a bit more likely to buy back stock at a given discount than WEB and would maybe repurchase more shares than Warren, especially since he made it pretty clear that dividends are Not coming anytime soon, which I found reassuring.

I don't know anyone else who uses this model, so do with it what you will. Personally, I think it makes perfect sense that price and IV should track metrics such as sales, earnings and BV. Berkshire's market cap has tracked BV almost perfectly over the last 60 years (r^2 = 0.98). In my opinion the trendline of price versus BV provides a reasonable estimate of IV at any point in time over that period.

Though I appreciate you don't make any big claims for it, this method has a gigantic fatal flaw, and I really suggest you stop calculating it or posting it. It's like being a vaccine skeptic: talking about it is not really a public service.

Yes, it's true that value has tracked a multiple of book very closely over very long periods. That makes decent economic sense, as long as the general structure of the firm is about the same. Some operating subs, some investments.

And yes, it's true that inflation-adjusted book per share has risen at a shockingly steady rate on trend since (say) 1998.

But there is a yawning and potentially fatal gap between those observations and assuming that the rate of growth will CONTINUE to be steady, which is what happens when you look at the current level of the long run trend line and think of it as a value metric. The ratio of price to *current* observable value is a meaningful thing, but the ratio of price to the trend line from the past is a trap. If the rate of value growth drops, your metric will insist for many many years into the future that the stock is getting cheaper and cheaper, and you'll buy more and more, and be surprised when you don't make any money. It just isn't meaningful at all, except in the special case that the value growth rate in future is *coincidentally* the same as the past. Not only is that uncertain, I would be very surprised if many informed observers expect that to be the case in the next 10-20 years.

Jim

"But there is a yawning and potentially fatal gap between those observations and assuming that the rate of growth will CONTINUE to be steady, which is what happens when you look at the current level of the long run trend line and think of it as a value metric."

The model does not assume that the rate of growth will be steady. It only assumes that if the value metric, in this case BV, increases, then price and IV will increase as well. From 1981 to 1999 BV grew 29%/yr, but since then BV has grown 10%/yr. Nevertheless, the share price, and I argue the IV as well, has continued to track BV almost perfectly. The model does not plot BV versus time, or stock price versus time. It plots stock price versus BV. It makes no assumption about the rate of growth of BV or IV.

"The model does not plot BV versus time, or stock price versus time. It plots stock price versus BV."

I apologize that I have not made myself clear. I did post the graph once, price versus BV over the period 1965 to present (log-log), and the trendline (power law), not of BV or stock price versus time, but of stock price versus BV.

"Yes, it's true that value has tracked a multiple of book very closely over very long periods."

A constant P/B is not required by the mode.

"And yes, it's true that inflation-adjusted book per share has risen at a shockingly steady rate on trend since (say) 1998."

A steady rate of increase of BV is not required by the mode.

I like and admire you, Jim, but you keep insisting that I am saying something that I am not saying. I am not talking about rate of growth of BV or stock price. I am not talking about constancy, or lack thereof, of P/BV. I'm just talking about price versus BV.

It makes no assumption about the rate of growth of BV or IV

Perhaps I have misunderstood what you're doing, in which case I apologise.

But you said "In my opinion the trendline of price versus BV provides a reasonable estimate of IV at any point in time over that period."

Simply put, no it doesn't.

Or rather, it may have by coincidence in the past, but there is no reason at all to think that it will do so now or in future. The current point on any trendline built from past data should be ignored, whether it's price or book or any other value metric. It's the BMW chart fallacy.

Jim

Here's the graph. The trendline, of course, is the trendline price. It's only an assumption that the trendline is a reasonable estimate of IV.

https://docs.google.com/document/d/1zl_L--clGQlFSj...

Here's the graph...

Indeed, that graph doesn't depict what I thought it did, so my apology applies.

But I still can't for the life of me figure out what you're trying to do here.

