Data downloads
The research findings as CSV files. Every file is built automatically from the archived research results, with no manual typing. The files at the top come from the central specification (step 281) and the steps after it (282–304).
- Download (CSV)
Central specification: all 432 cases
The assumptions of each case, whether both displaced cohorts kept 90% of consumption, whether the benefit was cut or the profit levy used, peak debt, and each cohort’s gap to the economy without automation (mean of the last five years).
- Download (CSV)
Central specification: preservation by factor
In how many cases both displaced cohorts kept their consumption, by the value of each factor. Counts, not probabilities.
- Download (CSV)
Productivity gain needed for each combination
For each of the 144 combinations of assumptions, the smallest productivity gain above which consumption is preserved, under three financing rules: rationing the benefit only, a levy of up to 15% of profits, and a levy of up to 30%.
- Download (CSV)
Pass-through of labour cost to prices
Pass-through estimates from data on 50 industries in 15 European countries, 2000–2023: within a year and over five years, with country-clustered confidence intervals.
- Download (CSV)
Demand, interest rates and fiscal support: all scenarios
Steps 282–285: in how many of the 432 cases both displaced groups kept their consumption, and how many fiscal crises occurred, by the strength of the investment slump, the interest-rate rule, the slope of the Phillips curve, and fiscal support and its financing.
- Download (CSV)
Cost of the levy, responsive wages and an indexed benefit: all runs
Steps 301–304: for each run, in how many of the 432 cases both displaced cohorts kept their consumption and how many fiscal crises there were, with and without a levy, when the levy has costs, when wages respond and when the benefit is indexed to wages.
- Download (CSV)
Israel’s unemployed by occupation
Share of the unemployed, share of the employed and relative risk, 2017–2024 (CBS, Labour Force Survey).
- Download (CSV)
AI exposure by occupation group
ILO 2025 index for 40 occupation groups: score, rank and range across occupations in the group.
Previous versions
Files built from earlier versions of the model (steps 101–205). They are kept for replication and comparison, and their numbers are not the central specification.
- Download (CSV)
All 576 scenarios, previous version
The assumptions of each scenario and the classification of each household group in the original grid of step 101.
- Download (CSV)
Preservation by factor, previous version
In how many paths both displaced cohorts kept their consumption, by the value of each factor. Counts, not probabilities.
- Download (CSV)
Gap to the control path, previous version
Final consumption against the same economy without automation, in the matched comparisons of step 196 (percent).
- Download (CSV)
Sensitivity to the preservation floor, previous version
Preservation at floors of 80%, 85%, 90% and 95%, in the paths completed in step 196.
- Download (CSV)
Preservation by housing tenure and pass-through, previous version
Renters, mortgage holders and owners, at each level of cost savings passed to prices.
- Download (CSV)
Israel housing calibration, previous version
Share of paths where consumption was preserved, with Israel’s tenure mix and housing burdens (household expenditure survey 2024).
Documents
How to cite
Parnas, I. (2026). Cheaper Goods, Thinner Budgets: Can Lower Prices and Tax-Financed Benefits Protect Workers Displaced by Automation? (MacroVision) Working paper. Zenodo. https://doi.org/10.5281/zenodo.24134846
The PDFs on this site are draft 3.2, newer than version 3.1 archived on Zenodo.
Simulation results are illustrations from a model partly calibrated to Israel, and counts of paths are not probabilities. CBS and ILO data are processed and descriptive.