# Data audit

• File: data/firm_panel.csv, SHA-256 ae54378f63f4c9527cb2a1061d96d42eb1e45173bb8762a5998966cbfc9672e0 (matches bundle.json)
• 1200 rows x 4 columns: firm_id (int64), year (int64), x (float64), y (float64); dimensionless.
• firms 1001-1100 (100 firms), years 2008-2019 (12 years); each firm has 12 rows, each year 100 rows -> balanced.
• Missing values: 0 in all columns. Exact duplicate rows: 0. Duplicate (firm_id, year) keys: 0.
• x range [-4.52, 3.30], mean -0.16, sd 1.26; y range [-6.29, 5.33], mean -0.42, sd 2.00. No implausible values.
• Rows in random order (irrelevant for estimation).
• No information-availability violations (single permitted input).
• Exclusions: none needed (see out/exclusions.csv; 1200 -> 1200).