|
CAPM in Insurance
The key underlying assumption in applying a Surplus allocation in rate regulation is not
the authors’. The regulatory model in Massachusetts and other states is that insurers
should only earn profits consistent with those earned on other investments of
comparable risk. This is based on the U.S. Supreme Court’s Hope Natural Gas decision,
which is usually interpreted to mean the profit margins which would be calculated using
the Capital Asset Pricing Model (CAPM) method. That is, only the systematic, nondiversifiable
risks within an insurance portfolio will (and should) earn a return above
Treasuries, and that return is the same as a stock with equal correlation to the overall
market would earn. MBA students are traditionally taught that a dynamite factory is a
systematically less risky investment than a diversified, but leveraged, stock portfolio.
CAPM argues that the capital market cannot allow the dynamite factory to earn a higher
return than an investment with a comparable beta factor.
A recent empirical challenge to the use of CAPM in insurance pricing is the market for
Catastrophe bonds. These bonds earn significant excess returns on their total risk, even
though that risk is seen as diversifiable. Catastrophe bonds have a low beta: some
authors (example: Kenneth Froot) suggest that catastrophe insurance risk has a zero
beta. However, Catastrophe bonds actually earn returns very significantly above
equivalent Treasury bonds (see the recent work of Morton Lane). Contradicting the
dynamite factory analogy, investors require an extra reward for the clear, but
diversifiable, chance of a catastrophic loss of their investments.
We see that the CAPM results do not currently hold for insurance investments. This is
only possible in the long term if the essential assumptions of CAPM do not apply to
insurance investments. Do they? Insurers may not borrow or short sell without limit.
Capital does not move freely into and out of insurance companies. Insureds and insurers
and their investors and regulators all have different time horizons. Insurance contracts
are not transferable or divisible. Several of the key assumptions underlying CAPM rate
regulation clearly do not apply, although regulators persist in applying them out of
respect for precedent or for lack of a practical alternative.
One alternative possible now for regulators is to model a notional portfolio representing
a mix of Catastrophe bonds and Treasury bonds that matches the degree of total risk of an insurance product. The excess return on this notional portfolio would be an
appropriate profit provisions to include in rates. This provides an objective measure of
return without relying on the CAPM’s assumptions, which we can see are violated both
in theory and by market results. The alternative would however provide a nonconfiscatory
return “of equivalent risk” as required by Hope.
A sample calculation is
shown on page three of the exhibit.
A more direct analysis, that was also impossible in the past, is to look at the long-term
loss ratio for the same product in the many states that now have vibrant and effective
competitive market. This would provide a direct benchmark of the relationship between
price and risk that Hope says insurers should earn. This comparison would also allow
regulators to ignore differences of expense and product mix that insureds do not care
about. Insureds only value what they receive as expected loss recoveries, defense and
cost containment expenses, loss control services (and perhaps premium taxes, as a
surrogate for sales tax that they might pay to replace their damaged property). This
different approach could be viewed as “demand side” regulation.
|
Previous