Analyst Calls Stack: Block, Equifax, Lockheed, Spotify—but What Do the Models Say?

# Analyst Calls Are Noise Until You Test Them

Monday’s roundup of Wall Street research calls — Block, Boise Cascade, Equifax, FirstEnergy, Lockheed Martin, Lumentum, NetApp, Roper, Spotify — is the usual buffet of price target revisions and ratings theater. The quant’s job is to treat these as data points, not directives.

Start with the base rate. Sell-side rating changes have a documented tendency to cluster after price moves rather than precede them. Analysts downgrade weakness and upgrade strength, which means a fresh “Buy” on Block or Lumentum often encodes momentum that’s already priced. If you want to trade these calls systematically, don’t read the thesis — measure the event. Pull the timestamp, compute abnormal returns over a 1-day and 5-day window post-publication, and check whether drift persists after controlling for sector beta. `pandas` plus a CRSP or yfinance feed gets you 80% of the way in an afternoon.

The nuance is in the dispersion. A price target change matters more when it diverges from consensus — a lone upgrade on Equifax amid a hold-rated street carries more information than the fifteenth analyst marking to market. Build a consensus-dispersion factor: standard deviation of estimates, direction of revision, and revision breadth across the coverage set. Revision momentum is one of the more robust post-earnings anomalies in the literature, which is more than can be said for most “AI-powered” rating aggregators repackaging the same feeds.

Defense and industrial names like Lockheed and FirstEnergy come with an extra trap: thematic capital inflows distort short-horizon signals. Narratives about geopolitics or rate cycles show up in flows before fundamentals, and backtests that don’t segment regime-dependent flow effects will happily overfit a strategy to one macro cycle. Walk-forward validation with purged k-fold cross-validation isn’t optional here — it’s the difference between a signal and a curve-fit souvenir.

For the tech-adjacent names — NetApp, Roper, Spotify — watch how analyst revisions interact with post-earnings announcement drift. A target raise within a week of earnings, in the direction of the estimate revision, historically carries more signal than a standalone call. Encode it: revision direction × days since earnings × distance from consensus target. Backtest it honestly, including transaction costs and the uncomfortable reality that many of these calls drop at the open when spreads are widest.

The uncomfortable truth about Monday’s research calls is that they’re mostly marketing for the bank’s coverage footprint. The measurable edge isn’t in the rating — it’s in the second-order behavior of the traders who react to it. Model the reaction, not the recommendation.