← Every exhibit

Figure Figure 3 From the research bench

Leavers and stayers, coded on one rule

The point estimates are close; this descriptive check does not establish the absence of selection bias.

Departed companies disclosed a retention metric in 60.0% of their final annual reports and survivors in 61.5%. The point estimates are close; no uncertainty test is shown.0%20%40%60%80%100%60.0%Departed, final filing61.5%Survivors, FY2025Disclosing a retention metric (%)
Vector SVG Cite Embed

Reference & Evidence

Source: The final annual report of all 30 departed companies, pulled from SEC EDGAR and coded on the same rule as the panel: 18 of 30 disclose, 60.0% (17 net, 1 both, 12 none). The survivor bar is the balanced panel at fiscal 2025, 59 of 96, 61.5%. This is a descriptive selection check for the four-year series. Single coder; a second pass is owed. Corpus: departed-coding-2026-08-18.json.

Related exhibits