A momentum screen that looks for recent analyst recommendation upgrades, especially those "bold revisions" that move away from the broker consensus.
Although investing on the basis of broker recommendation alone does not appear to be a successful strategy, research suggests that focusing on recent changes in broker recommendations is more fruitful, particularly in combination with other signals. Looking at the relative level of recommendations, rather than the absolute level, may help to eliminate the upward bias. It also ensures that the information is fresh, unlike recommendations which can be up to 6 months’ old.
Researchers have distinguished between bold revisions, i.e. ones that move away from the consensus and which may provide new information vs. herding revisions that merely move toward the consensus. The results indicate that “bold” revisions have a greater impact. It also seems that the most influential revisions tend to be "revisions issued contemporaneously with earnings forecasts, analysts with greater relative experience, and those with more accurate earnings estimates”. Growth firms, small firms, and high turnover firms are also apparently more likely to have influential stock recommendations.
It’s unclear why this post-recommendation revision drift phenomenon exists. Investor inattention is probably the most plausible explanation given that it appears greatest for stocks with smaller market capitalizations and with lowest analyst following. It may be that recommendation changes captures new qualitative information that it takes time for the market to assimilate (e.g., managerial changes, strategic alliances, or other growth opportunities). An alternative theory is that the recommendations themselves cause the subsequent price effect through the publicity surrounding them.
Definition of a Recommendation Revision Screen
Based on various research papers, some indicative criteria for a Recommendation Revision Screen might be:
- At least one net revision in the recommendation over the last month (or preferably the last week). Another option might be, say, a 5% or greater change in rating over past 4 weeks.
- Market cap in the bottom half of the database (Jegadeesh found the strategy was more profitable for smaller firms) but above £15m to avoid “the long tail” of illiquid stocks
Possible additional filters:
- Relatively low analyst coverage (there tends to be greater investor inattention and information dissemination may take longer).
- Revision moves away from the consensus – a “bold revision”. This would involve specifying however that at least two or more analysts cover this year’s forecast in order for there to be a “consensus”.
- Recommendation upgrade must skip a rank, i.e. from Hold to Strong Buy.
Does it Work?
Overall, research suggests that the profits from a pure recommendation change trading strategy exist but are probably lower than transaction costs, especially if it’s not possible to trade on the day the recommendation is made. Research by Jegadeesh across many markets, including the UK, found that stock prices do react significantly to recommendation revisions. The bulk of this tends to be on the day of recommendation and the one following that but stock prices continue to drift up over the next six months. The price reaction appears to be strongest in the US market – a recent study found that the average return on the revision date is 2.03% for all upgrades (increasing to 4.85% in six months). A UK specific study by Ryan and Taffler looking at the London Stock Exchange found that share prices are significantly influenced by analysts’ recommendation changes but apparently concluded that the effect of “buys” is rather weak and only new “sells” and recommendation revisions for smaller, less followed stocks have value for investors.
However, perhaps this might not be the case using some of the additional filters outlined above. Furthermore, it seems possible that investors may profitably use recommendation revisions in combination with other signals. JP Morgan’s analysis found that its value was best captured when changes in analysts Buy/Sell Recommendations act as a confirmation or filter of other signals. Specifically, they tested a sector relative one year forward PE ratio, and found that it was made significantly stronger when augmented with the three month change in Mean Analyst Buy/Sell Recommendations.
Interestingly, in their paper, “Analysts, Industries, and Price Momentum”, Boni & Womack looked at a long-short strategy buying analyst upgrades and short analyst downgrades within each industry. They were trying to dilute or remove the momentum associated with changing industry to isolate the pure analyst upgrade and downgrade effect. This industry- neutral portfolio yielded 1.23% in the next calendar month, or about 15% annualized and its return-to-risk ratio was up almost 100% relative to portfolios in which industry weighting is ignored. However, the strategy is likely to be a difficult strategy for individual investors to implement, given its long/short nature and the number of trades involved.
How can I run this Screen?
For further info, it’s worth reading the 2004 paper by Jegadeesh - Analyzing the Analysts:When do Recommendations Add Value? as well as the subsequent analysis across international markets - Value of analyst recommendations: International evidence. See also:
- JP Morgan: The Search for Alpha
- Investor Inattention and the Underreaction to Stock Recommendations
- Analyst Forecast Revisions and Market Price Discovery
- Do Brokerage Houses Add Value? The Market Impact of UK Sell-Side Analyst Recommendation Changes
- Do Analysts Herd? An Analysis of Recommendations and Market Reactions
- When Are Analyst Recommendation Changes Influential?
- 12 Stocks Upgraded by Analysts That Hit Their Targets