Attention-Grabbing Stocks and Buy-Sell Asymmetry

Attention-Grabbing Stocks and Buy-Sell Asymmetry

Buy-sell asymmetry is the prediction, due to Barber and Odean ($2008$), that limited attention should affect purchase decisions much more strongly than sale decisions for individual investors. The asymmetry arises because the choice sets are fundamentally different on the two sides of the trade.

• A retail investor selling chooses among the small number of stocks she already owns — typically a handful. She does not need attention to know what is in her own portfolio.
• A retail investor buying must choose from the universe of public stocks — thousands of names. She cannot evaluate them all, so she purchases from the small subset that has recently grabbed her attention.

The asymmetry implies that attention-grabbing events should produce strong net buying pressure from retail investors, not just elevated turnover. Barber and Odean ($2008$) define attention-grabbing stocks operationally as those experiencing any of three triggers:

• Extreme prior-day returns (large positive or large negative).
• High abnormal trading volume.
• News announcements that day.

Using $1991$–$1996$ trading records from a large discount brokerage and a large retail brokerage, they confirm that retail investors are net buyers of stocks in all three categories, with buy-sell imbalances reaching $20$–$30$ percentage points in the most extreme deciles. Critically, the asymmetry holds for stocks with extreme negative prior-day returns as well: retail investors buy the biggest losers more than they sell them, because attention is what triggers the trade and bad news is just as attention-grabbing as good news.

Implication for prices. Net retail buying pressure on attention-grabbing stocks creates a temporary mispricing that should reverse — the attention-driven demand is not a fundamental signal. Empirically, attention-grabbing stocks underperform over the subsequent weeks, consistent with the Barber-Odean ($2000$, $2001$) result that retail investors lose money on average (see active fund underperformance for the related institutional analogue and the more general retail-trading-loss literature).

Intuition

The choice problem facing a retail investor is fundamentally different on the buy and sell sides. To sell, she must choose one of the (say) $10$ stocks she owns. The choice set is small enough that she can evaluate every option without much cognitive effort. To buy, she must choose one of the $5,000$ public stocks. The choice set is far too large to evaluate exhaustively, so she relies on a salience heuristic: she buys whatever recently grabbed her attention. This is not irrational on its own — using salience as a search algorithm is a reasonable response to a search-cost problem — but it produces predictable distortions when salience is uncorrelated with fundamental quality.

The cleanest way to test the asymmetry is to construct portfolios sorted on attention-grabbing triggers and compare the net buy-sell imbalance among retail traders. If attention drove both sides of the trade equally, one would see higher turnover in attention-grabbing stocks but no net imbalance. The actual finding — substantial net buying — implies that the buy decision is attention-constrained while the sell decision is not.

The Barber-Odean asymmetry is closely related to the disposition effect (the tendency to sell winners and hold losers), but it is conceptually distinct. The disposition effect is about which stocks among one’s portfolio one sells; buy-sell asymmetry is about the direction of net flow into attention-grabbing stocks across the entire universe. Both are individual-investor anomalies but they operate on different stages of the decision process.

Properties

(Three Attention Triggers) Extreme returns, high volume, news.

Barber and Odean ($2008$) operationalize “attention-grabbing” via three independent measures, each of which produces the same buy-sell asymmetry result. (i) Extreme prior-day returns: stocks with returns in the top or bottom decile of the previous day’s distribution. (ii) High abnormal volume: stocks with prior-day trading volume in the top decile relative to their own historical average. (iii) News-event flag: stocks with any news article in a major financial news index that day. The three triggers correlate but are not redundant; the asymmetry is robust to using each individually or in combination.

(Asymmetric Buy-Sell Imbalance) Imbalance reaches $20$–$30$ pp in extreme deciles.

The headline figure: in the top decile of attention-grabbing stocks, the percent buy-sell imbalance among individual investors at large discount brokerages reaches $25$–$30$ percentage points (where $0$ would indicate equal buying and selling). The bottom decile shows imbalances near zero. The middle deciles trace a U-shape: stocks with extreme returns in either direction (high or low) draw net buying, while stocks with average returns draw none. The shape is exactly what limited-attention asymmetry predicts.

(Retail-Specific) The pattern does not appear among institutional traders.

Barber and Odean confirm the asymmetry is a retail phenomenon. Institutional traders, who use specialized screening tools and are not attention-constrained on the buy side, show no comparable buy-sell imbalance in attention-grabbing stocks. The contrast strengthens the limited-attention interpretation: institutions face the same news flow but a different search technology for new investment ideas, and the imbalance disappears when search is not attention-constrained.

