In last Friday’s Links of Interest, I posted a link to a research document by Ineichen Research and Management. The document contains an article by Joe Taussig entitled “In search of permanent capital”.
Taussig reversed engineered Buffett’s investment record within Berkshire Hathaway and came to some rather intriguing conclusions about the sources of Berkshire’s dramatic outperformance.
“Had the shareholders of Berkshire Hathaway sold all of their holdings in 1969 and reinvested the proceeds in the S&P 500, their $70 million would have compounded at 8.9% for 40 years and be worth $2.1 billion today. But what of Warren Buffett, the “world’s greatest investor”? We have reverse engineered his investment record within Berkshire Hathaway: 12.0% per year. In investment parlance, his investment “alpha” is 3.1% per year. This is pretty good, but does it qualify him for the reputation he has as the “world’s greatest investor”? Had the same investors liquidated their holdings in Berkshire Hathaway and had Buffett the asset manager manage the proceeds in the Buffett Partnership, the $70 million would have grown to $4.4 billion in 40 years. This is a far cry from the $153 billion of market cap that BRK/A enjoys today. This difference between $153 billion and $4.4 billion is what we call “structural alpha”.
In the (re)insurance businesses, the industry standard is that underwriting profits (or losses) equal premiums, minus claims, minus operating expenses. These generate an average underwriting loss of 3% per year (also known as the cost of insurance or “COI”) for each dollar of reserves. The industry generally invests these reserves in long-only fixed income (“because that is how we have always done it”). Assume that the fixed income generates 5% per year. Thus, for every dollar of reserves in a traditional (re)insurer, returns are 2% per year (5% for investments minus 3% COI). In terms of ROEs, the key is the ratio of reserves to equity (leverage), which runs around 5x in the industry. With its equity invested in the fixed income portfolio at 5% plus 5x of reserves earning 2% per increment of reserves, pre-tax ROEs tend to be 15% and after-tax ROEs are roughly 10%.
Under Buffet’s leadership, Berkshire never had a cumulative underwriting profit until 2006 (after which time his cumulative cost of “float” or COI became less than 0.0%). Up until that time, Berkshire’s underwriting losses were still better than the industry norm (his COI was 1% to 2% p.a.). Furthermore, at 2x, his level of leverage was far less than the industry standard of 5x. Investing the equity at 12% and adding 10% for each increment of reserves (investment returns of 12% minus the 2% COI), the total was a pre-tax 32% (12% + 2×10%). Taxes reduced it to an after-tax 20.3%. 20.3% compounding for 40 years turns $70 million into $120 billion. A price to book of 1.29x brings it to $153 billion. Thus the structure generated $149 billion of alpha ($114 billion in better ROEs and $35 billion in a premium to book value).
Stated another way, if Buffett had been run over by a truck 40 years ago but Berkshire had done all of the same things that it did in the meantime, except that it invested in the S&P 500, Berkshire would still be worth $26 billion (versus $2.2 billion in the S&P or $4.4 billion with a manager who could consistently generate returns of 12%). Substitute the HFRI Composite Index (a random selection of hedge funds) for the S&P 500 and the amount is $101 billion. While $153 billion seems like a lot more, Buffett’s share of the difference is far greater than any major hedge fund manager’s performance fees, save Steve Cohen.”
Taussig credits much of Berkshire’s massive outperformance to the way Buffett ingeniously levered the balance sheet 2x. Tausigg appears to be using the debt/equity ratio.
Because Buffett has been explaining it for decades, we’ve understood for a long time that insurance float provides Berkshire with a huge advantage. It functions as quasi free, permanent capital that Buffett can use to invest. What we may not have fully appreciated is the magnitude of the impact.
Taussig does not provide the details of what went into his calculations, which would be fascinating to walk through.
I did take a look at Berkshire’s balance sheet since 1994 using all the data available on line. The average leverage since 1994 is 1.2x. Interestingly, float only accounts for 40% of the leverage during the same period. Although Buffett doesn’t put a spotlight on it, Berkshire definitely benefits from cheap debt financing owing to its fortress balance sheet. (Note that, for simplicity, equity includes minority shareholders’ interests.)
What are some of the takeaways from this?
Trying to get 20% plus returns buying the types of stocks Buffett has typically purchased for Berkshire’s investment portfolio will probably not work. Buffett had the advantage of very low cost, permanent leverage working in his favor. (In fairness, most Buffett watchers feel that he does not invest this way in his personal account.)
The market is focused primarily on Buffett’s skill as an investor and what will happen when his services are no longer available. The market under-appreciates the degree of Berkshire’s structural advantage – access to massive insurance float and low-cost long-term debt, which should be a source of “alpha” for many years to come.
Taussig’s insights give investors a template to use in search of other insurance companies that enjoy the same structural advantages (along with demonstrable investing skill) with the benefit of having a much longer runway.