Deconstructing Buffett

By Felix Salmon
February 28, 2010
BusinessWeek cover story on Warren Buffett, she explains the hidden upside:

" data-share-img="" data-share="twitter,facebook,linkedin,reddit,google" data-share-count="true">

Alice Schroeder has two interesting takes on Berkshire Hathaway’s acquisition of Burlington Northern. In her BusinessWeek cover story on Warren Buffett, she explains the hidden upside:

Buffett always likes a sweetener, and Burlington gives him one in the form of information. He learns about wallboard demand from USG and consumer-credit trends from American Express, but Rose has called the railroad a kaleidoscope of the economy. Rail traffic patterns are a window on commodity, wholesale, consumer, and international trade flows. Buffett is adding this kaleidoscope to what his other CEOs tell him about the “reset of the consumer” to a lower level of spending. They feed him data from Berkshire’s portfolio of companies—sales of building materials, jewelry, furniture, real estate, credit, fractional jets, vacuum cleaners, fabricated steel, newspaper ad lineage, and other products and services. He may now command as much information about the state of the U.S. economy as anyone, including the Federal Reserve—and probably gets his faster.

This makes sense — and it might also provide a hint as to whom Buffett will choose to succeed him as chief investment officer. He’ll want a data hound, someone who can leverage the huge amounts of information that the Berkshire conglomerate provides.

On her blog, Schroeder then adds that she’s disappointed in the way that Buffett is going to hide BNSF’s earnings in his annual reports:

Berkshire has been growing less transparent year by year. Now it is going to combine BNI with its utlity segment for financial reporting next year. Buffett made the argument for combining them (regulated, high capex etc.), but the result is less transparency. BNI is in the transportation business. If being a regulated and capital intensive business is what creates an operating segment for financial reporting, the insurance businesses would also be combined with Mid-American.

Schroeder gives good accounting reasons why BNSF should not be lumped in with the utilities, adding, wearing her CPA hat, that “this is not the spirit of the rules” and that “this is pretty annoying”.

Meanwhile, in his annual letter to shareholders, Buffett does a pretty weak job of responding to Jeff Matthews’s criticism of his decision to pay for BNSF in large part using BRK’s own stock. Or, to put it another way, he actually spends a lot of time and space laying out Matthews’s criticism in his trademark ultra-lucid manner. And then he tries to answer his own criticisms thusly:

In the end, Charlie and I decided that the disadvantage of paying 30% of the price through stock was offset by the opportunity the acquisition gave us to deploy $22 billion of cash in a business we understood and liked for the long term. It has the additional virtue of being run by Matt Rose, whom we trust and admire. We also like the prospect of investing additional billions over the years at reasonable rates of return. But the final decision was a close one. If we had needed to use more stock to make the acquisition, it would in fact have made no sense. We would have then been giving up more than we were getting.

This can essentially be boiled down to two words: “trust me”. Which is maybe sensible, given that trusting in Buffett is exactly what he’s been asking his shareholders to do for decades. But the letter does make it seem that it’s a lot easier to argue against the BNSF acquisition than it is to argue for it.

For me, however, the weirdest part of the letter is where he talks about Clayton Homes, the manufacturer of modular and mobile homes which famously was giving away a free can of pork and beans with every house bought. Clayton doesn’t live in the category of “Manufacturing, Service and Retailing Operations”, where you might suspect it would be found; instead, it’s part of “Finance and Financial Products”, since the real business here isn’t selling homes so much as it’s selling the financing to buy them.

As Buffett explains:

Currently buyers of conventional site-built homes who qualify for these guarantees can obtain a 30-year loan at about 5 1⁄4%. In addition, these are mortgages that have recently been purchased in massive amounts by the Federal Reserve, an action that also helped to keep rates at bargain-basement levels.

In contrast, very few factory-built homes qualify for agency-insured mortgages. Therefore, a meritorious buyer of a factory-built home must pay about 9% on his loan.

