Three simplistic valuation models give IV as of 3/31/22 of about $520K. The first model is simply IV = 1.5*BV, a multiple that has worked reasonably well for some years now. This model gives IV of $518K. A second model is IV = highest repurchase price/0.95, or a 5% discount to IV, = $521K (estimating the hight price paid to be $495K. A third model is the trendline of a 30+ year regression (log-log) of price/share vs BV/share, what I call the weighing machine model. “In the long run the market is a weighing machine.” The regression fit is extremely good. The current trendline value is about $513K. FWIW.
9 Likes
“The regression fit is extremely good.”
Deviations of the price from the trendline are generally small and short lived, although in June 1998 P/B reached 2.8 and in Mar 2000 and Mar 2009 P/B hit 1.1, which represented significant deviations from the trendline. Since 1965 the trendline value has increased from 0.8BV in 1965 to 1.5BV today.
Thanks. FYI, attended a Tilson dinner in Omaha last night after the meeting. He included his latest BRK Q1 2022 update on his IV which is now 542K/share. He bases it on 335K cash and investments/ share and 18.8B pretax earnings excluding all investment income. I believe he chose a 16 pretax multiple. He said he would release his more detailed update later this week. I believe he predicted IV would be 600K by year end.
2 Likes
Tilson seems to have become a promoter. Ads with “the legend who called…”
I have also found his valuations to be a bit lazy.
3 Likes
“I have also found his valuations to be a bit lazy.”
I don’t fault the two-column method. It’s logical for an asset rich company. He should, however, compare it to other methods and to the same method with different assumptions. Maybe he’s done that and just not presented the comparisons. Over the years he’s used multiples of pre-tax earnings ranging from 10 to 16. Hopefully he’s not choosing a multiple to get the answer he wants, but rather has some justification, such as changes in tax rate. I’ll give him the benefit of the doubt, but I much prefer to read mungofitch’s analyses that discuss the assumptions and give comparisons to alternative valuation models.
There are many, reasonable, valuation models for Berkshire, including discounted cash flow, sum-of-the-parts, look-through earnings, etc. We don’t want to fall into the trap of portraying one model with one set of assumptions within the model as definitive.
1 Like
“A second model is IV = highest repurchase price/0.95, or a 5% discount to IV, = $521K.”
I think the fact that Buffett did not repurchase shares in April gives us some idea of his IV estimate. While other factors besides price affect Buffett’s decision to repurchase shares, such as whether he is spending money to buy other companies or other stocks, price is certainly a major factor.
If one looks at the price chart for Berkshire, it appears that Buffett bought all of his shares for March in the first couple of weeks. His average price paid was $485K, and the low price available in March was on Mar 1 at $476K, after with the price rose to $539K. Buying from Mar 1 at $476K to Mar 14 at $494K would give about the right average of $485K. It appears that Buffett decided not to buy shares above about $494K.
Perhaps it’s just coincidence that Buffett stopped repurchasing shares at 1.43x BV, but if that’s his stated limit of a 5% discount to IV, then his estimate of IV is about 1.5x BV. Buffett’s future repurchases may support or refute this conclusion.
5 Likes
2 column method by Whitney Tilson in his daily email:
“I have used a consistent method to estimate Berkshire’s intrinsic value for the past two decades, which I believe is similar to the one Buffett uses: take the cash and investments per share and add the value of the operating businesses.
At the end of the first quarter, cash and investments were $363,000 per A-share, but Berkshire’s stock portfolio declined in April by $28,000 per share, so that’s $335,000 today.
Berkshire’s pretax operating earnings over the past 12 months through the first quarter, including an estimate for normalized insurance earnings, were $18.8 billion per share.
To this, I apply a conservative below-market multiple of 11 times to arrive at a value of $207,000 per share.
Thus, my estimate of Berkshire’s intrinsic value is $335,000 plus $207,000, for a total of $542,000.
The A-shares closed Friday at $484,340, meaning that I believe the stock is trading at an 11% discount to its intrinsic value.”
It baffles me how a failed fund manager is qualified to dish out stock research / recommendations.
https://www.bloombergquint.com/markets/his-hedge-fund-shut-w…
I guess it’s just good business for promoters. Tilson’s networth was reported to be around 50 mil.
p.s. I read he just changed his “research business model” again but I had no interest to read on.
2 Likes
Always people who want to be told the future. And have their hand held.
11x pre-tax operating earnings. That’s better.
At the end of the first quarter, cash and investments were $363,000 per A-share, but Berkshire’s stock portfolio declined in April by $28,000 per share, so that’s $335,000 today.
Berkshire’s pretax operating earnings over the past 12 months through the first quarter, including an estimate for normalized insurance earnings, were $18.8 billion per share.
To this, I apply a conservative below-market multiple of 11 times to arrive at a value of $207,000 per share.
Thus, my estimate of Berkshire’s intrinsic value is $335,000 plus $207,000, for a total of $542,000.
The A-shares closed Friday at $484,340, meaning that I believe the stock is trading at an 11% discount to its intrinsic value.”
Here’s how I prefer to analyze the same data:
-
Investments today are trading at $335k per A-share. You have to decide for yourself whether those shares are overpriced or underpriced, but 2 things should be kept in mind: (i) the Apple stake is about half the investment, and they are at about 25 times earnings, for a business that is growing very slowly. And (ii) those shares come with a baked-in additional debt, which is the tax that will someday have to be paid on the capital gains, which are substantial. That matters less if none of the big positions are ever sold, but it’s not nothing.
