The Value of a Changed Opinion: Analyzing Stock Price Reactions to Analyst Recommendation Revisions

Christoffer D. Brammer & Victor F. Näsström

Student thesis: Master thesis


This thesis investigates the market reaction in stock prices to changes in analyst recommendations in Denmark and the United States. We examine a dataset of 102,417 analyst recommendation revisions for 1,403 stocks listed in Denmark and the United States for 2000 through 2019.
According to the market efficiency hypothesis, if analyst recommendations contain price-relevant information to the market, then prices should adjust immediately to incorporate the new information. We test whether there is any apparent drift or mean reversion after the announcement of recommendation revision as has been found around, e.g., earnings announcements and previous studies.
We find an immediate effect determined by both direction of change as well as the rating level of the recommendation revision. A positive (negative) recommendation revision, defined as an upgrade to “buy” (downgrade to “sell”), results in three-day buy-and-hold abnormal returns of 1.86% (-1.26%) in Denmark and 1.62% (-1.91%) in the US which are all significantly different from zero at the 1% significance level but is generally not significantly different between the two markets.
The results indicate an asymmetric reaction to positive and negative recommendation revisions in the U.S. market but not in Denmark, where upgrades to “buy” are associated with the most positive abnormal returns and a larger effect in absolute terms than downgrades to “sell.”
We observe those recommendation revisions that skip rating level(s) experience higher abnormal returns than revisions that do not skip levels. We also show other significant determinants of the reaction to recommendation revisions such as market capitalization, trading volume, broker, year, the potential of the stock, and the distance of the recommendation from the consensus.
Furthermore, we find that stock prices do not respond immediately to recommendation revisions as abnormal returns are significant for the day following the announcement, and different forms of drifts and mean reversions are indicated depending on the type of recommendation change.
Lastly, we show that an investment strategy with long (short) positions in positive (negative) recommendation revisions and daily trading leads to significant abnormal returns before accounting for transaction costs.

EducationsMSc in Applied Economics and Finance, (Graduate Programme) Final Thesis
Publication date2020
Number of pages94
SupervisorsPeter Feldhütter