We can never trust them completely. That's why we have a margin of safety.
Firstly, if two accountants prepare the figures they will get different results. Secondly, we have to ask ourselves what is the purpose of the company when preparing their financial statements. I think the most realistic view is that it is to make the company as good as possible. It is not to educate the general population. It is like a courtroom. The job of the lawyer for the defendant is to make him/her look as good as possible. They will deliberately promote the positive side and avoid the negative side.
What we do from our side is to put a substantial margin of safety.
Could investors ever trust earnings figures? Again the answer is no. Back in 1934, Benjamin Graham wrote in "Security Analysis", "But these earnings per share, on which the entire edifice of value has come to be built, are not only highly fluctuating, but are subject also in extraordinary degree to arbitrary determination and manipulation."
The point is that nothing has changed. We always have to be wary of earnings figures.
Some commentators have asserted that it is worse now. Perhaps, I suspect that if Benjamin Graham were alive now, he would say the complete opposite.
Another point is that doomsday articles have their fashionable topics. Criticizing earnings figures are the current fashion. So analysts and writers try very hard to come up with arguments about how bad is it now.
In "Beating the Street", Peter Lynch talks about a panel he was in each year. And every year there was a different doom and gloom story: killer bear market, world-wide depression, trade deficit, unemployment and the budget deficit. Pessimism sells newspapers and analysts reports.
So I think the question should be not whether we can trust EPS figures. Rather, are there companies for which we can have more trust in the figures? Some of the companies we avoid are those that change their accounting methods too often, that continually make earnings forecasts, that are always in the press as "the company of the moment", that have any hint of rule bending about them, that are actively involved in takeovers, that write unintelligible footnotes in their annual report, and that have products and services that we have no idea about. Also it is best to avoid companies whose growth seems too good to be true because it probably is.
If we do this, and if we add a healthy margin of safety, and if we avoid companies with earnings, sales, debt levels, etc that vary significantly, then we are reducing to a low level, the possibility of being involved with a company that has earnings figures that are dubious to a significant degree.
If I might end with a quote from Peter Lynch, "The key to making money in stocks is not to get scared out of them."
Regarding the EPS figures we use, they are supplied to us by MGFS, the same company that supplies data to Microsoft Investor, Hoover's, Quicken, SunGuard and other major financial organizations
