Berkshire: Part 4 (Boring Port) December 9, 1998

<THE BORING PORTFOLIO>
Berkshire: Part 4
Super-Cat

by Dale Wettlaufer (TMF Ralegh)

ALEXANDRIA, VA (Dec. 9, 1998) -- Before going forward today, I wanted to get the bad news to you. The bad news is that I committed an error in logic the other day. In talking about combined ratio, I went on to conflate that with return on capital. The underwriting loss or gain expressed as a percentage of earned premiums is not return on float or cost of float, as expressed as a percentage, unless the average float was exactly the same as earned premiums. Which would be a coincidence.

The point that I did want to make was that you can treat the after-tax underwriting gain or loss, as a percentage of the average float, as either a return (positive or negative) on float or you can treat it as a cost of the float. Looking then at the after-tax return on investment the float earns in stocks, bonds, or whatever gives you a further return on float measure. To ascertain the spread between return on capital and cost of capital, you can either treat cost of float as zero and add the return on float provided by the underwriting gain to the return on float from investments and look at that at the spread between cost of capital and return on capital. Or, you can treat the underwriting gain or loss as a cost of float (in the case of an underwriting gain, the cost of float will be negative) and compare that with investment returns attributable to the float. Either way, for the purposes of Economic Value Added analysis, the spread between cost of capital and return on capital will be the same.

Say, for instance, a company generates an after-tax underwriting gain of $100 million and has an average float of $5 billion. The cost of that float will be a negative 2%. And say the company generates an after-tax investment return on the float (including unrealized capital gains) of 15% over the course of the year. If you treat the cost of capital as -2% and the return on capital of 15%, then the difference between return on capital and cost of capital will be 17 percentage points.

If you treat the cost of capital as zero and treat the underwriting gain and the investment return as a 17% return on capital, then the spread between return on capital and cost of capital will be 17 percentage points.

Sorry for the confusion, if there was any.

Continuing with Berkshire

OK, on with other things. We've talked about GEICO Direct and other lines of specialty insurance at Berkshire. Let's talk about super-cat underwriting. Super-cat is super-catastrophe insurance, which primary insurers such as Allstate (NYSE: ALL) or reinsurers such as PartnerRe (NYSE: PRE) take out to cover themselves against a very large catastrophic loss event such as a hurricane or an earthquake. For instance, Berkshire underwrote the California Earthquake Authority when other financial interests couldn't give the CEA an answer quickly enough. In this instance, Berkshire received very large premiums (which are now part of the float) in exchange for a promise to cover the CEA above and beyond a certain layer of losses it will cover. What Berkshire super-cat super-genius Ajit Jain saw in this opportunity was that it would take a rare, very large earthquake for the losses to reach the layer of coverage for which Berkshire is responsible. The large premium Berkshire received for the coverage, if invested wisely, would more than offset the company's exposure over the life of the contract.

The maximum exposure Berkshire faces for this contract is about $1 billion pre-tax and $600 million after tax. Here's part of the story on that from Berkshire's 1996 Chairman's Letter to Shareholders:

"A few facts about our exposure to California earthquakes - our largest risk - seem in order. The Northridge quake of 1994 laid homeowners' losses on insurers that greatly exceeded what computer models had told them to expect. Yet the intensity of that quake was mild compared to the "worst-case" possibility for California. Understandably, insurers became - ahem - shaken and started contemplating a retreat from writing earthquake coverage into their homeowners' policies.

"In a thoughtful response, Chuck Quackenbush, California's insurance commissioner, designed a new residential earthquake policy to be written by a state-sponsored insurer, The California Earthquake Authority. This entity, which went into operation on December 1, 1996, needed large layers of reinsurance -- and that's where we came in. Berkshire's layer of approximately $1 billion will be called upon if the Authority's aggregate losses in the period ending March 31, 2001 exceed about $5 billion. (The press originally reported larger figures, but these would have applied only if all California insurers had entered into the arrangement; instead only 72% signed up.)

