What Earnings Figures Should I Use? by David Merkel, CFA of Aleph Blog
One of the challenges in investing is understanding whether stocks are really earning above expectations or not. Many companies, especially the large ones have their adjusted or modified GAAP earnings. I can sympathize with the companies if the adjustments make the adjusted earnings more like free cash flow. But if not, I disagree.
GAAP accounting is very good, better than IFRS, but not perfect. There are two uses for earnings figures, and they are different, because they serve different time periods.
Clint Carlson's Carlson Capital Double Black Diamond fund returned 3.34% in August net of fees. Following this performance, the fund is up 8.82% year-to-date net, according to a copy of the firm's August investor update, which ValueWalk has been able to review. On a gross basis, the Double Black Diamond fund added 4.55% in August Read More
When analyzing quarterly earnings, heed the adjusted earnings figure, but review that calculation to see that is is on the same basis as it was over the last year. Shenanigans are often played here.
Second, look over a 3-, 5-, and 10-year periods: see how much actual earnings lagged behind management estimates. Yes, there are temporary fluctuations in earnings. But when they repeat again and again, it indicates that management is sloppy. This is structural and not sporadic.
Don’t invest in companies that regularly have significant one-time disappointments. What is regular should be treated as regular. Companies that regularly have to adjust earnings higher then GAAP deserve lower valuations.
Here’s another way of thinking about it: writedowns usually reflect the past, indicating that past profitability wasn’t so high. If writedowns come frequently, it means that management has made bad/liberal accounting decisions. This could be a stock to avoid.
Thus in the short run, look at adjusted earnings, but critically. In the long run, look at unadjusted earnings, because managements should be responsible for their long-term errors.