Explaining Berkshire's Losses

In last week's column, I analyzed Berkshire Hathaway's (NYSE: BRK.A) dramatically improving cash flow and noted that operating profits showed improvement as well. Today, let's take a closer look at the results for GEICO and General Re, Berkshire's largest subsidiaries by far, accounting in the second quarter for 21% and 28%, respectively, of Berkshire's revenues. Both of these subsidiaries have incurred steep losses recently, which have weighed heavily on Berkshire's stock.

GEICO
GEICO's losses are due primarily to the company investing heavily in growth. As Warren Buffett wrote in Berkshire's latest annual report, "I think these expenditures are the best investment Berkshire can make... We would happily commit $1 billion annually to marketing [GEICO] if we knew we could handle the business smoothly and if we expected the last dollar spent to produce new business at an attractive cost."

GEICO's investments are paying off. Over three years through 1999, its national market share in personal auto insurance increased from 2.7% to 4.1%. Policies-in-force grew 21.5% in 1999 and 20.5% and 18.0% in Q1 and Q2 this year, respectively.

Despite ongoing heavy investments in the business, GEICO's underwriting losses declined from $86 million in Q1 to $65 million in Q2 this year, as both losses and underwriting expenses as a percentage of premiums declined (from 86.5% and 20.1% in Q1, respectively, to 86.2% and 18.5% in Q2). I view this a good news, but be aware that an analyst at CS First Boston argued that losses would have been $108 million had it not been for seasonality and lower profit-sharing payments.

For its part, Berkshire has become more bullish recently on GEICO's future prospects, based on a careful comparison of the Q1 and Q2 reports. (Reading Berkshire's reports is like reading official government newspapers in China -- you have to read between the lines and carefully compare current writings with past ones to really understand what's going on.)

Q1's news release stated: "During the past 12 months, GEICO voluntary auto policies in-force increased 20.5%. Growth in policies in-force is expected to continue to be strong during the remainder of the year. However, it is anticipated that the rate of growth will decline modestly."

Q2's news release stated, "Underwriting results also improved from the preceding quarter at GEICO. Unit growth slowed with policies-in-force at June 30, 2000 up 18.0% from last year. This slowing is expected to continue, though premium growth should continue strong due to higher prices."

Note that both predict a slowing of growth in policies-in-force, but the more-recent release predicts that "premium growth should continue strong due to higher prices."

Equally importantly, the following statement, which appeared in the Q1 Interim Report, did not appear in the Q2 report:

"[GEICO's]ongoing underwriting losses are anticipated throughout 2000."

These two factors lead me to conclude that Berkshire now believes that GEICO could begin generating operating profits again by the end of this year due to better prices driving robust, higher-margin premium growth.

General Re
Unlike GEICO, it's hard to put a positive spin on Gen Re's losses, which have been large and persistent -- due to inadequate premium rates and unexpectedly large claims -- ever since Berkshire acquired the company at the end of 1998. In 1999, Gen Re's underwriting loss was $1.2 billion, plus another $504 million in the first half of this year.

But, the trend may be improving as Gen Re's loss in Q2 of this year declined to $231 million, compared to a $273 million loss in Q1. Losses as a percentage of premiums were unchanged, but underwriting expenses declined (losses were 86.4% of premiums in both periods and underwriting expenses fell from 29.8% to 26.3%).

That's not much to write home about but, as with GEICO, positive statements about the future are buried in the text of Berkshire's Interim Reports. Both the Q1 and Q2 Reports state that Gen Re's underwriting results are "unsatisfactory" and will remain so during the remainder of 2000 despite increased pricing, since this takes a while to be reflected in the financial results. However, only the Q2 report adds that "there are signs of improvement in certain markets" and that "absent a mega-catastrophe, Berkshire expects that General Re's underwriting results will be improved during the second half of 2000 as compared to the first half of 2000."

Good news related to Berkshire's float
Berkshire's Q2 news release noted that float -- the enormously valuable cash from insurance premiums that Berkshire can invest before claims must be paid -- grew by $408 million in Q2, to $25.8 billion. This compares to a tiny $92 million gain in Q1. More importantly, the company stated that, "It is anticipated that float will grow at a significantly greater rate during the remainder of 2000 and that, absent a mega-catastrophe, the cost is likely to decline further."

The news release also noted that "National Indemnity's reinsurance underwriting loss for the quarter includes about $50 million which stems from business that will generate significant amounts of 'float' for many years." This means that $50 million of Berkshire's underwriting loss in Q2 was not, in fact, a genuine loss. Yes, Berkshire wrote certain reinsurance policies on which it expects to lose money -- and must book these non-cash losses up front -- but, according to the Interim Report, due to "time-value-of-money concepts... this business is accepted because of the large amounts of policyholder float generated for investment."

If that doesn't make any sense, consider this analogy: would you pay me $1.10 in 10 years if I gave you $1.00 today? While at first glance, you might appear to be losing 10 cents, you would be wise to make this trade since you could invest the dollar in a 100% secure 10-year Treasury note paying 5.7% and earn 74 cents in interest, earning yourself a guaranteed 64 cents in profit.

Similarly, Berkshire sometimes writes insurance wherein it agrees to pay more in the future than it is receiving in premiums, so it must book this loss now, but it can't report the expected gains from the float at this time because they are uncertain (though the premiums appear in the cash flow statement). Writing policies like this hurts current reported income, but almost certainly increases Berkshire's intrinsic value.

Independent of Wesco, Berkshire has not sold a material amount of its Freddie Mac stock
In last week's column, I speculated that Berkshire had been selling its Freddie Mac stock in the first half of this year, above and beyond what Wesco sold (Wesco is 80.1% owned by Berkshire, so its stock holdings are consolidated with Berkshire's). Well, so much for my theory. On the day my column appeared, the company filed a Form 13F-HR/A with the SEC that showed that while Berkshire (Lou Simpson at GEICO to be exact) had sold some Freddie Mac stock in the first half of the year, it was only 45,100 shares, equal to a mere 0.15% of its holdings at the beginning of the year.

Conclusion
As I have argued in previous columns (see my home page for links), Berkshire Hathaway is my largest holding because I believe it is an exceptional company that is widely misunderstood and significantly undervalued. Yet, the stock remains substantially below the high above $80,000 it reached in early 1999 largely because of the big losses at GEICO and especially Gen Re. With one quarter of improving results at these subsidiaries -- and bullish statements about their future (from a company known for underpromising and overdelivering) -- plus continued big increases in cash flow, I believe Berkshire Hathaway's stock could break out of its doldrums.

Share your opinion of the quarter's results on the Berkshire Hathaway and the Boring Stocks discussion boards.

-- Whitney Tilson

Whitney Tilson is Managing Partner of Tilson Capital Partners, LLC, a New York City-based money management firm. Mr. Tilson appreciates your feedback at [email protected]. To read his previous guest columns in the Boring Port and other writings, click here.

Related Links: