In considering the stock ratings of expert analysts, should investors focus more on the level of the ratings or changes in ratings? In their December 2007 paper entitled “Ratings Changes, Ratings Levels, and the Predictive Value of Analysts’ Recommendations”, Brad Barber, Reuven Lehavy and Brett Trueman investigate the potential value to investors of both levels of (strong buy, buy, hold, sell, strong sell) and changes in analyst stock ratings. Using real-time analyst stock ratings from two databases spanning 1986-2006 (more than 1,000,000 ratings) and contemporaneous daily stock returns, they conclude that:
- Buy/strong buy rating announcements outnumber sell/strong sell announcements by a factor of about 5.5. Upgrades and downgrades each comprise roughly one fifth of all rating announcements.
- Average daily abnormal returns generally decrease with rating, ranging from +1.0 basis point for strong buys -2.4 basis points for strong sells.
- Upgrades (downgrades) are associated with an average daily abnormal return of +1.9 (-1.0) basis points. Returns do not increase with upgrade magnitude (number of levels), but they do decrease with downgrade magnitude.
- For a given type (upgrade or downgrade) and magnitude (number of levels) of a ratings change, buy and strong buy ratings outperform hold, sell and strong sell.
- For a given rating level, upgrades outperform downgrades. For example, upgrades to buy beat downgrades to buy by an average 2.7 basis points per day.
- Both rating levels and changes predict future unexpected earnings, indicating that analysts have some ability to discover valuable private information about future earnings. Their rating actions do not simply elevate or depress investor demand for specific stocks.
The following chart, taken from the paper, plots the cumulative values of $1 initial investments in three hedge portfolios over the period 1986-2006. The Levels-based Portfolio is long (short) all stocks rated buy or strong buy (sell or strong sell). The Changes-based Portfolio is long (short) all upgraded (downgraded) stocks. The Levels- and Changes-based Portfolio is long (short) stocks receiving a double upgrade to buy or strong buy (a double downgrade to sell or strong sell). The investments in the first two portfolios each grow to just over $7, while that in the third grows to over $24. Results do not incorporate trading costs/frictions, which are likely significant.

In summary, investors may find edges by considering both the levels of and changes in analyst stock ratings, with the combination more powerful than the separate indications.
It would be useful to know whether these indications have weakened in recent years, especially since the information-leveling Regulation FD (Fair Disclosure) effective October 2000.
Frequently Asked Questions
What is an analyst stock rating and how do rating levels differ from rating changes?
An analyst stock rating is an expert's recommendation on a stock, typically strong buy, buy, hold, sell, or strong sell. The rating level is the current recommendation itself, while a rating change is an upgrade or downgrade from a prior level. In the 1986-2006 study of more than 1,000,000 ratings, upgrades and downgrades each made up roughly one fifth of all rating announcements.
Do analyst ratings and rating changes predict stock returns?
Analyst ratings and rating changes both predict stock returns, and combining them works best. In the 1986-2006 data, average daily abnormal returns ranged from +1.0 basis point for strong buys to -2.4 basis points for strong sells, and upgrades averaged +1.9 basis points versus -1.0 for downgrades. A hedge portfolio combining levels and changes grew $1 to over $24, versus just over $7 for levels-only or changes-only portfolios.
How are analyst rating strategies tested using hedge portfolios?
Analyst rating strategies were tested by pairing real-time ratings from two databases (1986-2006, over 1,000,000 ratings) with contemporaneous daily stock returns. Three hedge portfolios were built: a levels-based one long buy/strong buy and short sell/strong sell stocks, a changes-based one long upgraded and short downgraded stocks, and a combined one long double upgrades to buy/strong buy and short double downgrades to sell/strong sell. Performance was tracked as the cumulative value of $1 invested.
Is an upgrade to a buy rating better than a downgrade to a buy rating?
An upgrade to buy outperforms a downgrade to buy, by an average of 2.7 basis points per day in the 1986-2006 study. More generally, for a given rating level upgrades outperform downgrades, and for a given type and magnitude of change, buy and strong buy ratings outperform hold, sell and strong sell. Returns do not rise with upgrade magnitude, but they do fall with downgrade magnitude.
What are the limitations of trading on analyst rating levels and changes?
The reported analyst rating strategy results do not include trading costs or frictions, which are likely significant and could erode the returns. The study covers 1986-2006, so it is unclear whether the predictive power has weakened in recent years, particularly after Regulation FD took effect in October 2000. The findings suggest analysts discover private earnings information rather than merely shifting investor demand, but this does not guarantee net profits.