Wow, sounds like a great business! Rather than see its financing profits competed away in a commoditized mortgage market, Clayton essentially has a monopoly on providing financing on its homes, and can lend out money at rates much higher than prevailing mortgage rates.

But weirdly, Buffett seems to be unhappy about this:

Berkshire can’t borrow at a rate approaching that available to government agencies. This handicap will limit sales, hurting both Clayton and a multitude of worthy families who long for a low-cost home.

Really, Warren? I’d like to see some numbers on this, because I always thought that the great thing about being Berkshire Hathaway, even without a triple-A credit rating, was that you could borrow at a rate approaching that available to government agencies. What’s the spread on Berkshire debt over agency debt? When Berkshire recently borrowed $8 billion, it paid between 2bp and 43bp over Libor on the floating-rate bonds, and between 63bp and 93bp over Treasuries on the fixed-rate bonds. In comparison, agency debt recently narrowed all the way to 66bp over Treasuries, albeit at longer maturities.

In any case, I’d say that the funding advantage that agencies have over Berkshire will be no more than about 50bp, while the financing rates that Clayton’s buyers are paying to Berkshire are, by Buffett’s estimation, about 375bp more than those offered by Frannie. And remember here that Buffett is adamant that Clayton’s buyers are good credit risks, and that “Clayton’s delinquencies and defaults remain reasonable and will not cause us significant problems”.

If you take his words at face value, then, Buffett is saying that Clayton’s buyers are no less creditworthy than buyers of site-built homes. So something over 300bp of the spread that Clayton is charging over other mortgage providers should be pure gravy for Clayton and Berkshire. I’m sure that Buffett would love Clayton to be able to sell more homes than it’s doing right now, but if you remember that Clayton is basically in the business of finance more than it is in the business of manufacturing, I can’t quite see what Buffett is complaining about. Sure, he could sell more mobile homes if you could get conventional mortgages on them. But then he’d lose all of his financing profits, and I’m sure he wouldn’t want that.


We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see

It’s easier to trust Buffett when he’s coherently arguing from the same principles as he always has than when he seems to be trying (badly) to justify a position instead, as he did as quoted in the Matthews post earlier.

Posted by dWj | Report as abusive

I think this is one of those cases where Warren Buffet genuinely don’t see, or simply don’t want to see, what he knows perfectly to be true. Which is, his manufacturing business is making money from financing rather than from actually making the products, and on some intellectual level he’s not comfortable with that as he has been a fierce critic of the Wall Street economic model.

Which is all fine and dandy, except that the Wall Street economic model isn’t some nefarious scheme dreamed up in some corporate attic. The financialized economy is the reality wrought by the changes in economic structure.

Posted by SGKingsley | Report as abusive

By the way, 300 basis points. Jesus Christ. That’s an amazing spread. I would kill for that.

Just unbelievable. I would kill for that spread.

Posted by SGKingsley | Report as abusive

Well, I’m not sure exactly what Mr. Buffett means by “meritorious”, but manufactured homes have a significantly poorer loan repayment record than site-built homes, on average.

So maybe Clayton buyers are “good credit risks” relative to other buyers of manufactured homes and, sadly, relative to many buyers of site-built homes during the boom, but they probably aren’t so good compared to the traditional, long-term average for buyers of site-built homes.

Posted by karen2 | Report as abusive

Felix, for what it is worth, if Berky wanted to issue debt today, they would have to issue at around 0.75% +/- 0.15% over agency yields. More around 5 years, less around 30.

While I’m here, here are 2 curiosities — Bloomberg’s DLIS function doesn’t work with Berky, which gives a list of maturities, probably because of all the nonguaranteed debt, and EETCs [enhanced equipment trust certificates] from BNSF.

But, doing BRK , a list is easily available. Sorting it by size of issue outstanding, what is fascinating is that most of the holding company debt has a short tenor. My estimate is an average maturity of 4.4 and an effective duration of 2.8. 90% of it comes due by 2015.

Now, Berky doesn’t have that much debt at the holding company level, but it is remarkable that they are financing so much short. It is a negative arb, because he has a little more cash on hand than holding company debt.

