To The Moon

Summary

• Berkshire Hathaway’s book value growth is losing relevance as a measure of intrinsic value growth.
• The economic value of BNSF and insurance operations significantly exceeds their book value.
• Buybacks at levels above book value have hindered both balance sheet and per-share book value growth.

John Pollock/iStock Editorial via Getty Images

Introduction

Per my February Berkshire Hathaway (NYSE:BRK.A) (NYSE:BRK.B) article, the Berkshire Hathaway Energy ("BHE") subsidiary is not as attractive as it was in the past. Since the time of that article, we now have 1Q24 and 2Q24 numbers, and this is a good time to discuss the erosion of book value growth as a proxy for intrinsic value growth.

Throughout this article, I mean the Berkshire Hathaway shareholders' equity line on the balance sheet when I use terms like book value or balance sheet book value. We're excluding noncontrolling interests, so this is different from the Total shareholders' equity line on the balance sheet.

Berkshire's per-share intrinsic value has been growing faster than their per-share book value, which has been growing faster than their balance sheet book value. The 2Q24 10-Q shows a balance sheet book value of $601,697 million, but the intrinsic value is much higher.

My thesis is that Berkshire's book value keeps losing relevance as a yardstick for their growth.

Erosion History

Berkshire's 2017 annual report was the last one with an opening tied to book value growth in the shareholder letter. This was the way the opening paragraph went for almost three decades:

To the Shareholders of Berkshire Hathaway Inc.:

Berkshire's gain in net worth during 2017 was $65.3 billion, which increased the per-share book value of both our Class A and Class B stock by 23%. Over the last 53 years (that is, since present management took over), per-share book value has grown from $19 to $211,750, a rate of 19.1% compounded annually.*

CEO Warren Buffett addressed this in the shareholder letter in the 2018 annual report, three circumstances which make book value growth less relevant (emphasis added):

First, Berkshire has gradually morphed from a company whose assets are concentrated in marketable stocks into one whose major value resides in operating businesses. Charlie and I expect that reshaping to continue in an irregular manner. Second, while our equity holdings are valued at market prices, accounting rules require our collection of operating companies to be included in book value at an amount far below their current value, a mismark that has grown in recent years. Third, it is likely that - over time - Berkshire will be a significant repurchaser of its shares, transactions that will take place at prices above book value but below our estimate of intrinsic value. The math of such purchases is simple: Each transaction makes per-share intrinsic value go up, while per-share book value goes down. That combination causes the book-value scorecard to become increasingly out of touch with economic reality.

The divergence between book value and intrinsic value has widened over the years as BNSF and insurance operations have become more important. Also widening the gap, buybacks have increased dramatically since the time of the 2018 annual report.

BNSF

The balance sheet in the BNSF 2Q24 10-Q shows a book value of $49,526 million, which we can round to $50 billion for some estimates. Meanwhile, Union Pacific (UNP) has a market cap of $152 billion based on their September 6 share price of $249.63 and the 609,197,628 shares as of July 19 from their 2Q24 10-Q filing. If BNSF's economic value is 80 to 90% of UNP's market cap, and we round to the nearest $5 billion, then their economic value is $120 to $135 billion, which is $70 to $85 billion above their book value.

Insurance

In the Spring of 1996, CEO Buffett said the following in the shareholder meeting for the 1995 fiscal year (emphasis added):

Our insurance business, though, is the most dramatic case of dollar difference between book value and intrinsic value. I mean, the number has gotten very big over time there. I personally think it will tend to get bigger, because I think GEICO will grow, and I think our other businesses will do well.

This dollar difference has grown prodigiously over time. Per the 2023 annual report, insurance float increased from $164,109 million in 2022 to $168,895 million in 2023. A March 2024 article talks about how one can get the 2023 figure from the 10-K. The total is different by $1.1 billion, which is a relatively small percentage:

Berkshire insurance float (Berkshire insurance float (Berkshire Hathaway Fanpage))

The full amount of float is deducted when calculating book value, but this is not so when looking at intrinsic value.

Buybacks

For many years, Berkshire grew book value quickly as they eschewed dividends and buybacks. It is important to make a distinction between balance sheet book value and per-share book value. Suppose we start with a fictitious company which has never issued a dividend and never done any stock buybacks. The company has a balance sheet book value of $100 billion, and it has 100 million shares outstanding, trading at 1.5 times book value. The share price is $1,500 and the per-share book value is $1,000. This fictitious company will make about $6 billion in net income the next year, and their cash flow is around the same amount.

I like to think about 3 oversimplified scenarios where the company doesn't do any cash acquisitions. Assumptions are made but stating them all here would be confusing and pedantic. I'm guessing these hypothetical scenarios aren't perfect in some respects, but they make it easy to see key points. We have conclusions, with all other things being equal, as follows:

1. Balance sheet book value growth is fastest when there are no dividends and no buybacks.

2. Per-share book value growth is fastest when there are no dividends. When buybacks are made at 1.5 times book value, this hurts per-share book value growth, but the pain is less for per-share book value growth than balance sheet book value growth.

In scenario A, the company doesn't do any dividends, and they don't do any buybacks; they add $6 billion to retained earnings. Their balance sheet book value goes up to $106 billion and their per-share book value goes up to $1,060.

In scenario B, the company doesn't do any dividends, and they add $6 billion to retained earnings. However, they do $6 billion worth of buybacks, such that the balance sheet now has a Treasury stock line of $(6) billion. Per the balance sheet, retained earnings went up $6 billion but the Treasury stock line is negative $6 billion such that the balance sheet book value is still $100 billion. The stock was flat during buybacks, such that the $6 billion allowed them to purchase 4 million shares at $1,500 each. We started with 100 million shares, but 4 million were bought back, so there are now 96 million shares. The per-share book value has gone up to $1,041.67. The per-share book value has increased, but it hasn't increased as high as scenario A. In this case the per-share book value growth was nearly 4.2% while it would have been 6% without buybacks.

