Research Firm: Thomson Reuters StarMine
About the Firm
Thomson Reuters StarMine is focused primarily on building quantitative factor models for institutional investors. Thomson Reuters StarMine's equity analytics and research management tools help investment firms around the globe generate alpha and process equity information more efficiently. They are one of the largest and most trusted sources of objective equity research performance scores. Their performance scoring helps investors anticipate trends in analyst sentiment, anticipate surprises, evaluate financial statements for measures of earnings quality, and more.
Thomson Reuters StarMine's sophisticated scoring system facilitates fair comparison of firm recommendation performance across widely disparate industries and market conditions. Thomson Reuters StarMine uses recommendations from research providers and the past performance of their recommendations at the sector level to create a quantitative metric, the Thomson Reuters StarMine Relative Accuracy Score, for a given stock.
The Equity Summary Score provides a consolidated view of the ratings from a number of independent research providers on Fidelity.com. Historically, the maximum number of providers has been between 10 and 12. However, some stocks are not rated by all research providers. Since the model uses a number of ratings to arrive at an Equity Summary Score, only stocks that have four or more firms rating them have an Equity Summary Score. The Thomson Reuters StarMine model underlying the Equity Summary Score has a few key components:
Normalize – Look at the research providers' buy and sell ratings distributions to understand which ratings are scarce and therefore more important.
The distribution of ratings from each of the independent research firms are normalized to make them more comparable with each other. For example, some research providers may issue a large number of buy recommendations and few sell recommendations, and vice versa. Thomson Reuters StarMine adjusts for this by overweighting "scarce" ratings and underweighting "plentiful" ratings. By normalizing the distribution of ratings, the model can recognize the "scarcity value" of ratings that are infrequently given which adds additional information to the model.
Weight – Look at the 24 month relative firm/sector ratings accuracy and use that information to determine which firms' ratings have the most weight in the aggregated Equity Summary Score.
For over five years, provided on Fidelity.com, Thomson Reuters StarMine has run its sophisticated scoring system to facilitate a fair comparison of research provider recommendation performance across widely disparate industries and market conditions. The Thomson Reuters StarMine Relative Accuracy Score for each research provider uses the past performance of the provider's individual stock recommendations with that of its peers in each sector to calculate a statistical aggregation ranging from 1 to 100. It is calculated over a 24 month period based on the performance of a research firm within a given sector against its peer set of other firms in the market rating stocks in this sector. The calculation is analogous to a "batting average score", that is how often stocks rated "buy" outperform the market and stocks rated "sell" underperform the market as a whole. To get a score higher than 50, the industry-relative return of a firm's recommendations within a sector must, when taken together, be greater than those of the median provider.
Calculate – The normalized analysts' recommendations and the accuracy weightings are combined to create a single score. For the largest 1500 stocks by market capitalization, these scores are then forcibly ranked against all the other scores to create a standardized Equity Summary Score on a scale of 0.1 to 10.0 for the 1500 stocks. This means that there will be a uniform distribution of scores provided by the model and thereby assisting investors in evaluating the largest stocks [in terms of capitalization], which typically make up the majority of individual investors portfolios. Finally, smaller cap stocks are then slotted into this distribution without a force ranking, and may not exhibit the same balanced distribution.
Learn more about the Equity Summary Score Methodology (PDF).
A monthly Equity Summary Scorecard is available to view the aggregate, historical, theoretical performance of the Equity Summary Scores/Sentiments across all covered stocks.
Research firms covering over 6,000 stocks.
What's Provided on Fidelity.com
Equity Summary Score
The Equity Summary Score provides a consolidated view of the ratings from a number of independent research providers on Fidelity.com. Historically, the maximum number of providers has been between 10 and 12. However, some stocks are not rated by all research providers. Since the model uses a number of ratings to arrive at an Equity Summary Score, only stocks that have four or more firms rating them have an Equity Summary Score. Review Understanding and Using the Equity Summary Score Methodology (PDF) to learn more.
Equity Summary Scorecard
Each month Thomson Reuters StarMine provides the historical, theoretical performance of the Equity Summary Scores across all covered stocks. The scorecard includes Total Return by Sentiment; Performance by Sector and Market Cap; and Universe Distribution to help you understand and compare the performance and composition of the Equity Summary Score model. View the current Scorecard.
Thomson Reuters StarMine Relative Accuracy Score
The Thomson Reuters StarMine Relative Accuracy Score measures the historical accuracy of analysts' recommendations. This independent and objective score can help you focus on firms with the strongest track record in deciding which stocks are investment-worthy.
The Thomson Reuters StarMine Relative Accuracy Score is a relative measure that compares a firm's sector recommendation performance with that of its peers over the last 24 months. Scores range from 1 to 100 and are based on a statistical aggregation of how well a firm performed on all of its individual stock recommendations in a given sector. Thomson Reuters StarMine first calculates the individual scores for each set of recommendations for a single stock over a 24 month period. Then Thomson Reuters StarMine rolls up these individual performance scores into a firm's sector score. A firm scores well on an individual stock if it recommends a "Buy" and the stock outperforms its industry, a "Hold" that stays in-line with its industry, or a "Sell" that underperforms its industry. The Thomson Reuters StarMine Relative Accuracy Score is computed by taking a simple average of the firm's single-stock recommendation scores on all stocks in a sector, and statistically adjusting to compensate for the differences in coverage among brokers. To get a score higher than 50, the industry-relative return of a firm's recommendations within a sector must, when taken together, be greater than those of the median firm.
Thomson Reuters StarMine SmartEstimate and Predicted Surprise
Thomson Reuters StarMine utilizes their sophisticated scoring system to provide a smart-weighted earnings estimate. The SmartEstimate is a calculation of the analyst's historical accuracy and timeliness of estimates that provides a weighted estimate rather than a simple average of analyst estimates (Current Consensus). The Predicted Surprise equals the difference between Current Consensus and the Thomson Reuters StarMine SmartEstimate.