Posted by DavidMerkel | Report as abusive

Sorry, to finish — some of what I wrote got lost in the HTML — I downloaded it all to Excel and ran my own calculations.

The non-guaranteed debts of Berky are much longer — an effective duration of around 8.4 — three times as long as the holding company debt.

I’m not sure if any of this is significant, but it makes me wonder what Buffett’s philosophy is with holding company debt.

Posted by DavidMerkel | Report as abusive

TxLIHIS’s ( response to Buffett’s discussion of the “punitive differential” on MH:

On Friday Warren Buffett released his annual letter to shareholders of Berkshire Hathaway. Berkshire Hathaway owns Clayton Homes, a major manufactured home manufacturer. In the letter, Mr. Buffett laments that “very few factory-built homes qualify for agency-insured mortgages”

Mr. Buffett compares the market rates of 9% on manufactured housing loans to the current rate of 5.25% and states “If qualifications [on conforming loans] aren’t broadened, so as to open low-cost financing to all who meet down-payment and income standards, the manufactured-home industry seems destined to struggle and dwindle.” In doing so, he implies that the 375 basis points (bp) (3.75%) differential between manufactured home loans and conforming loans is due to market intervention by the “all powerful” FHA, Fannie Mae, and Freddie Mac.

Mr. Buffett is being disingenuous. The current market rate on “non-conforming” site built housing is about 80 bp (0.8%) higher than conforming loans. This 80 bp is the interest rate subsidy provided to loans purchasable by Fannie Mae. The other 295 bp is the market discount for manufactured housing.

Why does the market discount manufactured housing?

Research by Consumers Union, the publisher of Consumer Reports, shows that prices of manufactured homes over time have more volatility than conventional homes. (Full disclosure, I was the leader researcher on Consumers Union’s Manufactured Housing Research Project and author of the linked report.) This means that a greater percentage of manufactured home buyers are under water at some point in the life of a loan. Homes worth less than the loan balance have a higher risk to the lender. Ergo, they demand a higher interest rate.

But Mr. Buffett must know this, because he also owns Vanderbilt Mortgage, a manufactured home lender. And with $30.5 Billion in cash on hand, if he wanted to make cheaper manufactured home loans, he would do it. If the subsidy was the only difference, he could be making loans at 6.05% on all manufactured homes and profiting tidily.

But he’s not.

We should note that manufactured homes can be eligible for conforming loans. They generally have to have a permanent foundation and meet other guidelines. Even more subsidy is available through the USDA 502 loan program, which accepts new manufactured homes on permanent foundations, and offers highly subsidized interest rates for very-low, low, and moderate income borrowers. However, this program is almost never used: in 2008, USDA made or guaranteed just 9 loans in Texas.

The reality is most dealers and consumers don’t structure their purchase transactions to meet the lending requirements of these programs. For example, they may place the home on a short-term foundation and lease in a park rather than on a permanent foundation on owned land. Such homes are generally not eligible for subsidy.

Whether or not the government should provide an 80 bp subsidy to all manufactured homes would need to be the subject of a much longer blog post. (The short version: the answer depends on whether the purpose of the lending subsidy is to put roofs over peoples’ heads or to promote access to the asset building opportunity of homeownership. (see this report)) But what Mr. Buffett says the industry needs to stay alive is not an *equal* 80 bp subsidy, but a much larger 375 bp subsidy.

In short, he wants a 468% higher loan subsidy for manufactured housing than conventional housing. And that’s not leveling the playing field, that’s tilting the table.

(Version with links available at: tt-wants-to-tilt-the-market-towards-manu factured-housing/)

Posted by jeweke | Report as abusive

The reason that conforming loans are so low versus manufactured housing, is that they are lending against the land, which is valuable, and doesn’t usually depreciate, versus the manufactured housing, which does depreciate rapidly. Also, the conventional housing borrowers are higher quality, and thus, the loss statistics on MH are a lot higher than that for conventional lending.

Posted by DavidMerkel | Report as abusive