It is in essence a graph of the price to book value over time. The use of log/log makes sense to get all the points to fit on the graph, but I don't see any way in which it's meaningful to the problem. The growth rate over time has slowed making the data points more clustered where the numbers are bigger in log space, so the fit gives higher weight to recent data points, but that's a small effect.

But all it is really saying is that
* the average price to book ratio has been in a relatively narrow range (or at least seemingly so if you compare it to the 4-5 orders of magnitude of growth over time), and
* that formula fit suggests that P/B has been rising over time. e.g., with book value 50k the fit formula suggests P/B of 1.34, with book value of 100k it suggests 1.46, and with book value of 500k it suggests 1.545.

If you consider the current point on that trend line, feeding in book to get a notion of the "normal" price to expect, you're implicitly assuming that the very long run log trend of P/B will hold, and that P/B will continue to rise. I for one would not place any weight at all on either of those assumptions.

My two concerns here are that (a) it doesn't seem to be a very wise way to come up with the trend of P/B compared with simply looking at P/B figures through time, and (b) it doesn't seem to be true in any useful way. Consider the P/B in the last 30-or-so years. In the first half it averaged 1.726, and in the second half it averaged 1.400. Not much of a long term up-trend.

I think you're better off with KISS. Estimate a reasonable guess of what the "normal" P/B might be now and in the next few years, informed by history. "Informed by" meaning in whatever way seems most appropriate to your observations of the trajectory of the slowly changing structure of the firm. Multiply current book (or peak-to-date book) by that number.

Lately I've been estimating price trajectory by looking at possible future average P/B numbers of 1.37, 1.40, 1.45 and nominal growth rates of 1.5%, 2.0%, and 2.5% per quarter, just to get a cloud of reasonably plausible futures over the next couple of years. As a hobby I've been writing calls whose exit dates/prices are entirely above the cloud, and puts whose entry dates/prices are entirely below the cloud.

Jim

Tight graphs like that are intrinsically interesting, nice work!

I'm not a dog in this hunt, but FWIW, a walk-forward analysis comes to mind:
1) Use, e.g., the first ten years of data to start off. Determine the trend line from just that past ten years of data.
2) Then advance a quarter and redetermine the trend line from just that past data that now includes one additional past quarter.
3) Repeat - keep determining the trend line up to time 't' using only data up to 't' (never using data to the future of 't').

The 'burn-in period' need not be ten years, that just came to mind.

It might be of interest to segment the data into two pieces: pre-1998 and post-1998.

IIRC, Jim used post-1998 data to show show that BV/share was predictive of future smoothed three and five year returns.
What you do with the walk-forward data depends on what your goal is, but at least 'walk-forward' addresses the issue of 'future-peaking' in what you're doing.

Tight graphs like that are intrinsically interesting, nice work!

I really don't think it deserves that view. If it weren't for the log-log format, the wild variation of P/B would be apparent (expensive days have been around 3.3 times the valuation multiple of cheap days). All it's really saying is that both price and book per share rose by more than four orders of magnitude in the last 50 years. Since those end-to-end rises are much bigger than the 3:1 variation range of valuation multiples, it looks like a meaningful fit.

The bigger problem is that the suggested interpretation of the end result, with the fit line suggesting that a "normal" P/B right now would be 1.545 on that trend line, doesn't pass the sniff test given the 15 year average of 1.40. No, I'm sorry, it's utter nonsense and it's just plain wrong.

If you want to have an idea of what a normal price would be now, pick any decent metric of observable value, ponder its typical relationship to market price in the modern era, and apply that multiple to the current level of your metric.

Jim

Appreciate the comments. Can we take another swipe at comparing 5 groves.

For me.

