Actuarial Review Return to Main Page

How DFA Can Help the Property/Casualty Industry, Part 6
DFA—A Capital Idea

by Susan T. Szkoda

Part 1 , Part 2, Part 3, Part 4, and Part 5 are available online.

This article is the sixth and final in our series on Dynamic Financial Analysis (DFA). The previous articles covered Business Planning, Underwriting, Pricing and Reserving, Reinsurance, and Investment aspects of DFA. This article pulls these topics together to discuss one of the major uses of DFA-determining optimal levels of capital necessary to support the company’s business plans.

The question of how much capital is "enough" has been answered in a variety of evolutionary ways. When I started in the business about twenty years ago, that question was usually answered by the "Kenney Rule"-the premium to surplus ratio should not exceed 2:1 or 3:1. Over the years this proved to be an insufficient answer. In the early 1990s the NAIC introduced the concept of Risk Based Capital (RBC). RBC was a quantum leap forward from the old Kenney Rule. In fact, RBC examines many of the same factors as DFA-premium growth, line of business mix, reserve adequacy, reinsurance and investments (among others). However, it relies on a formula approach based in part on broad industry averages. RBC also excludes explicit consideration of catastrophe risk and reserve/pricing risk arising from mass torts. While RBC is of benefit to regulators, company managements and rating agencies generally view this measure as somewhat simplistic.

I believe that DFA represents the next step in this evolutionary process. Sophisticated DFA models can capture many of the most critical variables that will affect the company’s capital needs over some unknown future business plan horizon and generate thousands of scenarios that, in the aggregate, can help management determine capital confidence levels. Ultimately, management (and rating agencies) will need to determine how much is "enough" for any individual company. A rating agency using a DFA model will need to consider the confidence level necessary to merit AAA rating vs. AA rating and so on.

Richly complex DFA models will consider not only internal company factors but broad external economic factors. These include rates of economic growth in various economies, the capital markets, interest rates, inflation, the general business cycle, the specific insurance cycle AND appropriate linkages between these factors and internal insurance company business concerns such as frequency and severity trends by individual lines of business, and ability to achieve adequate prices.

Sophisticated models will also ultimately consider new products and new competitors that may enter the market and their impacts on existing market participants.

One of the great benefits of DFA models is their ability to incorporate the notion of variance (risk) around each of the potentially dozens or even hundreds of "expected value" variables affecting an insurance company.

Most business plans, even today, consist simply of a base case (expected value) scenario. This base case may be augmented by one or two pessimistic scenarios and one or two optimistic scenarios. Business planning as we know it today is unequivocally not up to the challenge of helping management set reasonable capital levels and capital allocation strategies. In many companies the business plan is little more than a moving target or goal that is subject to constant revision during the course of the year.

There are a variety of DFA models currently in existence. Based on company feedback, several seem to have a fairly strong capital orientation. These models seem to generate more interest from the CFO and senior management of the company than those that present themselves with less capital focus.

The models available today are helping companies to analyze their capital adequacy and to set capital allocation strategy. However, I still see too many models that randomly pick loss ratios from year to year without embedding them holistically and cohesively in multi-year external economic scenarios or even recognizing that year-to-year loss ratios for a particular line/company are not usually random independent events.

I believe many of the next generation DFA models will be more powerful. The next generation models will rely heavily on proprietary research linking external economic events with changes in internal insurance company trends. I believe the CAS should actively sponsor significant levels of research to help identify these critical "linkages."

Looking ahead, I anticipate that DFA models will become indispensable in strategic business planning, merger and acquisition valuation, rehabilitation/runoff evaluation and capital planning.

The future is indeed bright and intellectually challenging for DFA actuaries!

(We hope you will attend the DFA seminar in July in Boston and our hands-on Limited Attendance DFA seminars(s) to be given this fall).