Read the definition, not just the rating label
FINRA Rule 2241 requires a firm using ratings to explain their meaning, horizon, and benchmark in its research reports. A regulator-facing Buy/Hold/Sell grouping does not make every underlying rating identical.
| Term or field | What to inspect | What not to assume |
|---|
| Overweight / Outperform | The stated comparator and expected performance basis | A universal return hurdle or automatic buy instruction |
| Equal Weight / Neutral / Market Perform | The firm's middle-category definition | The stock cannot decline or is exactly fairly valued |
| Underweight / Underperform | Relative versus absolute expectation | A guaranteed negative return |
| Not Rated / Suspended | Why an opinion is absent | An ordinary Hold recommendation |
| Consensus | Provider mappings, observation window and sample size | One analyst's exact recommendation |
A concrete broker example: Morgan Stanley
Morgan Stanley defines Overweight, Equal-weight, and Underweight relative to a relevant country MSCI index or analyst coverage universe on a risk-adjusted basis over 12 to 18 months. Its disclosures explicitly say these are not equivalent to Buy, Hold, and Sell. The Buy/Hold/Sell correspondence used for regulatory disclosure is a different presentation purpose.
- Check the current report for exceptions and its own price-target horizon.
- Do not convert Overweight into a fixed allocation such as 8% merely because an index weight is 7%.
- Use the original wording when comparing two firms; retain the normalized category only as an organizing aid.
Relative performance can still lose money
These are hypothetical total-return examples using the same period and currency. Return differences below are percentage points, not a risk-adjusted alpha estimate or a test of any specific broker rating.
| Benchmark return | Stock return | Stock minus benchmark | Absolute result |
|---|
| +18% | +26% | +8 percentage points | Gain |
| −22% | −14% | +8 percentage points | Loss |
| +20% | +8% | −12 percentage points | Gain despite lagging |
| +1% | +12% | +11 percentage points | Gain |
Build a rating comparison record
For each note, record the firm, analyst, original label, rating definition, benchmark, publication date, target horizon, and price target. Then compare what actually changed.
- A rating change and a target change can point in different directions. Read the explanation for each.
- A price target is a model output. Check currency, share class, valuation assumptions, and the quote date used for upside.
- A maintained positive rating after a price fall does not by itself prove stubbornness; an upgrade does not prove buying by institutions.
- Compare conflicting assumptions in the reports instead of treating one vocabulary as inherently stronger.