In scenario C, the company pays the $6 billion out in dividends. Retained earnings don't increase. We don't see any increase in either balance sheet book value or per-share book value.

Per The Motley Fool, the only time Berkshire paid a dividend was in 1967, and it was only for 10 cents per share. As such, Berkshire was much like scenario A until 2011 when the balance sheet shows an initiation of buybacks of $67 million. It was the years leading up to serious buybacks when both balance sheet book value and per-share book value tended to grow the most. The cumulative buybacks appear in the Treasury stock line on the balance sheet, and they were $(79,720) million in 2Q24. We see how cumulative buybacks have grown over the years ($ in millions):

In most of the years above, CEO Buffett's buybacks weren't anywhere close to the 1.5x level of our hypothetical scenario. For years, he talked about 1.1 times and 1.2 times. Obviously, doing buybacks at levels under book value would actually help grow per-share book value faster than if no buybacks were made. This is true even at 1.0 times book value. At some point between 1.0 times and 1.1 times book value, an inflection point occurs and by the time we're at 1.1 times then the buybacks hurt per-share book value growth. The pain is less at 1.1 times and even 1.2 times than it is at the 1.5 times in our example.

The above buybacks substantially hindered balance sheet book value growth. It is my understanding that virtually all the buyback dollars above were at levels higher than 1.1 times book value. As such, they also limited per-share book value growth by varying amounts depending on how high the purchases were above 1.1 times. Looking at the balance sheet in the 2Q24 10-Q, we see $(79,720) million Treasury stock, at cost. Today's balance sheet book value would be $79.7 billion higher were it not for these buybacks and the per-share book value would be higher as well, but this is hard to calculate as we don't know precisely how high the above purchases were above the book value at the time.

Valuation

I like to use a sum of the parts framework when valuing Berkshire.

BNSF $120 Billion

We showed above why this number makes sense if we view BNSF through the same lens as Union Pacific.

BHE $69 Billion

My thoughts on Berkshire Hathaway Energy are still consistent with my February article where I explained why I think Berkshire's portion is about $69 billion and non-controlling interests are about $6 billion for a total of approximately $75 billion.

Other Non-Insurance Subsidiaries $250 Billion

We have trailing twelve months ("TTM") pre-tax income for various segments as follows:

PTC: $915 million or $269 million + $968 million - $322 million

Manufacturing: $11,774 million or $6,043 million + $11,445 million - $5,714 million

McLane: $520 million or $307 million + $455 million - $242 million

Service and retailing: $4,115 million or $1,877 million + $4,721 million - $2,483 million

-----------------------------------------------------

Total: $17,324 million

There are considerations such as acquisition accounting driving down the pre-tax income figures above. Berkshire now owns all of Pilot Flying J ("PTC") but an October Bloomberg article talks about CEO Buffett buying 41% of Pilot Flying J for $8.2 billion in January 2023. Doing a little algebra, this implies all of PTC was worth about $20 billion at the time. This is nearly 22x their TTM pre-tax income figure of $915 million.

In my last article, I valued this group at 12x pre-tax income, but I think this was low given the things being done which keep the GAAP pre-tax income figures low. As such, I now think this group is worth almost 15x pre-tax income, which is about $250 billion.

Investments

We focus on the first five lines of the balance sheet when looking at investments.

Cash, T-bills & Fixed Maturities $288 Billion

This is the sum of the first three lines of the balance sheet.

Stocks $285 Billion

This comes from the fourth line of the balance sheet.

Equity Method Investments $30 Billion

This is the fifth line of the balance sheet, and much of it is Kraft Heinz (KHC) and Occidental (OXY).

Deferred Tax Liability $(38) Billion

We had a net deferred income tax liability table which broke things down in the 2023 10-K, but we don't have this table in the 2Q24 10-Q. That table showed a $56.8 billion Investments, including unrealized appreciation line. In my February article, I estimated Berkshire's economic reality from this to be about 2/3rds of the amount or $38 billion, based on Berkshire's history of being smart with tax planning. Given Berkshire's large sale of Apple shares this year, I'm sure things have changed here, but I'm hesitant to update the number without more details.

Non-Controlling Interests Outside Of BHE $(4) Billion

The 2Q24 balance sheet has a line of $(6,274) million for non-controlling interests, but I think the economic reality is more severe. One reason I think this is because the 2021 annual report said the economic reality of non-controlling interests at the time was about $11 billion, despite a balance sheet figure of $8.7 billion. We already accounted for $6 billion in non-controlling interests when looking at BHE. I'm guessing the grand total on an economic level is about $10 billion, such that there is about $4 billion remaining economically outside of BHE.

Other Considerations $0

I treat Berkshire's insurance operations and their float liability as a wash.

Valuation Table

Here is a spreadsheet summary of my valuation thoughts:

[1] Includes manufacturing, McLane, service and retailing

[2] Includes Kraft Heinz.

[3] Other than BHE.

Of course, I think of valuation more as a range than an exact number. I think Berkshire is worth about $1 trillion +/- 5% at this time.

Each class A share is worth 1,500 B shares. The 2Q24 shows 1,436,696 Class A share equivalents as of June 30, 2024, such that the market cap is about $990 billion based on the September 9 BRK.B share price of $459.61. Berkshire is trading in line with my valuation range, which is unusual as it typically trades at a discount to my estimated valuation range. Overall, I think the stock is mostly a hold, but I've been trimming a little bit as I think we've seen more attractive valuations in the past.

Disclaimer: Any material in this article should not be relied on as a formal investment recommendation. Never buy a stock without doing your own thorough research.