Grove 1: $23.10 AT operating earnings for 2028 FY non manufacturing, service, and retail. Assume P/E of 20.1. Slightly higher due to recent decline in 10 YR treasury, but lower than S&P 500 (non mag 7). Keep in mind Treasury rate of 3.99% implies P/E of 25X for no growth.
Gets per share value of $222.24

I got $23.20b and think of a multiple that I would buy at. I wouldn't buy these businesses at 20x.
Treasury rate is as risk-free as we can get. Earnings yield - real or expected - on stocks has to be higher to compensate for the risk.
I use 15x. A 6.7% earnings yield. Add some growth get to my 10% hurdle rate.
$161 per B

Grove 2: Equities less OXY & KHC stock. Included OXY preferred. $297.7 BB less anticipated taxes.
Gets per share of $117.12

Same, $117 per b share. Jim cuts the Apple stake a bit. I didn't but do see the reasoning.

Grove 3:
KHC, OXY & Berkadia.
Based on current prices.
Gets per share value of $10.23

I had been trying to use earnings, but it's a mess. Using value of $20 billion I get $9 per b share.

Grove 4:
Cash, etc.
I include railroad cash $4 BB
I include YTD estimated retained earnings of $7.1 BB
I exclude reserve of $26.4 BB
Net is $372 BB
Gets per share value of $172.45

I've been copying Jim and deducting 30% of float value. Jim explains it better. It's a similar figure to Bloomstran's maximum annual insurance claims deduction.
I don't include the subs cash - figure they need it, so it's included in their valuation.
I didn't include YTD estimated retained earnings.
Cash & Fixed Income Available for Investment $155 per b share.
Add your YTD estimated retained earnings $158 per b share.

Grove 5
Earned premiums times 3.3% average annual after tax underwriting profit performance per 2024 BRK Buffet annual letter. 2024 was 8% ish and 2025 was 6.5% ish.
So $88 BB earned premiums times 3.3% times P/E of 12.5 (estimated insurance average for peers, much lower than the operating units).
Calculates to $36.38 BB
Gets per share value of $13.43

I copy Jim again. 10 year average has been 1.67% of float. That would be $3 billion in 2025. At 15x that's $20 per b share.

Total share value of $539.03

My total "conservative" IV is $463 with deducting 30% of float value. $488 if don't make that deduction.

Current quote $483. Be careful what you wish for!

Kingswell’s take on the AR…

https://www.kingswell.io/p/a-closer-look-at-greg-a...

ciao

Lately I've been estimating price trajectory by looking at possible future average P/B numbers of 1.37, 1.40, 1.45 and nominal growth rates of 1.5%, 2.0%, and 2.5% per quarter, just to get a cloud of reasonably plausible futures over the next couple of years. As a hobby I've been writing calls whose exit dates/prices are entirely above the cloud, and puts whose entry dates/prices are entirely below the cloud.

Jim, care to part the clouds and spill a few more details about that cloud range of yours?
It looks like prime put-selling weather out there, my own nebula whispers May or June around $440.
TIA

"I think you're better off with KISS. Estimate a reasonable guess of what the "normal" P/B might be now and in the next few years, informed by history...Multiply current book (or peak-to-date book) by that number."

I also track P/B over time, 1965-2025, and I plot it versus time, but I don't think that that is simpler or better than plotting P versus BV. If you want to see how price tracks BV, plot price versus BV.

I do think that using one's best estimate of IV/BV and multiplying BV by that number to get IV is quick, easy and reasonable. However I don't think that the average P/B, which, as you point out, is about 1.4 since 2000, is the best estimate of IV/BV.

"The bigger problem is that the suggested interpretation of the end result, with the fit line suggesting that a "normal" P/B right now would be 1.545 on that trend line, doesn't pass the sniff test given the 15 year average of 1.40. No, I'm sorry, it's utter nonsense and it's just plain wrong."

I respect your opinion, but I guess we'll have to disagree there. I do think that IV is closer to 1.54 than to 1.40 at present, based on the methodology I've described. However there's uncertainty in both of our estimates.

Apparently Buffett agrees with me, as he repurchased stock in Nov 2023 at 1.51x known book. Since Nov 2023 he has repurchased stock at higher P/B, too, as high as 1.60x known book, but those purchases were from long-time owners, and may have been slightly above his estimate of IV.

