Which Wall Street firms' buy ratings actually beat the market?
We evaluated 42,000+ analyst rating windows across ~500 US large caps against the S&P 500. The results argue for reading the byline, not just the rating.
"JPMorgan upgrades" and "boutique shop upgrades" land in the same headline font. They should not carry the same weight in your research, but until you measure it, you don't know which one deserves more.
So we measured it. StockRank evaluated the historical performance of buy-equivalent analyst ratings across roughly 500 US large caps: for each rating old enough to judge, we took the stock's return over the following 12 months and compared it to the S&P 500 over the exact same window. A rating "beat the market" if the stock outperformed the index. As of July 5, 2026, that's more than 42,000 evaluated rating windows across 300+ firms, one of the larger open tallies of analyst accuracy you'll find outside a Bloomberg terminal.
What the leaderboard shows right now
A sample from the top of the current ranking (minimum 5 evaluated 12-month ratings; full live table on the Top Firms page):
| Rank | Firm | Beat S&P (12M) | Avg excess return | Evaluated (n) | |---|---|---|---|---| | 1 | Northland Capital Markets | 62% | +53.5% | 29 | | 3 | Rosenblatt | 54% | +42.4% | 381 | | 5 | B. Riley Securities | 51% | +36.5% | 51 | | 8 | Melius Research | 58% | +25.8% | 26 |
Historical statistics computed by StockRank from published third-party ratings (via Yahoo Finance, delayed) and daily prices, versus SPY over matched 12-month windows. Recomputed nightly; the live table is the source of truth.
Three honest observations from the data
1. Sample size is everything. A firm at "86% accuracy" on 7 evaluated ratings is a coin-flip streak, not a skill estimate. A firm at 54% on 381 evaluated ratings is a statement. That's why every accuracy figure on StockRank carries its n, and why we'd weight Rosenblatt's 381-rating record very differently from a boutique's 7.
2. Beating the market half the time is good. The S&P 500 is a brutal benchmark. Most buy ratings in our data do not beat it over 12 months, which is exactly why the firms that consistently clear 50%+ with meaningful sample sizes stand out. If you expected star analysts to be right 80% of the time, the data will recalibrate you.
3. Excess return and hit rate tell different stories. Some firms beat the index barely more than half the time but win big when they're right: a high average excess return on a modest hit rate. Others are steadier but smaller. Neither profile is "better"; they're different signals, and the leaderboard shows both.
How to actually use this
When a rating change hits one of your stocks, the two questions worth asking are who said it and what's their record. On every StockRank ticker page, each firm in the ratings table carries its track-record badge, so "Rosenblatt reiterates Buy" arrives with the context "54% of 381 buy ratings beat the S&P over 12 months" attached. The live ratings feed does the same for every upgrade and downgrade across our coverage, daily.
The necessary caveats
These are descriptive historical statistics about third-party opinions, computed at the firm level (our source data doesn't attribute ratings to individual analysts). Past accuracy does not guarantee future accuracy. Ratings are the opinions of the named firms, not StockRank's view, and nothing in this article is investment advice or a recommendation. The methodology (windows, benchmark, exclusions, minimum samples) is published in full here.
See the complete, live leaderboard of 80+ ranked firms on the Top Firms page.