Why Everyone Gets Contrarian Investing Wrong

I ran into this problem about three years ago when a client wanted to run a Dreman Contrarian Investment Strategies screen on a universe of 4,000 mid-cap stocks. The raw data came back with 800 candidates. Not a typo. The issue was that Dreman's original criteria — low P/E, high dividend yield, low price-to-sales, stable earnings growth — are each individually popular. When you stack them together as he prescribed in his 1998 work and subsequent updates, you still get an enormous pool because the metrics don't move in lockstep the way people assume they do. A stock can have a low P/E and a low dividend yield simultaneously if it's a growth stock that just happened to get crushed on earnings. The filter didn't catch it. My workaround was to add a momentum filter that Dreman never really emphasized. I screened for stocks that were contrarian by fundamentals but had stopped declining on price over the prior 60 days. That cut the universe from 800 down to about 47. The logic is simple even if it isn't in his original text: a low P/E stock that is still falling fast has a reason. It might be deteriorating. Waiting for the price to stabilize before buying the value gave us better entries without sacrificing the contrarian premise.

Building a Dreman Contrarian Investment Strategies Screen

Jeremy Dreman's framework is fundamentally about exploiting investor psychology through fundamental screening. The core idea is that markets overreact to bad news, pushing stocks to valuations that don't reflect their actual earning power. You buy when everyone else is selling. You sell when the valuation gets ridiculous again. It sounds straightforward until you actually run it. Here is how I build the screen, step by step, using data from Finnhub: First, define your universe. Dreman worked with large-cap and mid-cap stocks primarily. If you start with small caps, the liquidity issues alone will destroy your returns. Pick stocks with at least 500,000 average daily volume and a market cap above $2 billion. This eliminates the junk that occasionally looks cheap for very good reasons. Second, run the four fundamental screens. The first is the P/E screen. Look for stocks trading below 10 times trailing twelve-month earnings. The second is the dividend yield screen. Target yields above 3%. The third is the price-to-sales ratio screen. Under 1 is the traditional threshold, but I find under 1.5 works better in the current environment where revenue multiples have expanded across the board. The fourth is the earnings stability screen. Require positive earnings growth in at least three of the last five years. Dreman originally used a revision component here — stocks where analysts haven't been slashing estimates — but I handle that separately in my workflow. Third, combine the screens. The key insight most people miss is that you should not require all four filters simultaneously at the start. Each filter independently identifies mispricing. Running all four together is aggressive and leaves you with maybe two dozen stocks in a bull market, sometimes fewer. I prefer to rank the universe by a composite score where each metric gets equal weight, then take the bottom quartile on each metric. This gives you a stronger signal without eliminating stocks that fail on one dimension for defensible reasons. Fourth, check the revision component. This is where Dreman's approach gets interesting and where most retail investors skip it. Look at whether analyst earnings estimates have been revised upward or downward over the past 90 days. Contrarian success depends partly on the gap between market perception and reality. If estimates are still falling, the stock may have more pain ahead. If estimates have stabilized or started creeping up while the price remains depressed, that is your entry zone.

The Mechanics Actually Matter More Than the Philosophy

Dreman's original research showed annualized returns in the high teens over long periods using this approach. Those numbers sound good until you realize they assume you hold for years and you can actually execute at the prices the screen predicts. In practice, the gap between what a screener shows and what you can buy is where the strategy breaks down for most people. I found this out the hard way with a name that screened perfectly in early 2021. Everything checked out. P/E around 7. Dividend yield near 5%. Price-to-sales under 0.8. Earnings had been flat but positive for four straight years. The stock was down 60% year to date on what looked like a sector rotation away from energy. I bought 12% of the portfolio. The stock went down another 35% over the next six months. The problem was not the fundamentals. The problem was that the company was facing a structural decline in its primary market that the aggregate earnings data had not yet reflected. Three years of flat earnings masked a fourth year that was going to be a disaster. The fix was to add a margin of safety that Dreman barely discussed. After the initial screen, I now require that the stock has not dropped more than 40% year to date unless there is a clear macroeconomic trigger for the selloff. If the stock is crashing and the fundamentals look cheap, I wait. The screen will still be there in two months. Usually it will be cheaper. Another thing nobody tells you about running these screens is the rebalancing frequency. Dreman suggested holding for a year or more until the mispricing corrects. But holding forever is also wrong because some stocks genuinely deserve their low valuations. The market is not always irrational. Sometimes a 7 P/E is a 7 P/E for a reason. I set a maximum holding period of 36 months for any position that comes out of the screen. If the thesis hasn't played out by then, I sell regardless of valuation. This forces discipline. It also means you will occasionally sell before a stock rallies, but that is preferable to holding a value trap for five years.

