Analyst Calls Monday: Block, Equifax, Lockheed—Alpha Signal or Just Noise?

Monday’s analyst research calls rolled in with the usual fanfare: Block, Boise Cascade, Equifax, FirstEnergy, Lockheed Martin, Lumentum, NetApp, Roper, Spotify. Somewhere, a sell-side analyst upgraded something, and somewhere else, a fund allocated capital based on a price target with a confidence interval wide enough to drive a truck through.

The quant’s honest assessment of analyst calls: they’re a sentiment factor, not a signal in isolation. Womack and Barber’s classic work on recommendation changes showed post-change drift exists, but the effect decays fast and is largely arbitraged away in large caps within days. If your pipeline ingests these calls, treat them as features — recommendation deltas, target-price revision ratios, breadth of revisions across covering analysts — not as trade triggers. A single upgrade on Block is noise; a cross-sectional shift in revision breadth across payments names is information.

The structural problem with sell-side targets is well-documented: persistent optimism bias, asymmetric distribution of ratings, and herding behavior that makes consensus nearly useless as an independent predictor. Any Python backtest worth running (backtrader, zipline-reloaded, or a hand-rolled pandas event study) will confirm that the raw signal is weak, but target-price revisions conditional on earnings surprise timing show measurably better discriminative power. Event-window alignment matters more than the call itself.

Practical filter stack: exclude calls published within 48 hours of earnings (the analyst is narrating, not analyzing), weight by revision magnitude rather than rating category, and check whether the price move on publication already exceeds the historical average drift — if the market front-ran the drift, you’re buying someone else’s fill. Slippage on these event-driven entries is the silent killer; measure it before you model around it.

For the names themselves — Lumentum and NetApp draw calls because optical and enterprise hardware sit on real AI capex flows, which is at least a cash-flow argument rather than a vibes argument. Lockheed is a policy beta dressed as a stock. FirstEnergy’s calls are regulated-utility rate-case arithmetic. Classify your signals by the mechanism generating returns; a factor model that lumps a defense contractor and a payments fintech into one “analyst sentiment” bucket is overfitting with extra steps.

The recurring lesson: the edge isn’t in reading Monday’s calls, it’s in systematically measuring whether they ever predicted anything after transaction costs. Most of the time, they don’t. Run the test anyway — negative results with tight confidence intervals are worth more than any price target.