Monday’s Analyst Calls: Signal or Noise? The Quant-Backed Screen on Block, Spotify, More
Monday’s analyst calls arrived on schedule, as they always do: Block, Boise Cascade, Equifax, FirstEnergy, Lockheed Martin, Lumentum, NetApp, Roper, Spotify. The headline list looks authoritative. The information content, at least as a trading signal, is closer to noise.
The well-documented pattern in sell-side research — Womack and later Barber, Lehavy, and Trueman quantified it — is that rating changes do carry short-horizon drift, but it’s small, front-loaded, and largely arbitraged within days. By the time an aggregated headline hits a feed, you’re trading the residual of a residual. If you want to harvest this, you don’t read the calls; you backtest them. Pull ratings history from an Estimize or I/B/E/S-style dataset, map rating revisions to returns via `pandas` + `vectorbt`, and measure event-window alphas at t+1 through t+5. The results are typically attenuated so badly by execution slippage that only the upgrade-to-Strong-Buy transitions with unusually large magnitude revisions survive transaction costs.
Two names on this list deserve more skepticism than the rest. Lumentum calls tend to trade on optics demand narratives that swing violently with hyperscaler capex commentary — the factor exposure there is mostly a leveraged bet on one industry’s procurement cycle, not company alpha. Spotify analysts argue about subscriber growth while the stock’s realized volatility clusters around earnings dates; any systematic approach there should be built around an options-implied move model (the standard `implied_vol` decomposition from the CBOE’s own data is sufficient), not analyst price targets, which historically have near-zero directional skill at the 12-month horizon per multiple academic replications.
The boring tickers are where the data gets interesting. Boise Cascade and Equifax are rate-sensitive, cyclical-beta names whose analyst revisions correlate strongly with macro factor moves — pull FRED data on building permits and consumer credit, run a simple OLS of revision direction against trailing macro surprises, and you’ll find the “analyst insight” is largely a lagged echo of data everyone already had. FirstEnergy’s calls carry regulatory idiosyncratic risk that no rating model prices cleanly.
The actionable takeaway: treat aggregated sell-side calls as a feature, not a signal. Ingest them, weight them, backtest their incremental value against momentum and revision-factor baselines, and expect to find the edge mostly evaporates after costs. Analysts are excellent at summarizing the recent past with confidence intervals they decline to publish.