Dreman's contention is that investors tend to overreact both to the positive and the negative (depending on the relative outlook towards a company or industry). He cites numerous bubbles to make his case, and also points out that the opposite of bubbles also tends to occur (situations where prices are well below where they should be).
Rather than rely on anecdotal data to make his case, however, Dreman describes a study he conducted with he believes proves that investors overreact. Once again he separated stocks into quintiles based on how cheap they were (this time by their Price to Book values). He then found that the operating performance (based on 5 factors including ROE and sales growth) over the subsequent five years of the lowest quintile group was much weaker than that of the highest quintile group. Despite this, the prices of the lowest quintile group showed remarkable gains both on an absolute basis and in comparison to the highest quintile group. As such, Dreman makes the point that while the market is able to determine which set of companies will have better operating metrics, it overvalues these companies and undervalues those which are expected to perform poorly.
Dreman ends by discussing how investors can profit off such overreactions in ways that go beyond simply following a low P/E (or low P/B or low P/CF or high dividend-yield) strategy. For one thing, junk bonds tend to show abnormally high returns over time, as default rates are over-estimated. However, Dreman suggests that investors deciding to enter this arena know what they are doing as rigorous financial analysis is required. On the other hand, Dreman argues an investor with little knowledge can likely do well by investing in junk bond mutual funds.
Earnings surprises are also a good source for finding overreactions from which the investor can profit. Often, a company missing estimates by a few million dollars (pennies per share) can see its market cap cut by billions of dollars. As traders overreact, such situations are ripe for finding bargains. Studies Dreman points to have also shown that stocks that drop the most in one period are more likely than other stocks to rise in the subsequent period, suggesting a strategy of buying beaten up stocks can also be successful.