< Return to All Blogs

Brief Buys…Slow Sells

In this article, Cameron Hight summarizes a paper titled “Selling Fast and Buying Slow” on return contribution of buy and sell decisions on over 4 million trades across 700+ portfolios.

Sean Stannard-Stockton, who runs Ensemble Capital, sent us a link to a recent paper titled “Selling Fast and Buying Slow” on return contribution of buy and sell decisions on over 4 million trades across 700+ portfolios. The result is that the portfolio managers in this study were great buyers and lousy sellers. On top of that, they compounded the problem by slowly buying, when they would have made more money by quickly buying, and quickly selling when they would have lost less money if they would have slowly sold.

Alpha Theory does this by design. It starts (buys) positions quickly and sells them slowly. We’ve measured a 4% difference between our clients’ ROIC of 6% and the 10% ROIC if they’d have traded according to Alpha Theory’s recommendations. We further break that down into initial position sizing and trading, where we find that about half, or 2%, comes from better trading. The “Selling Fast and Buying Slow” paper finds 170bps from better buying and selling decisions, which is close to the trading difference we’ve measured for our clients*.

Excerpt from “Selling Fast and Buying Slow” - Is there skill in buying and selling?

We examine this question using a unique data set containing the daily holdings and trades of sophisticated market experts—experienced institutional portfolio managers (PMs). Our data is comprised of 783 portfolios, with an average portfolio valued at approximately $573 million. More than 89 million fund-security-trading dates and 4.4 million high-stakes trades (2.0 and 2.4 million sells and buys, respectively) are observed between 2000 and 2016.

While the investors display clear skill in buying, their selling decisions underperform substantially. Positions added to the portfolio outperform both the benchmark and a strategy which randomly buys more shares of assets already held in the portfolio by over 100 basis points per year. In contrast, selling decisions not only fail to beat a no-skill random selling strategy, they consistently underperform it by substantial amounts. In our preferred specification, PMs forgo 70 basis points per year in raw returns.

Why would a majority of portfolio managers appear to exhibit skill in buying while at the same time underperforming substantially in selling? At face value, the fundamentals of buying and selling to optimize portfolio performance are similar: Both require incorporating information to forecast the distribution of future returns of an asset. Skill in both decisions requires the investor to look for relevant information and integrate it into the forecast. However, there is a reason to suspect that selling and buying decisions involve different psychological processes (Barber and Odean 2013). Recent work from the lab is consistent with this discrepancy: Buying decisions appear to be more forward-looking and belief-driven than selling decisions in an experimental asset market (Grosshans, Langnickel, and Zeisberger 2018). And indeed, anecdotal evidence from our sample points to PMs thinking differently about the two decisions; extensive interviews suggest that they appear to focus primarily on finding the next great idea to add to their portfolio and view selling largely as a way to raise cash for purchases.

We utilize a unique dataset and find evidence that financial market experts—institutional investors managing portfolios averaging $573 million—display costly, systematic biases. A striking finding emerges: While investors display skill in buying, their selling decisions underperform substantially—even relative to random sell strategies. We provide evidence that investors use heuristics when selling but not when buying, and that these heuristic strategies are empirically linked to the documented difference in performance.

As shown in Section 4, the comparison of trades on earnings announcement versus nonannouncement days suggests that PMs do not lack fundamental skills in selling; rather, results are consistent with PMs devoting more cognitive resources to buying than selling. When decision-relevant information is salient and readily available—as it is on announcement days—PMs’ selling performance improves substantially. We propose a mechanism through which overall underperformance in selling can be explained by a heuristic two-stage selling process, where PMs limit their consideration set to assets with salient characteristics (extreme prior returns) and sell those they are least attached to (low active share assets). A proxy for this heuristic strategy is associated with substantial losses relative to a no-skill random selling strategy.

The question remains of why professional PMs have not learned that their selling decisions are underperforming simple no-skill strategies. While we can only speculate, the environment in which fund managers make decisions offers several clues. As Hogarth (2001) notes, the development of expertise requires frequent and consistent feedback. While it is feasible to generate this type of feedback for both buy and sell decisions, anecdotal evidence from our interviews with PMs suggests that decisions are overwhelmingly focused on one domain over the other. In terms of time allocations, our understanding is that the vast majority of the investors’ research resources are devoted to finding the next winner to add to the portfolio. Moreover, standard reporting practices are well-suited for evaluating the performance of buying decisions: Purchased assets are tracked, providing salient and frequent feedback on the outcomes of buying decisions. This process appears successful in producing expertise—purchased assets consistently outperform the benchmark. In comparison, paltry resources are devoted to decisions of what to sell, and the relevant feedback is largely lacking: Assets sold are rarely if ever, tracked to quantify returns relative to potential alternatives such as our random sell counterfactual.

A recent paper by our friends at Essentia Analytics titled “The Alpha Lifecycle” confirms this conclusion with a different data set and a different approach.

Essentia is wrapping up a 5-month analysis of this phenomenon (Alpha Lifecycle), involving data from 42 portfolios over more than 10 years. The conclusions are clear: alpha has a beginning, a middle and an end. It tends to decay over time, reducing — or even reversing — the benefits it offered early on. Active managers who wish to deliver sustained alpha in their portfolios need to understand their own alpha lifecycles and adjust their investment decision-making processes accordingly.

Dominant Lifecycle from Essentia Analytics

Alpha has a lifecycle and tends to decay over time — frequently causing managers who fall in love with their stocks to suffer. On average, managers we analyzed experienced a 400 basis point peak-to-trough decay in return on each position.

These papers are great examples of the value of feedback for firms that want to improve. Capturing data, analyzing that data, and changing behavior based on empirical evidence is akin to Tiger Woods hitting golf balls while hooked up to a TrackMan and adjusting his swing to maximize the attributes that give him the greatest chance of success. The managers of the future are already adopting regimens that treat their process like that of an elite athlete. Those that don’t will get left behind.

*** We recognize this is not a perfect apples-to-apples comparison as our analysis measures the trading value as the difference between the Alpha Theory return and the return of keeping a consistent position size from beginning to end of the holding period.

Clare Flynn Levy
Essentia Analytics
Institutional Investor
Portfolio Strategy
Portfolio Optimization
Risk Management