Running the Screen with Real Tools

You do not need a Bloomberg terminal for this. I use Finnhub's stock screener and supplement it with their financials endpoint for the revision data. The screener lets you filter by P/E, dividend yield, price-to-sales, and earnings growth simultaneously. Export the results to CSV. Then pull the analyst revision data manually for the top candidates. It takes about 15 minutes per batch if you have a shortlist of 50 or so stocks. The Finnhub endpoint for earnings revisions is not as clean as you would want. You have to cross-reference consensus estimate changes from the fundamentals endpoint and compute the direction yourself. The API documentation is adequate but sparse on this particular feature. I built a small Python script that pulls the data and scores each stock on a 1-to-10 contrarian strength scale. The script takes about 20 seconds to run on a reasonable universe. Here is a practical example of what the output looks like in my workflow: I screen for stocks with P/E under 10, dividend yield above 3%, price-to-sales under 1.5, and positive earnings in at least 3 of 5 years. That returns roughly 300 names from a universe of 3,000 US-listed stocks. I then rank those 300 by analyst revision trend over the past 90 days. Stocks with upward revisions get a +1, neutral get a 0, downward get a -1. I drop anything with a score below -2. That leaves me with about 60 candidates. I apply the price decline filter. I drop stocks that are down more than 40% year to date without a sector-wide catalyst. That usually leaves 15 to 25 names. I do a quick qualitative check on each one — reading the latest earnings call transcript, checking for pending litigation, verifying that the dividend is actually covered by free cash flow, not just accounting. Then I allocate.

Where the Strategy Actually Fails

I want to be blunt about the limitations because Dreman's books tend to present the successes without much discussion of the failure cases. The strategy fails hardest in secular decline industries. A company in structural decline — think physical retail, legacy media, certain commodity sectors during a multiyear trough — will screen as a contrarian bargain repeatedly while the business genuinely deteriorates. The P/E stays low because earnings collapse. The dividend yield stays high because the price collapses faster than the payout. The stock is not mispriced. It is correctly priced for a dying business. I avoid entire sectors that show sustained structural decline based on revenue trends, not just valuation. The strategy also struggles in rising rate environments. High dividend yield stocks get crushed when rates climb because income-seeking investors rotate into bonds. The contrarian signal becomes noise because the fundamental pressure from rates is real and ongoing. During the 2022 rate hike cycle, roughly 60% of my contrarian screen candidates underperformed the market simply because they were dividend-heavy and rate-sensitive. The screening data did not capture macro conditions. A third failure mode is timing. Even when the stock is genuinely mispriced, it can stay mispriced for years. Dreman acknowledges this but downplays it. In practice, buying a cheap stock and watching it stay cheap while you lose opportunity cost is painful. I manage this by capping any single position at 5% of the portfolio and keeping at least 30% in non-screened positions that are generating alpha through other means. This way the contrarian screen is a satellite strategy, not the whole thing.

A Practical Entry Framework

When I do buy from the screen, I do not go all in at once. I split the allocation into three tranches. The first tranche is 40% of the intended position size. I buy immediately if the stock passes all filters including the qualitative check. The second tranche is another 30%. I buy it if the stock drops an additional 10% from my entry price within 60 days. This assumes the thesis is still intact and the drop is market-driven rather than company-specific. The third tranche is the final 30%. I buy it if the stock has either recovered and retested my entry level as support, or if the dividend has been reaffirmed and the next earnings report shows stabilizing or improving fundamentals. This tranching approach adds about a week to the execution timeline but significantly reduces the chance of catching a falling knife. I have lost money on contrarian screens before, but I have lost less money since I started tranching. That is the practical benefit, not some theoretical advantage.

What to Watch Instead of the Screen

Once you have positions from the screen, the monitoring changes. You are not watching the P/E anymore. The P/E is your entry signal, not your ongoing indicator. What matters now is whether the fundamentals are moving in the right direction. I track three things weekly for each screen-generated position: free cash flow generation relative to the prior quarter, gross margin trend over the last four quarters, and any changes in debt covenants or credit ratings. If all three are stable or improving, I hold. If one is deteriorating, I investigate. If two are deteriorating, I reduce the position by half. If all three are deteriorating, I sell. This is faster than waiting for the next earnings report and it catches problems before they show up in the P/E. The downside is that this system requires actual attention. It is not a set-it-and-forget-it strategy. I spend about 30 minutes per week reviewing the holdings across all screens combined. That includes both contrarian and non-contrarian positions. The time investment is manageable but nonzero. If you cannot commit that time, the strategy will underperform because you will hold losers too long and sell winners too early out of anxiety.

The Bottom Line

Dreman Contrarian Investment Strategies works when you treat it as a systematic framework rather than a magic formula. The screens identify opportunities. Your job is to separate genuine mispricing from genuine deterioration. The revision data and the tranching approach are the practical tools that make the difference between theoretical returns and actual returns. The strategy has real limitations in declining industries and rising rate environments, and it requires ongoing monitoring. But for investors willing to do the work, it remains one of the more reliable approaches to generating excess returns from market overreaction.