Let’s say my annual car insurance premium with AIG is $2,000. On January 1, my entire payment goes on AIG’s balance sheet as an unearned premium liability, since coverage has not yet applied. (This is analogous to unearned revenue in non-insurance companies.) By July 1 this amount is reduced to $1,000, and by December 31 it is reduced to zero. Unearned premium thus becomes earned premium in a prorated fashion.

Do you think that 100% of my premium can reasonably count as a liability? In other words do you think none of it will end up as profit for AIG upon expiration of the policy? We know from many years of experience that some of the premium will indeed end up as money in AIG’s bank. If the company’s combined loss and expense ratio is 95%, it will keep at least 5 cents of every dollar. (It can earn money on top of that by investing the premium.) Thus the reported book value of a profitable underwriter is always lower than its true value because unearned premium liability is put down in full. Especially profitable insurance companies, with combined ratios of 70% to 80%, can have book values understated by over 5%.

Reported earnings can also be understated if unearned premium liability increases. Let’s say AIG increases unearned premiums from $25 billion to $30 billion. Assuming no change in the combined ratio of 95%, true earnings would be an additional $250 million ($5 billion times 5%). That contributes significantly to intrinsic value. Here’s what Ben Graham said on the subject:

During the years 1945-1948, when the insurance business
expanded greatly, the large increase in unearned premiums
operated to hold down the reported earnings of all companies
and even to produce “statutory” losses for some of them.
This fact, in turn, probably contributed to the unusually
conservative dividend policies followed by most companies.
The two elements together made for a seriously depressed
market level for insurance-company shares in the face of
the tremendous growth of their business and the inherent
stability of the industry.  Thus the need for a proper
understanding and a proper interpretation of the financial
results has been particularly great in the field of fire-
and casualty-insurance-company stocks.

A second factor related to unearned premium liability—neglect of deferred acquisition cost—once created even greater distortions in valuation. In addition to covering my risk of getting into an accident, some of the premium goes to insurance brokers as sales commission. On AIG’s balance sheet this amount is called deferred acquisition cost—an asset resembling prepaid expense in non-insurance companies. DAC is a significant (usually more than 15%) portion of the unearned premium reserve.

When Warren Buffett first studied insurance companies, DAC was not reported on the balance sheet. (Insurance commissioners didn’t care whether certain assets were omitted; for the safety of policyholders they stayed conservative.) Therefore book values were understated by the entire amount of DAC. An investor who recognized this could buy insurance company shares at huge discounts to intrinsic value, thereby making high returns with very low risk. Too bad accountants finally recognized the stupidity of the established system. Since DAC is now reported, reported book values and intrinsic values are much closer.

The accounting treatment of unearned premiums resulted in one major inefficiency that Graham and Buffett were able to exploit. I hope this post has given readers a good idea of how technical their approach really is. Superb investing requires a deep understanding of matters that other people overlook. And people often overlook what is complex.