Displacement Report: replication materials
The code, parameters and derived data behind the NextChapter Displacement Report are public, so anyone can check our numbers. The scripts download Census Current Population Survey microdata directly and rebuild the index, breakdowns and margins of error published in the September 2026 edition.
What you can reproduce
| Output in the report | Where in the repository |
|---|---|
| White-Collar Long-Term Unemployment Index (Exhibit 1), monthly 2015–2026 | code/01_wc_index_monthly.py → data/wc-index.csv |
| Breakdowns by occupation, industry and age; older workers; graduates; blue-collar | code/02_cps_breakdowns.py, code/03_cps_aggregates.py |
| 90% margins of error (BLS generalized variance method) | code/04_margins_of_error.py with data/bls_gvf_parameters.json |
| Seasonal-adjustment robustness (fixed factors, X-13ARIMA-SEATS, STL, 12-month average) | code/05_seasonal_sensitivity.py, code/x13_wcltu.spc |
| AI in SEC restructuring filings (Form 8-K, Item 2.05) | methodology/sec_item205_search.md, data/sec-item205-ai-coding-2026.csv |
| Hires per job opening, 2001–2026 | data/hires_per_opening.csv (BLS JOLTS via FRED) |
| Revenue per employee, 27 large white-collar employers | data/displacement-report-2026-q3-revenue-per-employee.csv |
| Industry productivity vs long searches (Q3 feature) | code/08_industry_productivity.py, data/industry_productivity_vs_long_searches.csv |
| Nativity breakdown and age-sex standardized 55+ participation | code/09_nativity_and_age_standardization.py |
| Beyond-unemployment white-collar panel | code/10_hidden_displacement.py, data/cps_hidden_displacement_wc_jan_aug_2019_2025_2026.json |
| AI exposure vs outcomes by occupation (including the customer service test) | code/11_exposure_vs_outcomes.py, data/exposure_test_results.json |
| Education and experience breakdowns | code/12_education_experience.py, data/cps_education_experience_2019_2022_2025_2026.json |
| Total factor productivity decomposition (Q3 addendum) | code/13_tfp_decomposition.md, data/sf_fed_tfp_decomposition.csv (SF Fed quarterly TFP series) |
| Senior openings tracker (pilot, Appendix A) | code/07_openings_tracker_pilot.py (needs the NextChapter jobs database; not reproducible externally) |
Code: MIT License. Data files and documentation: CC BY 4.0. Census microdata are not redistributed; the scripts fetch them with a free Census API key.
How we record changes between editions
Every change between versions and editions (a new section, a revised figure, a correction or a new data source) is logged with what it was before, what it is now, why it changed and whether earlier numbers are affected. The rules are in methodology/change_record_standard.md, and the record itself is the repository's CHANGELOG.
Found an error?
Send corrections to Justin Kulla at jkulla@launchyournextchapter.com. Corrections are dated and noted in the report's version log and the repository's changelog.