Actuarial Review Return to Main Page

Brainstorms

Property/Casualty Insurance Company Investments

by Stephen W. Philbrick

In two previous columns, I talked about using our actuarial expertise to make informed personal investment decisions. In this column, I'll discuss the same concepts as applied to a property/casualty insurance company.

The CEO of a P/C insurance company has a long list of critical success factors, but two factors are very high on the list:

P/C actuaries have long been involved in helping the CEO with the first issue. Our compatriots in the life industry are more active in the investment arena than P/C actuaries are, but I anticipate that P/C actuaries will become more active in this area. I'm convinced that good actuarial training is an excellent, albeit incomplete, introduction to the key issues in the investment world.

In the underwriting world, there is a natural tension between the twin concepts of diversification and specialization. P/C actuaries often use the term "law of large numbers" rather than "diversification" but the concepts are related. Companies look for a spread of risk, by writing different lines of business in different locations. This must be balanced against the advantages of specializing—understanding a particular risk and becoming an expert in it. Both approaches can be successful, as witnessed by the simultaneous existence of single-state, single line companies alongside global multi-line companies.

On the investment side, the twin concepts also exist. Diversification is most famously embodied in a classic paper by Markowitz, while specialization is articulated well by Peter Lynch, the former manger of the Fidelity Magellan fund. One of his rules for investors is, "Be sure you know what you own, the company behind the stock." In the investment world, the balance is tipped in favor of diversification, at least in terms of educated advice, if not actual practice. Lynch urged investors to do homework into companies, and pick companies where they had some specific knowledge, but he still urged that investments should be spread over a number of companies.

P/C insurance companies are naturally most knowledgeable about P/C insurance companies. Following Peter Lynch's advice would mean that a fair portion of the equities purchases of a P/C insurance company would be in companies they know well: the stocks of their competitors.

In contrast, the diversification argument limits the amount that should be allocated to the equities of P/C insurance companies. Some proponents go so far as to advocate an indexing strategy. Rather than attempt to outguess the market, they advocate holding a market-basket of stocks, in an attempt to mirror either the S&P 500 performance or a broader market such as the Wilshire 5000.

We can test whether companies do this. A review of the Schedule D for P/C insurance companies reveals that approximately two percent of the equity investments are in P/C insurance companies. In comparison, P/C insurance company stocks make up about 3.5 percent of the S&P 500. The investment managers of P/C insurance companies, on average, have underweighted the P/C industry in their portfolios.

I talked to a portfolio manager who told me that this phenomenon was deliberate with some companies, but not for analytical reasons. Some companies deliberately chose not to buy the equities of their competitors because of the public relations reaction. How do you tell your stakeholders that you have a plan to outperform your competition, yet you are buying the stock of your competitors?

I think they are making the right decision, although for the wrong reason. In the case of an individual, holding a market-basket of stocks that mirrors the entire market is the ultimate in diversification. (Because of transaction costs, an individual will normally achieve this with an index fund.)

When investment experts recommend an index fund, they are making the implicit assumption that the individual's financial position is not correlated (to any meaningful degree) with any segment of the market. This is largely true for an individual (although the point of my earlier articles was to counsel against exposure to the stock of one's employer). However, this assumption is clearly not true in the case of a P/C company. The financial results of any particular P/C insurance company are very likely to be correlated with other P/C companies and, to a lesser degree with financial corporations in general. Note that this correlation can occur, even if your financial results are superior to those of your competitors.

As a consequence, even a company that has chosen to take an indexed approach to equity investing would be wise to consider a modified index, one that excludes P/C companies, or possibly all financial companies.

Those companies already excluding P/C companies from their portfolios can now point to a valid financial reason for this exclusion, rather than simply doing it for appearances sake.

I haven't run a quantitative analysis of this approach, measuring the correlation of particular companies with similar financial companies, and determining whether the exclusion improves the portfolio to a meaningful extent, but I think the logic appears sound. I would appreciate feedback, particularly from readers who may be involved in the investment decisions of P/C insurance companies.