WHEN a company goes bankrupt, recriminations tend to follow. Even so, the fury caused by the recent collapse of Carillion, a British contracting firm, is unusual. A report on the debacle by British MPs, which was released this month, savaged everyone from the firm’s executives to its regulators. But the MPs reserved special bile for the Big Four accounting firms—not just KPMG, which audited Carillion’s accounts for 19 years, but also its peers, Deloitte, EY and PwC, each of which extracted fees from the company, before and after its fall. The MPs have called for a review into the audit market and asked it to say whether the Big Four’s British arms should be broken up. The row is local, but concerns about the industry are global.

Critics of the auditors are right in two respects: that the industry matters, and that it needs reform (see article). It is in everyone’s interest that auditing works. If investors cannot trust financial statements, then companies’ cost of capital will rise, crimping growth and employment. It is also true that the industry has flaws. It is highly concentrated. The Big Four audit 98% of the companies listed on the S&P 500 and the FTSE 350 indexes. And auditors are paid not by investors, whom they serve, but by the company whose accounts they scrutinise. That raises questions about objectivity, especially since the Big Four earn nearly twice as much from consulting and other services as they do from auditing. Past reforms banned them from providing both an audit and certain consulting services to the same client, but conflicts of interest remain. In America non-audit fees charged to the same client amount to a quarter of audit fees; in Britain the figure is around a half.

A break-up, whether to separate the audit arms from the consulting businesses or to turn the Big Four into a Middling Eight, seems to offer a simple solution to these problems. It would at first affect only the British parts of the firms’ global networks, but the idea could spread.

Although a break-up might be justified as a last resort, it is premature. Investors have exaggerated expectations of auditors’ ability to detect fraud. Because audits rely on sampling, some skulduggery will inevitably slip through. There are also signs that the industry is improving. Many countries tightened the rules after a scandal in 2001 sank Enron, an energy-trading firm, and its auditor, Arthur Andersen. In America the number of accounts that are restated because of a material error has fallen sharply over the past decade. Break-up would bring unintended consequences. As the world economy shifts from making goods to selling services, auditing is becoming more complicated: scale and the multidisciplinary expertise of large firms count for more. Smaller firms risk being too reliant on a few large clients, which may cloud their judgment.

If you want radical fixes, there are better ways to correct the incentive problems at the core of the industry. You could sever the link between auditors and their clients by requiring securities regulators to pick firms’ auditors. Or you could introduce mandatory insurance of accounts, whereby companies must buy coverage for losses from accounting errors and the insurers would therefore appoint auditors to assess their risk.

One bean at a time

Such ideas have been floating around for years, but even these are too hasty. Instead regulators should sharpen tools that are already available in Europe. They could lower the cap on non-audit fees charged to an audit client from today’s generous level of 70% of the audit fee. Under rules introduced in 2016, British companies with the same auditor for ten years must re-tender; they are forced to rotate after 20. Such rules look draconian to American eyes, where the average auditor tenure for the first 21 companies in the Dow Jones Industrial Average to have made disclosures this year is a cosy 66 years. New research finds that auditors are most likely to find misstatements early in their tenure; by the tenth year, the benefits of a fresh pair of eyes are lost. Academics also find that the Big Four’s fees rise with tenure. Even Britain’s 20-year limit is too long.

Auditors in many countries are already required to add flesh to the bare bones of the audit opinion. That is to be encouraged. Transparency over the main points of contention with management, and the size of revisions made to the accounts as a result of scrutiny, would cast light on auditors’ successes, not just their failures. And that in turn would help investors to assess auditors’ performance.

For years shareholders have waved through a company’s choice of auditor at annual general meetings. A bit more bolshiness could be salutary. Last month, for instance, over a third of investors in General Electric voted against the reappointment of KPMG, its auditor for 109 years. The case for breaking up the Big Four is unproven. But every so often, shareholders need to remind the quartet who their main customers are.