Tucked inside this recent post joining the debate on who is the bigger contrarian, is an assertion that Wall Street is tricking us into thinking the market is attractive when it is really almost 30% overvalued. Barry Ritholtz invokes a big name, Cliff Asness, and states that he has researched the forward earnings comparison. This is not what Asness and his colleagues did, as they make clear in a footnote. The problem is that most people's eyes glaze over while reading complex research like this, despite some entertaining zingers from Asness.
Most people are not going to read the paper, so they take it on faith that people like Barry (and Mark Hulbert in his piece) are getting it right.
There are three things wrong with this interpretation of the Asness study:
- He did not look at forward earnings from 1871 onward because the data do not exist. One would expect people to know this, even without reading the footnotes! Looking at the time period where forward earnings data are available gives a dramatically different conclusion -- but then it is a more bullish era.
- We should not care about what happened to the market or what the relationships were in 1871. Or 1926. Or 1956. Or maybe even 1976. Start with the question of whether any market relationships before the era of derivatives should still be expected to hold true. The desire to use data that are not relevant, just because you have it, is one of the big problems in forecasting.
- Most importantly, prediction is, after all, about looking forward. It seems obvious, but many forecasters miss the point. Using trailing earnings is especially bad at a time when there is rapid growth (as we have seen for the last three years) or a sudden decline.
Those interested in this sensible notion might want to read my paper, "Should Experts Look Forward to Predict?"