"So what are the true odds of our having to make a payout during the policy's term? We don't know -- nor do we think computer models will help us, since we believe the precision they project is a chimera. In fact, such models can lull decision-makers into a false sense of security and thereby increase their chances of making a really huge mistake. We've already seen such debacles in both insurance and investments. Witness "portfolio insurance," whose destructive effects in the 1987 market crash led one wag to observe that it was the computers that should have been jumping out of windows.

"Even if perfection in assessing risks is unattainable, insurers can underwrite sensibly. After all, you need not know a man's precise age to know that he is old enough to vote nor know his exact weight to recognize his need to diet. In insurance, it is essential to remember that virtually all surprises are unpleasant, and with that in mind we try to price our super-cat exposures so that about 90% of total premiums end up being eventually paid out in losses and expenses. Over time, we will find out how smart our pricing has been, but that will not be quickly. The super-cat business is just like the investment business in that it often takes a long time to find out whether you knew what you were doing."

As someone that doesn't know that much about super-cat insurance, I cannot focus on individual contracts and the tiniest details of what Berkshire does here. What we focus on is how the company has conducted itself in the past with respect to risk and reward and how that might play out in the future. We do know this: shareholders are not exposed to one or a few bad decisions with Berkshire's insurance operations. Right now, the CEA exposure represents less than 2% of the Berkshire's tangible asset value and less than 4% of tangible book value as of September 31.

We also know that people at Berkshire think rationally about pricing, which is the component of the reward part of the risk/reward calculus: "The influx of 'investor' money into catastrophe bonds -- which may well live up to their name -- has caused super-cat prices to deteriorate materially. Therefore, we will write less business in 1998." As investors, we're not rooting for the company to show revenue increases or earnings increases year after year. We want people who know how to price risk appropriately, especially in the insurance and banking companies that will eventually be part of our portfolio.

Last year, super-cat underwriting provided $182.7 million in net earnings for Berkshire and $107.4 million the year before. As Chairman Warren Buffett warned, though, the industry has been very lucky with super-catastrophic events. Combined with great conditions in the bond market and a dearth of big super-cat losses, pricing has thus suffered. To learn about Berkshire's response to favorable pricing catalysts, I recommend the "insurance operations" section of the 1989 letter to shareholders.

Alex will be here Friday to talk some more about Borders. Hope to see you on the boards.

• Discuss Boring Investing on the Boring Port message board.
• 10/01/98: The New Boring Port Transitions Facts

Share Your Two Cents. Join the Fool Charity Drive.

Stock  Change    Bid
ANDW  +  7/32  17.91
BGP   -  7/8   23.75
CSL   +  3/16  47.75
CSCO  +  15/16 80.75
FCH   -  7/16  23.44
PNR   -  11/16 36.06
TBY   -  1/16  7.19

                   Day   Month    Year  History
        BORING   -0.38%   2.64%   0.52%  26.48%
        S&P:     +0.18%   1.71%  21.96%  90.39%
        NASDAQ:  +0.77%   5.17%  30.57%  96.97%

    Rec'd   #  Security     In At       Now    Change
  6/26/96  225 Cisco Syst    23.96     80.75   237.08%
  2/28/96  400 Borders Gr    11.26     23.75   110.99%
  8/13/96  200 Carlisle C    26.32     47.75    81.39%
  4/14/98  100 Pentair       43.74     36.06   -17.56%
  5/20/98  400 TCBY Enter    10.05      7.19   -28.45%
  1/21/98  200 Andrew Cor    26.09     17.91   -31.37%
  11/6/97  200 FelCor Sui    37.59     23.44   -37.65%

    Rec'd   #  Security     In At     Value    Change
  6/26/96  225 Cisco Syst  5389.99  18168.75 $12778.76
  2/28/96  400 Borders Gr  4502.49   9500.00  $4997.51
  8/13/96  200 Carlisle C  5264.99   9550.00  $4285.01
  4/14/98  100 Pentair     4374.25   3606.25  -$768.00
  5/20/98  400 TCBY Enter  4018.00   2875.00 -$1143.00
  1/21/98  200 Andrew Cor  5218.00   3581.25 -$1636.75
  11/6/97  200 FelCor Sui  7518.00   4687.50 -$2830.50

                             CASH  $11273.22
                            TOTAL  $63241.97