The risks of mark-to-market valuation are known and regulated for, but the main issue is whether the alternatives, historical cost and fair value, or cyclical PIT/TTC combined values, or what regulators are now called "hold to maturity" valuations, can be trusted that are not subject to gaming on or off balance sheet accounts, in or out of the trading book, banking book, or clients' investment funds?
On top of the historical cost to value assets (assets booked at the price at which they were bought), fair value can employ discounted cash flow (DCF) to take account of market turbulence plus VaR volatilities and re-price with a longer-run valuation than mark-to-market, but only if the assets are not held for trading at any time. In DCF models long run fundamentals, actual cash flow forecasts, have a greater role. Actually, this really depends on whether assets are held for sale (in the trading book) in which case mark to market and VaR should operate, or held to maturity in which case the assets may be transferred to the banking book, or more the latter than the former, in which case the assets may continue to be held in the trading book valued at 'fair DCF value'. The maturity periods are important, but some banks' risk accounting systems are weak on recording all these parameters in ways whereby they can be analysed, including collateral and other security, and various related legs to the dealings in and around the assets.
A DCF fair value model may seem appealing but it is not without problems. How to forecast future cash flows over an economic and credit cycle? Which discount rates to predict for such cash flows including setting the risk price margins for the various 'stakeholders'? Basel II recognises the role of judgement and insists on various valaidations and risk assurances underpinning the estimates given that those
using this model must to some extent value assets subjectively and may game the models. Internal risk rating models are allowed, but in practise they tend to be merely variations of external rating agencies gradings whose through-the-cycle and point-in-time models are opaque and hard to discern TTC from PIT, one from the other. Inevitably a DCF approach will vary among different firms. This is allowable gien differing strategies, but relating strategies to how valuations are arrived at is contentious and basically beyond the intellect of bankers who can do mathematics but cannot do economics. Within bounds (however these can be determined) DCF fair value should be allowed to price valuations differently of identical (or almost identical) assets and that firms should thereby have some degrees of freedom to succeed or fail according to their individual quality of approach.
The idea, in any case, that there is a reliable mark-to-market price that is certain at any time and not negotiable or varying across OTC markets and deal types is something of a myth, especially when these markets are insufficiently liquid. This is one reason why the problem is one of the regulation of markets and bringing the transparency of regulated stock markets and derivatives exchanges to the OTC capital markets.
Some, including the so-called gloomster Nouriel Roubini, and myself too, argue that the problem is not mark-to-market accounting but the pro-cyclical capital requirements of Basel II that can amplify economic and credit cycles. But even without such in-built pro-cyclicality banks do retrench leverage and credit if a bit late in a downturn and late again in an upturn, while today we see them retrenching too much and too fast during periods of turmoil when they suddenly panicked into being extremely risk averse. Roubini suggests solutions should be symmetric, i.e applied both during periods of rising asset prices and bubbles (when prices are above fundamentals) and when such bubbles go bust (and asset prices may fall below fundamentals as they have now).
But, such symmetry is likely only to serve as a base-case around which firms will vary according to their judgement and individual circumstances. Their ability, however, to know where they are in the context of the underlying economy of cycles etc. is not yet established. There is a weak literature and no detailed prescriptions from the regulators, only expressions of anxiety about what happens when banks act pro-cyclically. Why, because then only Government is left with enough financial clout to pull economies out of recession, which is precisely what we are seeing today.
Fair value, scenarios and stress-test models are part of accounting practise envisaged by IFRS-7. But this time-based dynamic forecasting view goes against the grain of an accountant's mind, perhaps why David Tweedie, the Scot who chairs the IASB, is so insistent that the only fair value is mark to market accounting.
The now obvious 100% responsibility for anti-cyclical spending residing with governments that is part of the Federal Reserve's remit, but not that of the ECB, is why there are voices in America cautioning against a central bank culture in Europe that constrains interest rate settings and governments' fiscal stance to an anti-inflation focus only.