Analyst Calls Pile Up Monday—Signal Density High, Conviction Yet Unproven

Monday’s analyst research call roundup — Block, Boise Cascade, Equifax, FirstEnergy, Lockheed Martin, Lumentum, NetApp, Roper, Spotify — is a list, not a signal. Treat it accordingly.

The sell-side price target is one of the least informative features in market data. Academic work going back decades ( Barber, Lehavy, McNichols, Trueman on analyst recommendations; Womack on recommendation revisions) finds the actionable information concentrates in *changes* and in *downgrades*, not in the steady drumbeat of Buy reiterations with nudged targets. If your pipeline ingests these feeds daily, the feature you want is the delta: target revision magnitude normalized by price, recommendation transition matrices, and time-since-last-revision. A “PT raised from $210 to $215” carries roughly the information content of a coin flip.

That said, aggregated revisions do carry drift. Price-target revision breadth — the fraction of covering analysts moving targets up versus down over a rolling window — is a documented post-revision drift factor and trivially computable in Python. Pull targets via your data vendor, resample weekly, compute breadth and dispersion, and you have a usable signal for a low-turnover overlay. Dispersion matters too: wide target dispersion is a proxy for fundamental uncertainty and historically correlates with underperformance in cross-section ( Diether, Malloy, Scherbina’s divergence-of-opinion result).

Practical caveats worth encoding into any backtest: analyst targets exhibit systematic optimism bias (targets historically overshoot realized prices on the upside), coverage initiation is correlated with banking relationships, and recommendation changes cluster around earnings, so avoid double-counting if you already trade post-earnings drift. Survivorship is the classic trap here — backtesting only on currently-covered names silently deletes every downgrade-to-delisting path. Point-in-time snapshots or nothing.

Names on today’s list that a quant would actually screen deeper: Lumentum, where optics demand cycles make revision momentum more informative than usual; and NetApp, where enterprise hardware upgrade cycles tend to show up in estimate revisions one to two quarters before the sell-side consensus catches up. The rest is narrative dressed in a rating.

The edge was never in reading the calls. It’s in systematically measuring how the crowd moves after them.