(Bad-News Symmetry) Retail investors buy big losers as much as big winners.

A common misconception is that attention-grabbing trade is driven by extrapolation — investors chasing winners. The data reject this: retail investors are net buyers of both extreme-positive-return stocks and extreme-negative-return stocks. The trigger is salience, not direction. This is what distinguishes the attention story from a momentum-extrapolation story (see 8.6.7. Extrapolative Beliefs) and makes the case that attention itself, not directional belief updating, is the binding cognitive constraint.

(Subsequent Underperformance) Attention-driven purchases reverse on average.

Retail investors who buy attention-grabbing stocks subsequently underperform the market. Barber-Odean ($2000$) document that high-turnover individual accounts underperform low-turnover ones by roughly $7\%$ per year, and the attention-grabbing-buy mechanism is one channel for this underperformance. The implication is that the salience-based search algorithm is not a fortunate heuristic — it systematically pushes retail investors into temporarily-overvalued names that subsequently revert.

(Modern Application) Robinhood-era trading data confirm the pattern at greater scale.

Barber, Huang, Odean, and Schwarz ($2022$) extend the original $1991$–$1996$ analysis to Robinhood data from $2018$–$2020$. They find an even more pronounced version of the asymmetry: the top $10$ stocks on the Robinhood “most-held” list show extreme net buying after attention-grabbing events, and these stocks subsequently underperform by roughly $5\%$ over the next month. The platform’s user-interface design — a “top movers” list prominently displayed on the home screen — amplifies the attention asymmetry by mechanically directing users to extreme-return stocks. This is also a clean illustration of the prescriptive flip-side: choice architecture matters.

Examples

Attention-grabbing GameStop in January 2021

On the morning of January $27$, $2021$, GameStop ($\text{GME}$) registered all three attention triggers simultaneously: an extreme prior-day return ($+135\%$), volume that was orders of magnitude above its historical average, and saturation news coverage. Retail-broker order flow from Robinhood, E*Trade, and TD Ameritrade was overwhelmingly buy-side that morning, with buy-sell imbalances reportedly above $80\%$ on Robinhood. The Barber-Odean framework predicts exactly this: salience-driven net buying from retail traders, decoupled from fundamental valuation. The subsequent reversal — $\text{GME}$ peaked at \483$andtradedbelow$$100$ within two weeks — is consistent with the post-attention reversal pattern.

Why doesn't the disposition effect generate the same asymmetry?

The disposition effect is the tendency to sell winners and hold losers once owned. It operates entirely within the existing portfolio — it influences which of the investor’s current holdings she sells, not whether she sells more often than she buys. Buy-sell asymmetry, by contrast, is about net flows into stocks the investor does not yet own. The two patterns are independent: an investor could exhibit strong disposition behavior (selling winners early) while also showing strong buy-sell asymmetry on attention-grabbing stocks (buying any salient name, regardless of her existing portfolio). Barber-Odean ($2008$) explicitly verify this independence by controlling for whether the buyer already owned the stock.

The buy-side / sell-side cost of inattention

Suppose a retail investor sees that a small biotech firm announced strong Phase II trial results yesterday. The stock jumped $40\%$. The investor buys $100$ shares at the open. By the Barber-Odean evidence, this investor is one of many making the same trade for the same attention-driven reason; the order flow imbalance pushes the price further temporarily. Three weeks later, when professional and institutional analysis catches up and the price re-equilibrates around the new fundamental — say, $20\%$ above pre-announcement levels rather than $40\%$ — the retail investor has lost roughly $14\%$ on her position. The mistake was not in believing the trial result was good news; it was in being part of the attention-driven crowd that systematically overpays at the moment of peak salience.

See Also

• 8.11. Bounded Rationality — parent hub.
• 8.11.1. Limited Attention and PEAD — sibling: attention failure on the announcement-incorporation side.
• 8.10.3. Disposition Effect — different individual-investor anomaly that operates on the sell side; conceptually independent.
• 8.10.7. Active Fund Underperformance — companion result on retail and institutional trading underperformance.
• 8.6.7. Extrapolative Beliefs — alternative explanation of attention-related buying that the bad-news symmetry helps to reject.
• 8.12. Prescriptive Behavioral Finance — choice-architecture implications of the attention asymmetry.