-
Berkshire’s pretax operating earnings over the past 12 months through the first quarter, including an estimate for normalized insurance earnings, were $18.8 thousand per share. (Tilson said ‘$18.8 billion’; ignore this error!) At Friday (April 29th)'s closing price of $484,340 per A-share, that means, the operating businesses are valued at $484k - $335k = $149k per A-share. That means that if you are happy with the market value of the investments, you can buy the operating businesses for $149k/$18.8k = 7.9, so the non-investment side of Berkshire is trading for about 8 times pre-tax earnings (or 10x net income, if income tax rates are about 20%).
Tilson’s approach is similar, but the 11x multiple he slaps on pre-tax operating earnings is completely arbitrary, as should be obvious from the fact that he keeps changing that multiple; sometimes it is 10, sometimes 11, sometimes 12. Better, I think, is to just divide the earnings from non-investment operations by the value of the operating side of the business, and let the reader judge whether she wants to pay 10 times earnings or not, for this collection of businesses.
Slightly better still would be to deduct the cost of the tax owed on unrealized capital gains. That also would involve assumptions about when it is paid, if ever, and it doesn’t change things enormously. But as an example, the Apple position was $159b at quarter end, and had a cost basis of $31b (!), so there’s a future tax bill of about $25b right there. The total unrealized gains in the investment portfolio were $245b on March 31st (a little less now), so the rest of the investment portfolio has a similar level of tax that may have to be paid someday.
I don’t know how to best handicap this roughly $50b in tax on future sales of these positions, but if we took a first stab and said that maybe that $50b in future tax is far enough in the future to only constitute a $25b present liability, that would mean that the value of the investment side of the portfolio is worth $25b/1.471m A-shares = $17k less per A-share, or $318k. Subtracting that from last Friday’s $484k per share now gives us $166k per A-share for the operating side of the business, and dividing that by the same 18.8k in pre-tax earnings per A-share now gives us an 8.8 multiple on pre-tax earnings, or a 11x multiple on after-tax earnings. Still good, and not very different from the 10x multiple we got when we ignored future taxes on capital gains.
Regards, DTB
15 Likes
nice work DTB. We tend to gloss over the capital gain liability.
On Apple, we bought more. Not a ton ($560m) but he doesn’t relax his standards for acquiring pieces of a business.
We tend to gloss over the capital gain liability.
On Apple, we bought more. Not a ton ($560m) but he doesn’t relax his standards for acquiring pieces of a business.
Yes, and the two points may be related: Buffett may be prepared to pay such a high price for Apple because he feels that, although the multiple is quite high, he has a very high degree of confidence in the sustainability of the earnings, so he can just hold this position pretty much forever. That, and perhaps also confidence that Apple will not squander those earnings, given their history in the last 10 years of making practically no acquisitions and paying out the majority of their earnings in dividends and share repurchases.
My iMac that I have owned for the last 10-12 years finally expired last week, and I hesitated about 20 seconds before buying another that looks almost identical. Of course, the processor is far faster, the screen quality is no doubt much better, and I will be able to update the operating system again. But don’t ask me what the numbers are or how the thing looks: I don’t know what the specs are, and I trust Apple to give me complete confidence when they mail it to me. I wouldn’t have bought any other computer even at half the price. In fact, for something that I use for hours a day, cost is almost not a consideration at all. I probably would have paid twice as much if I really had to.
This may seem like a silly waste of money to some people, but I am just relating this because I think there are millions of people just like me who are insensitive to price increases for these products, within some limit of course. When Buffett was asked what kind of company you would want to own if we get a lot of inflation, he gave no concrete advice, but I wonder if he had one in mind.
dtb
5 Likes
In fact, for something that I use for hours a day, cost is almost not a consideration at all. I probably would have paid twice as much if I really had to.
Along similar reasoning, I admit to considering purchasing a Mac Pro. The Cheese Grater. https://i.pcmag.com/imagery/articles/06s0SjxC2lXUoTLGAaJOFwE…
Not because I really need it, but because I’m a computer geek and I can afford it.
I’ve never had a Mac PC on my desk as it happens, but it’s no big leap, I’m an old Unix guy.
The odd thing is that their flagship computing device no longer has their chosen CPU architecture.
Jim
4 Likes
The odd thing is that their flagship computing device no longer has their chosen CPU architecture.
Not “no longer”, but “not yet”. This supposedly changes this year so maybe you should wait a little to not eventually regret your buy.
2 Likes
Thanks. FYI, attended a Tilson dinner in Omaha last night after the meeting. He included his latest BRK Q1 2022 update on his IV which is now 542K/share. He bases it on 335K cash and investments/ share and 18.8B pretax earnings excluding all investment income. I believe he chose a 16 pretax multiple. He said he would release his more detailed update later this week. I believe he predicted IV would be 600K by year end.
Admittedly I have never looked at his analysis in detail, but is not not skipping over a big chunk of liabilities here? Is that perhaps how his valuations tend to be a bit too rosy? I know DTB raised the point of deferred taxes but I suspect there are more missed liabilities than that.
StevnFool