{"id":1986,"slug":"imboden-2023-self-employment-income-divergence","title":"Imboden 2023 (US Census/IRS) — survey self-employment income roughly DOUBLE the admin figure","kind":"reference","scope":"business","status":"current","audiences":["kevin","candid-team"],"topics":["self-report-validity"],"reference_body":"**Claim:** Imboden 2023 (US Census Bureau CES Working Paper), using linked survey + administrative records, found **median self-employment income of $37,500 in survey data versus $18,500 in administrative records** — survey reports roughly **double** the administrative figure.\n\nCabral & Gemmell 2021 (*International Tax and Public Finance*), a matched register-and-survey study, found self-employment income **underreporting** of ~20% against tax-return data — but only ~10% in survey data, meaning survey measurement error halves the apparent gap.\n\n**Source:** Imboden 2023 CES Working Paper; Cabral & Gemmell 2021 ITPF. Two independent primary validation studies.\n\n**Confidence:** Verified.\n\n**Why this matters for Candid:** Even \"objective\" business figures like revenue carry large self-report error. The widget should favor binaries and BANDS over precise self-reported numbers (e.g., \"how many reviews? 0 / 1-9 / 10-49 / 50-199 / 200+\" rather than \"how many reviews?\") and read-off-screen counts over recalled ones.","rationale_body":null,"metadata":null,"links":{"outgoing":[],"incoming":[{"slug":"research-brief-smb-widget-capture-layer-june-2026","title":"Research brief: SMB widget capture layer — what owners can vs cannot self-report (June 2026)","kind":"reference","scope":"business","link_type":"relates-to"},{"slug":"rule-r7-read-off-screen-beats-recall","title":"R7 — When the owner can read a number off a screen, have them do so","kind":"rule","scope":"business","link_type":"depends-on"}]},"created_at":"2026-06-23T19:16:00.843Z","updated_at":"2026-06-23T19:16:00.843Z"}