Traditional accounting is a wonderful system for counting beans. For reporting business performance, it has a fatal flaw: reported profitability is affected by the volume of activity. Not only does this create an almost irresistible incentive to “short-termism”, it pollutes national statistics and it reinforces swings in the economic cycle by exaggerating news, whether good or bad.

Suppose, for example, Short-Term Exploitation Ltd is on course to deliver an annual return of 15 per cent. Long-Run Development Ltd is on course to deliver 30 per cent. If the bulk of STE’s activity occurs early in its life, whereas LRD’s occurs later, the return reported in STE’s early accounts can be much higher than 15 per cent and it may be many times greater than the return reported for LRD. Because they routinely send such false signals, accounts prepared strictly in accordance with International Accounting Standards Board rules probably cause more economic damage than all the high-profile financial scandals put together.

Instead of repairing this fundamental defect, “fair value” accounting – where some items are marked to the market price at which they could be sold or transferred – is likely to make matters worse. Since the mere presence of an item in a company’s accounts is proof that its sale or transfer has not, in fact, taken place, a fair value represents the hypothetical proceeds of an imaginary transaction. This risks introducing more distortions.

Conventional academic wisdom that an increase in an asset’s value necessarily represents a gain to its owners is a fallacy: it may actually make them worse off. Not only does fair-value accounting encourage even more exaggerated swings in the economic cycle, it can misreport actual losses as fair value gains and actual gains as fair-value losses.

There is no objection to the disclosure of fair values on company balance sheets (best practice under the historical system); but to incorporate them throughout the accounts produces a mishmash of actual transactions and hypothetical market values – sometimes unintelligible even to the company’s financial directors.

Since the reporting of changes in fair value as gains or losses in the profit and loss account can make a low investment return look better than a high one, fair value accounting is useless for comparisons of corporate performance. By promoting the fallacy that an increase in the value of an asset necessarily makes its owner better off, it has repercussions throughout the economy.

The housing market is an obvious casualty. The amount homeowners can afford to borrow depends on their ability to service the loan – normally out of income. The fair value fallacy encourages the mistaken belief that it is safe to borrow or lend, provided the loan is covered by the market value of the property. The result is a spiral of ever-increasing loans pushing up property values, which make possible ever-increasing loans. Pyramid schemes are normally against the law. Pyramid lending, however, has been made respectable by a fundamental error in accounting theory.

It is easy to put all the blame for the credit crunch on irresponsible lending, overexuberant banking and creative accounting. Yet the only lasting cure for bad financial practice is the elimination of bad financial theory. Removal of the perverse incentives of the accounting system requires a clear distinction between what is fact and what is opinion. That would make possible a system for publishing the return on capital that companies are actually planning to deliver and for monitoring progress towards the goal.

Irresponsible lending can be discouraged by making legal enforcement of loan contracts, particularly repossession of property, conditional on due diligence on the part of the lender. Faulty risk analysis can be corrected by the exposure of wishful thinking masquerading as mathematical rigour. However, false accounting cannot be eradicated as long as the IASB insists on promoting as “fair value” a system that is, quite frankly, fraudulent.

Anthony Rayman is the author of Accounting Standards: True or False? Routledge 2006