When a company and its auditors both shred documents, something has gone badly wrong with the accounting.

The events at Enron require reforms both in the nature of the accounting rules and in the way those rules are policed.

Take the accounting rules first. Much of what went wrong at Enron concerned deals between the company and special-purpose entities (SPEs) in which it held minority stakes. These shovelled debt off the Enron balance sheet, flattering underperforming businesses with incestuous trans-actions.

Such self-dealing has a long, undistinguished pedigree. The Financial Accounting Standards Board's standard FAS 57 requires disclosure of such relationships. At Enron this was both selective and impenetrable.

Even if FAS 57 had undergone its long-overdue updating, it is questionable whether disclosure would have improved. Creative accountants in Houston were putting a cosmetic veneer over economic reality, observing the letter, not the spirit of the law.

The Enron saga reveals the limits of rule-making. Clearly there have to be detailed standards. But very detailed prescriptions about how transactions should be treated have led to an accounting quagmire.

In taking the prescriptive approach, the US has been at odds with other countries. The alternative view, known as "substance over form", requires auditors to pass professional judgment on how far the accounts reflect economic reality.

It is true that this assumes auditors will act properly, a brave assertion in the current climate. But the alternative, legalistic approach will always provide loopholes for financial engineers. A traditional response to Enron would thus generate more rules, more bureaucracy, higher transaction costs, less economic efficiency and more scandals.

The reform that is needed, therefore, leads away from US Generally Accepted Accounting Principles' very detailed prescriptions towards the approach of the new standards devised by the International Accounting Standards Board. The IASB approach, which emphasises substance over form, has been treated with mild suspicion in the US. But the Enron case, by revealing the limits of the prescriptive approach, has given impetus to the case for replacing national standards with IASB ones, or at least with rules sympathetic to the IASB framework.

It would be a pity to miss this opportunity, in other countries as well as the US. And to achieve this convergence, the IASB will need stronger, more independent sources of finance.

A change of approach does not solve all accounting problems. In reporting on its SPEs, Enron failed to follow even the flawed disclosure rules in place. But by giving a precise (and, at 3 per cent, absurdly low) figure for the equity a partner must own in order for the partnership to be taken off the parent's balance sheet, US standards encourage companies to fall just the right side of the line.

It would be preferable to focus, as some other countries' standards do, on questions of control. The SPEs run by Andrew Fastow, Enron's chief financial officer, would arguably have failed the control test. But even where this test is used, off-balance sheet vehicles have proliferated. It is debatable just how independent many of the SPEs set up by financial institutions, for example, really are from their bank parents. Accountants should consider much higher hurdles before these vehicles are allowed to escape consolidation into the parent's balance sheet. A corollary is the need for clearer disclosure of derivatives positions.

There are other areas of accounting weakness. Telecommunications networks have been swapping capacity with each other, to add artificial revenues. Companies have emphasised pro-forma earnings - sometimes known as "earnings before the bad stuff" - to show a prettier picture. If accounting is accurately to reflect corporate activity, accounts will be much less predictable. Investors will have to adjust to bumpier earnings.

Improved standards require improved policing. Auditing, the first line of defence, failed at Enron. Exactly how Andersen fulfilled its duties as Enron's auditor deserves close examination. There is no doubt, however, that accountancy firms face a conflict of interest in offering consultancy to audit clients. Accountants have emasculated proposed reforms in this area but Enron leaves no option: this practice must end. Regulators around the world must ban auditors from providing non-audit services - including tax advice - to their audit clients.

Even then, auditors can become too cosy with longstanding clients: Andersen's Houston office audited Enron since 1984 and the company hired Andersen personnel into senior finance jobs. Auditors must be compulsorily rotated every three to five years. Companies should not hire finance staff who have worked as their auditors. Because there are only five big global accounting firms, choice of auditors is restricted. Competition authorities must stop further shrinkage.

The legal fiction that auditors are appointed by shareholders, not by corporate management, must become reality, with the audit committee seizing control of the appointment process. And the auditor's statement at the end of a company's accounts must highlight issues that shareholders should be aware of.

Last, the accounting profession needs tighter supervision. Around the world, peer review of auditing practices has proved inadequate. Statutory regulation is required, to global standards.

As a backstop to auditing, corporate regulators must monitor corporate accounts better, with special emphasis on complex companies. Depending on the country this task might fall to securities regulators, company registrars, or stock exchanges.

These reforms will raise the costs to companies. But shareholders must understand that that is a small price to pay for accounting accuracy.