No Easy Answer in Debate over MTM versus HC Accounting standards

It’s probably an indication of this country’s wealth that among the many policy controversies that consume our limited attention, accounting rules could ever be one of them.

A few years ago, a proposal from the Financial Accounting Standards Board (FASB) to treat employee stock options as expenses on the corporate income statement drew lots of attention and heavy lobbying from industry. And a few years prior, the much ballyhooed Sarbanes-Oxley Act responded to perceived problems in auditing public companies. So the current furor over mark-to-market (MTM) versus historical cost (HC) rules -- especially in light of the current crisis -- should probably surprise no one.

But the debate has progressed in polemic fits and starts, with adherents on both sides clinging to what may be considered myopic concerns about the role MTM has played or not played in creating the crisis we now find ourselves in. To wit, back in March, the chief accountant of the Securities and Exchange Commission and the chairman of FASB were brought before a subcommittee of the House Financial Services Committee and publicly harangued about their failure to make significant changes to current accounting practices. When the witnesses pushed back that knee-jerk reactions to the crisis could lead to worse policies, members insisted they be given a timetable of less than a month in which MTM would be changed more to Congress’ liking. 

What the debate could use is an agreement on exactly what accounting regimes can reasonably do, followed by an assessment of how each one accomplishes those goals. Financial crises will continue to occur, but they may not look exactly like this one. And tailoring policies to prevent current problems may setup different future ones. So the best strategy is to set policies according to immutable economic principles. Accounting standards -- from an economic point of view -- ought to accomplish two things: reduce transaction costs and reveal truthful, useful information.

Transaction Costs and Information Revelation 

An apocryphal history tells of English medieval knights determining the flow of traffic. Being predominantly right-handed, the knights tended to ride on the left-hand side of the road in order to have ready access to their weapon should a foe pass the other way. Later Napoleon, being left-handed and something of a megalomaniac, decreed all French territories would travel on the right. Finally the US, being heavily colonized by the French and wanting to cast off any British customs, adopted the right-hand side rule of the road. 

Having now laid down our swords, and the car being neither a right-hand or left-hand dominated operation, it’s pretty arbitrary which side of the road we drive on – except that you better go along with everyone else. 

And so it goes for many of the standards we live by, be they shoe sizes, Internet protocols, or accounting rules. Standards provide coordination among many potential competing ways of doing something. And with respect to this coordinating function, it’s almost irrelevant what the standard itself is, so long as everyone understands and follows it. The whole point of a coordinating standard is to minimize transaction costs – the costs surrounding the actual transaction or economic activity you wish to accomplish. 

In the example of driving, for instance, one can suppose it would be possible not to have any set rules on which side of the road to drive on (and setting aside that a natural pattern might emerge on its own over time). The transaction for each driver is bypassing each other driver without getting into an accident. In this scenario, each driver would have to continually “negotiate” with every other adjacent driver about which side on which to pass. If you add up all the extra time and effort to get from point A to point B, the sum would be the total transaction costs incurred. 

In accounting, coordination around a standard way of reporting means interested parties do not need to continually learn and interpret each company’s specific way of presenting information, which minimizes the transaction costs of acquiring information. With or without a standard (and separate from the question of whether it’s a good standard or not), the ultimate goal of acquiring useful information is the same. So accounting rules must be judged according to the criteria of how well they minimize the amount of effort necessary to get the information you want. 

The other criterion by which accounting rules should be judged is the degree to which they reveal useful information. In auction theory, for example, scholars have devised clever ways to induce bidders to reveal their private, subjective valuation for a given object. The advantage of such innovations in auction design is that sellers can maximize their surplus and the bids themselves are information-rich. 

Let’s assume for the time being that companies are not systematically attempting to deceive investors and the public, in part because strict accounting and auditing protocols make it difficult, and also because reputational capital is difficult to replenish. So then accounting rules should not so much be devised to tell the truth (whatever that means) but simply to reveal the most useful information. Taken together with the first criterion, investors should be able to use the information presented with as little manipulation or interpretation as possible. 

The Debate 

Each accounting regime has its own limitations and advantages. As Arnold Kling pointed out in The Chess Game of Financial Regulation, MTM was itself a response to the long lags and inaccuracy of HC. In the run-up to the Savings & Loan crisis of the late ‘80s, it was discovered that principals of S&Ls were able to hide the true financial health of their institutions, especially from capital reserve requirements, because loans that began to fail did not show up on the balance sheet as such. As a result, a MTM regime was thought to give the most up-to-date, accurate information on the financial health of an institution. The market transaction price for a given class or type of asset, so the argument goes, is the closest thing we have to a timely and accurate valuation. 

But then came the credit crunch of 2007-08. As demand for liquidity suddenly surged, banks were forced to sell off assets into an already declining market. As prices declined, and the market itself became thinner and thinner, MTM created a vicious cycle which many blame for severely exacerbating the crisis. At its lowest point, there simply was no market for certain types of securities which made valuing them nearly impossible under current rules. 

So given the desire to minimize transaction costs and maximize the usefulness of information, which regime is preferable? 

Broadly speaking, both minimize transaction costs of acquisition as long as they employ the fewest number of exceptions. For example, a few weeks after the recent House Financial Services Committee subcommittee hearing, FASB issued a rule that allows institutions to mark-to-model (as in use an internal valuation method, not based on either MTM or HC) certain assets, as long as they make note of it in their reports and give some guidance about how they went about coming up with those non-MTM and non-HC figures. MTM would still be the default rule. So even assuming investors are able to back out the numbers they really want from those mark-to-model figures, this entails a non-trivial increase in costs because each company’s model is different. 

As far as usefulness, it’s safe to say that investors will only rarely look for someone else’s internal model-generated numbers. Investors, in fact, often use their own models to determine the worthiness of a given balance sheet. What then is most useful is data they can input that has been manipulated as little as possible and is based on some uniform reference or standard that applies to everyone (such as a pure MTM or HC regime would require).

A deep academic literature has explored the information-rich nature of market prices, which would seem to score a point for MTM over non market-based standards. But more recent work by Guillaume Plantin (London Business School), Haresh Sapra (University of Chicago) and Hyun Song Shin (Princeton University) demonstrates in a formal way that for the very balance sheet items of banks and insurance companies, MTM leads to suboptimal decisions which ultimately degrades the information value of prices, such as in the case of the downward price spiral in the current crisis. Because markets themselves are not without transaction costs, actual prices do not reflect the full information that theoretical “perfect world” prices would. Ultimately this reinforces the notion that both regimes (and perhaps all regimes) necessarily involve tradeoffs. 

Conclusion

In the light of these economic realities, the debate over MTM versus HC becomes much less clear than many proponents of each would have one believe. It is not clear that one is clearly superior over the other. While MTM was introduced largely to correct certain deficiencies of the previous HC system, MTM itself introduces a new circumstantial problem in certain environments like the current one. The deep-rooted desire to now correct this new problem by returning to HC is understandable. But a rush to do so is ill-advised until additional research and a full accounting (as it were) of the various concerns and tradeoffs is made.

About the author