Verstavo › The Track Record
Would we have flagged it?
OUR EARLY-WARNING SIGNAL, REPLAYED ACROSS 24 YEARS OF FDIC DATA · 2001–2024
Every year we take the banks that look clean today and flag the ones quietly
registering early 30–89 day delinquency — then check who actually crossed into CRE distress a year
later. The flagged banks deteriorated at a median 2.2× the rate
of the ones we didn’t flag — in every one of the 24 cohorts — and
the signal fired before the 2008 crisis, not during it.
Signal holds — every cohort
≈2.2× flagged vs control
Into the GFC · 2008 cohort
39% vs 24% reached distress
Firing now · 2024 cohort
5.7% vs 2.0%
Flagged (clean today, top-quartile 30–89)
Control (clean, not flagged)
Distress = noncurrent CRE reached 3% within one year, among banks below 1% today (total CRE ≥ $50M). Shaded band = 2006–2010.
What this says — and what it doesn’t
- Testable from 2001 only — the 30-89 bucket was not collected earlier, so the replay spans one full prior cycle (the GFC) plus today, not the S&L crisis.
- Survivorship understates the edge: failed banks leave the panel and have no one-year-later reading, and failures are the worst outcomes — so the true lift is higher than shown.
- A signal, not a verdict. It says a flagged bank is likelier to deteriorate, not that it will — read it as awareness, alongside the bank's own numbers.
- The one weak year was 2021 (1.2×) — forbearance and stimulus compressed everyone; we show it rather than sand it off.
See the full backtest — and the banks showing this signal today.
Every cohort, the methodology, and the current watchlist of clean banks with elevated early
delinquency — the live end of this same signal — are in the Verstavo platform.
See the whole picture, not just the pulse.
Market Pulse is the free, public read. The Verstavo platform goes loan‑by‑loan — stress
scores, maturity walls, special‑servicing transfers, bank CRE, and your own portfolio benchmarked
against the market.