Why do so many large US companies diverge from generally accepted accounting principles when reporting their results? In 2017 no fewer than 97 per cent of S&P 500 companies used home-made measures in earnings releases — up from slightly more than half two decades before, says Financial Times.
The number of non-Gaap items in each one has also been rising sharply, from slightly more than two per statement in 1996 to more than seven in 2016. The conventional explanation is that bosses are trying to massage the share price. Going off the accounting piste allows them to put the most favourable spin on the numbers, slaloming round inconvenient non-cash or supposedly non-recurring items to focus on metrics they believe will support their stock. While that all may be true, it is not the only reason — maybe Gaap numbers themselves just aren’t much use.
That is one possible conclusion from research conducted by the economist Andrew Smithers. He compares corporate after-tax profits contained in US national accounts going back to the 1940s with the reported earnings of the S&P 500. The former are top-down economic data extracted from national statistics, the latter accounting figures those large companies churned out. The contrast is instructive. In the decades from 1947-57 to 1992-2002, the volatility of the two series is very similar, Mr Smithers observes. While current accounting rules remain unaltered, opportunities for non-Gaap bamboozlement will flourish unchecked.
Indeed, that of the S&P earnings data are slightly lower, perhaps reflecting some “smoothing” on the part of more cautious public company managers. But in more recent years, there has been a big divergence. After 1992-2002, the profits published by quoted companies have been five times more volatile than those in the national accounts. “For the volatile data to be an improvement, the relative stability of national profits data would need to be misleading,” says Mr Smithers, who regards that outcome as a remote possibility. “If they are not, corporate data now provide worse information than before.”
The uptick in volatility follows reforms to US accounting rules in the 1990s. These were designed to make company figures more “useful to users” and sprang from an intellectual movement that sought to apply financial valuation theory to accounting. To channel capital to the most productive outlet, the logic went, accounts needed to give traders of securities a clearer understanding of a company’s current worth. Out went boring old concepts such as verifiability and prudence. So-called fair values based on market prices and valuation models started to supplant historical cost numbers in the balance sheet. Banking assets held for trading started to be reassessed regularly and with unrealised gains recorded as profits. Contracts were sometimes valued as discounted streams of income, stretching seamlessly into the future.
Whatever these changes have led to, they have not obviously achieved the framers’ intention: to make accounts more “value relevant”. The use of fair values tends to accelerate the recognition of profits, introducing procyclicality because the prices themselves bake in estimates of anticipated earnings as well as what has actually been realised to date. Studies of US firms show both increasing volatility and the declining “persistence” of earnings — meaning they no longer follow a steady path. The declining value of the figures churned out is not lost on shareholders.
One consequence of this shift is a decline in the correlation of Gaap earnings changes in stock market valuations. The ironic consequence of all this is that investors increasingly rely on non-Gaap numbers for valuation. These are not only idiosyncratic, and thus not always capable of comparison, they are also devised by bosses whose views may well be richly coloured by their own outsize incentives. The ability of managers to present misleading information about their companies has increased.
The US stock market authorities frown on non-GAAP data and have attempted in recent years to devise more rules to govern their use. Instead they might profitably look at the real usefulness of the GAAP figures these are increasingly supplanting. “Accountants have no comparative advantage trying to guess at the current valuations of companies,” says Karthik Ramanna, professor of business and public policy at the University of Oxford’s Blavatnik School of Government. “Their job should be to establish some sort of verifiable lower bound.” That’s not what the current accounting rules are set up to deliver. While they remain unaltered, not only will their legitimacy dwindle, but opportunities for non-GAAP bamboozlement will flourish unchecked.