Minority Report on Fair Value
by Philip E. Heckman
I thank The Actuarial Review for this opportunity to respond to some comments made by Steve Lowe after his excellent presentation on fair value accounting at the CAS Spring Meeting’s closing session. These remarks had to do with the issue of whether recorded liabilities should be discounted to reflect the debtor’s credit standing. Steve enjoys well-earned regard and influence in the Society, making these comments impossible to ignore.
First, a little background is in order. Fair value is a global accounting reform that has been a long time in the making. The International Accounting Standards Board (IASB) is the body chiefly responsible for formulating and promulgating the changes, though FASB has had a forward role in forming opinion. In traditional practice, accountancy has centered on a set of rules for recording and summarizing transactions, leading to a balance sheet as an end product. Under fair value, this paradigm is reversed. The aim is to place realistic market or economic values on an enterprise’s assets and liabilities. Liabilities, which do not trade, can be placed on a market basis by using guarantee pricing. This is where the credit standing issue comes in, since IASB has defined the fair value of a liability so that the guarantor’s credit standing is unspecified. Traditional accounting for debt records the transactions in such a way that the first recognition of a debt liability is discounted for credit standing. IASB/FASB would extend this to all liabilities, on update as well as at first recognition.
Debt issuance is probably the clearest way to outline the problem. Suppose there are two companies, A and B. A has superior credit, B inferior. Each issues debt on the same day, promising to pay $10,000 in 10 years with no buyout option. A receives $6,000 in cash; B receives $4,000; a default-free borrower receives $8,000 for the same obligation. Traditional accounting for debt says that A posts an initial liability of $6,000; B posts an initial liability of $4,000; a default-free borrower posts $8,000. I submit that they are all three under the same obligation and should all three post $8,000 and amortize at the default-free rate. The defect I see in traditional accounting is that the borrowing penalty—$0 for the default-free borrower, $2,000 for A, $4,000 for B—is amortized over the term rather than being treated as a (deductible) expense (prepaid interest, if you will) impinging at inception. The closest parallel I can see for the traditional treatment is handicapping in sport, where a superior contestant is placed under an intentional disadvantage in order to make the outcome harder to predict. (A superior racehorse carries extra weight; a superior borrower posts a higher liability.) This makes for good sport and terrible finance. I maintain that simply participating in an economy where other companies keep their books this way puts insurers at an unfair disadvantage.
The CAS has been alert to the Fair Value reform issues since 2000 when Ralph Blanchard chaired a task force that produced a white paper on fair valuation of P&C liabilities. The white paper addressed several topics, including risk adjustment of reserves, but withheld comment on the credit standing issue. In 2002, the American Academy of Actuaries (AAA) convened a task force on fair valuation of insurance liabilities with a mandate including life and pension as well as property/casualty. This too was inconclusive on the credit standing issue, but with evidence of vigorous debate.
As the 2005 deadline for the IASB insurance project approached, the CAS Committee on Theory of Risk drafted an RFP for quantitative research on the fair valuation of P&C liabilities, chiefly loss and expense reserves. When the CAS Executive Council (EC) reviewed the RFP, the project was deemed so important that it was handed over to a task force chaired, again, by Ralph Blanchard and staffed by EC members. The project was awarded to two proposers: Tillinghast, a business of Towers Perrin; and PricewaterhouseCoopers LLP. The instructions to the contractors specified that discounting for credit standing was to be omitted.
Thus it is that the project reports presented at the May 2004 meeting show loss reserve valuations discounted on a default-free basis and loaded for uncertainty in timing and amount. This is the procedure advocated by Bob Butsic in his 1988 Michelbacher Prize paper. It is a valuation basis that many colleagues and I agree with and that gives few if any casualty actuaries any trouble. It also represents a plausible regulatory outcome. Let me emphasize, however, it is not IASB/FASB fair value. Unless they have very recently changed their tune, these bodies would prescribe, unfathomably, that, in GAAP fair value financials, the reserve balances should be discounted further to reflect the insurer’s ability to pay. This amounts to discounting at the insurer’s own borrowing rate. Rightly or not, the CAS has skirted the issue of the effect this would have on insurer balance sheets, and, I would argue, has deprived the membership of a real eye-opener.
Enough background. Let me now respond to Steve Lowe’s remarks.
Steve remarked, first, that the current credit standing issue is a peripheral matter that should not concern the CAS and its members. This is certainly true in the narrow context of CAS members reporting to regulators under statutory accounting. If this were the only arena where CAS members are or will be active, we would have no cause for concern. Our members chiefly make their living valuing certain specialized liabilities, notably casualty loss reserves. These reserves appear on various balance sheets and take their meaning and purpose from the context, whatever our original intentions. Outside the shelter of statutory reporting, that context is dictated by IASB/FASB. The problems attending the proposed valuation rules are not yet firmly on the CAS radar screen, lying around a corner we haven’t turned yet. For our brethren on the life and pension side, they are a matter of very active concern.
Steve also remarked that insurance and other regulators would never allow reporting of statutory liabilities discounted for own credit risk. This is very likely true, but it solves no problems. If statutory reserves are accepted in GAAP statements, it means that insurance companies will report liabilities at full default-free value while competing in the equity markets with companies in other industries who report liabilities discounted for their own credit standing, at first recognition and on update, booking phantom equity and spurious earnings (and paying tax on the float). If statutory reserves are not accepted under GAAP, actuaries will be forced to report and opine on GAAP reserves discounted for credit risk–a situation not considered in the studies sponsored by the CAS. Insurers will also own the debt of companies who keep their books this way and may slide into bankruptcy with little or no warning. Unless we want to spend half our working life drafting disclaimers, we must take notice of the situation outside the insurance industry. There is no substitute for a level playing field but considerable doubt as to whose job it is to ensure it.
Steve also observed correctly that GAAP is intended for reporting by going concerns and not for bankrupts. We must wonder, then, why current GAAP requires first reporting of debt liabilities discounted for the debt holder’s own credit standing. The inconsistency is clear. Would that we could rely on IASB/FASB to get this right, but as remarked above, liability valuation is a new thing in accounting practice. In fact, there is no consistent theory of liabilities and no global consensus as to what reported liability balances should mean. This condition puts our own profession in peril since liabilities are what we do. It also means that IASB/FASB, want it or not, need all the input they can get, on general as well as insurance-specific issues, and sooner rather than later. In my paper for the 2003 Bowles Symposium, published with revisions in the North American Actuarial Journal (NAAJ) in January 2004, I have attempted to address this issue—somewhat stridently, perhaps, but no more so than the need and occasion required.
It has been widely suggested that the public information problem can be solved simply by requiring reporting of the current value of the insolvency put in the financials. This is true, but it is not simple. The insolvency put is the corporate owners’ option to walk away from the wreck of a defunct corporation without further payment. It can be valued by taking the difference of the default-free valuation of liabilities and the risky valuation. The difficulty is in convincing the accounting community that liabilities valued on a default-free basis are desirable, meaningful, and worth the trouble. In many financial economic treatments of the liability problem, this put—clearly an asset of the owners and not the corporation—unaccountably shows up as an asset on the corporate balance sheet, contributing phantom equity and lending false support to the proposed IASB/FASB fair value treatment of liabilities. Louise Francis, Don Mango, and I are putting together a paper on this subject for presentation at the August 2004 ARIA meeting in Chicago. We’re not accountants; we’re not financial economists. It’s not our job, but somebody has to do it.