Apparently Buffett agrees with me, as he repurchased stock in Nov 2023 at 1.51x known book. Since Nov 2023 he has repurchased stock at higher P/B, too, as high as 1.60x known book, but those purchases were from long-time owners, and may have been slightly above his estimate of IV.

I believe there haven't been any buybacks of B shares at a P/B higher than 1.426 times the lowest known book value during the published date range of the purchase.

From what I can read in the runes, I don't know of any A share purchases done above 1.48 times book that were for economic purposes, as opposed to buybacks from shareholders or there estates.
Buybacks in volume tended to start only below 1.45, basically the level today.

The future may differ, of course. As mentioned previously, I think the lack of buybacks in the last price dip was in part because they were waiting for the annual report to be out. So they might be in the market now or soon, if this dip lasts.

Jim

Nonetheless, I still think Greg would be a bit more likely to buy back stock at a given discount than WEB and would maybe repurchase more shares than Warren, especially since he made it pretty clear that dividends are Not coming anytime soon, which I found reassuring.


I agree with you, as long as you replace the word 'think' with the word 'hope'. Abel used exactly the same language that Buffett has used for the last 10 years or so, so we are probably just projecting our wishes onto the new guy in charge, as opposed to having any new evidence that the guy shares our viewpoint. So for repurchases at higher prices than what Buffett has been prepared to pay, I will believe it when I (hopefully) start to see it.

As for making it pretty clear that dividends are not coming soon, Buffett has also been completely consistent in saying this, so it is not really evidence that the balance has shifted in favour of repurchases.

Regards, DTb

A simple minds simple comment re the AR: Perfectly unchanged design and structure => Spelling a simple "Nothing changes. No worries, the culture is deeply engrained. Rest assured we will continue the way we always did".

But in the media huge disappointment with the numbers (and probably the continuance of the stance towards dividends and buybacks), resulting in -5% today (and me having used the opportunity to sell the majority of my puts. Thank you, media :)

Thanks so much for taking the time to compare.

The biggest variance seems to be Grove 1.

We have basically the same earnings but different p/E.

FWIW.

My P/E of 20 ish is derived from a small benchmark exercise.

The 23.1 BB is a combination of 3 industries.

Approximately 25% Rail, 15% Utility, & 60% service.

When I look at the other 3 large rails they have an average P/E of 24.8? Wouldn't we be similar. i understand that a pending merger is influencing one of them, but the lowest of the group is 23. That is what the market thinks they are worth?

Utilities similar story. You may not be willing to pay up, but the market is paying 20+ P/E for Utilities in the current market.

That leaves service. For that I use 17+ for a p/E.

Any ideas on how to improve this?

except in the special case that the value growth rate in future is *coincidentally* the same as the past.

What is the right P/B for a "bond-like stock that grows earnings steadily at 6-9% (*) per year" as I once heard a wise man in this board say?

(*) In real dollars, one assumes but one does not remember now, dince that post was probably 10 years ago. However this wise man / wise guy rightfully disdains calculations in nominal dollars.

My P/E of 20 ish is derived from a small benchmark exercise.

The 23.1 BB is a combination of 3 industries.

Approximately 25% Rail, 15% Utility, & 60% service.

When I look at the other 3 large rails they have an average P/E of 24.8? Wouldn't we be similar. i understand that a pending merger is influencing one of them, but the lowest of the group is 23. That is what the market thinks they are worth?

Utilities similar story. You may not be willing to pay up, but the market is paying 20+ P/E for Utilities in the current market.

That leaves service. For that I use 17+ for a p/E.

I wondered how you got to 20.1. Makes sense now.

For me, the purpose of the exercise is to come up with an IV estimate that would be a decent buy for a long-term hold. I'm not interested in paying 20x, and it's clear that Berkshire isn't, either.

Post New

Macroeconomic Trends and Risks FAQ

---

Contact Shrewd'm
Contact the developer of these message boards.