מסמך טכני מיומן המחקר · באנגלית, כפי שנכתב בשלב העבודה
Step 269: the Israeli base year under the new closures, with four fixes
Why. The second referee report raised five points about step 249:
- The 1% interest rate is labelled real, but the model charges it on nominal debt, so it is a nominal
rate. Under deflation the real rate is higher.
- VAT is levied on all of output, including investment.
- Year 0 balances by construction, so the room for any benefit starts at zero. Israel ran primary
deficits in 2021–2024.
- The survey consumption share (0.76) belongs next to the model's (0.90).
- Two results were left unexplained: no case is preserved at a 25% productivity gain (0 of 48), and
fewer cases are preserved at a 100% gain than at 50%.
Generated by scripts/israel_v269.py; archived in data/processed/israel_v269.json. The base year is step 249's. The closures are the step-266 firm closure and the step-268 central government rule: surpluses returned, risk premium, 90% debt cap. Illustrative; not a forecast.
Fixes
- Interest: relabelled. The 1% is a nominal rate on nominal debt, equal to the real rate only while
prices are stable. A variant uses 3% nominal, with existing spending recalibrated so year 0 balances.
- VAT: now charged on consumption. VAT revenue moves with last year's nominal consumption and is
paid out of firms' after-tax profit. The corporate tax base still deducts the output-based VAT, an approximation worth about a tenth of the VAT difference.
- Year-0 room: variants with a primary surplus of 1% of GDP and a primary deficit of 2% in year 0.
- Consumption share: every variant is run at 0.90 and 0.76.
- Duplicates: the 144 Israeli cases include 36 renter and outright-owner cells that are identical
across stress settings (step 250). Distinct counts are shown too.
Results (144 cases; 108 distinct)
| Variant | Preserved, share 0.90 | Preserved, share 0.76 | Distinct, 0.90 | Israeli tenure mix, pass-through 0 / 0.5 / 1 (0.90) |
|---|---|---|---|---|
| Step 249 settings (reproduced) | 82 | 127 | 59 | 65% / 64% / 55% |
| Central | 85 | 90 | 62 | 65% / 65% / 56% |
| VAT on output | 85 | 90 | 62 | 65% / 65% / 56% |
| Interest 3% nominal | 85 | 90 | 62 | 65% / 65% / 56% |
| Primary surplus 1% in year 0 | 90 | 117 | 66 | 65% / 65% / 65% |
| Primary deficit 2% in year 0 | 71 | 85 | 55 | 52% / 52% / 38% |
| Reach 0.60 | 90 | 134 | 66 | 65% / 65% / 65% |
| Rate 30% | 90 | 120 | 66 | 65% / 65% / 65% |
What changes.
- The headline number hardly moves (82 → 85 of 144), but it is now stable. In step 249 the survey
consumption share raised preservation from 82 to 127. That came from "saving becomes investment" with free credit. Under the new closures it adds only 5 cases (85 → 90). The 0.76 result no longer needs a separate warning: both shares give about 60%.
- The VAT base and the interest label do not change the counts. They change the relative gap a
little.
- The starting budget matters most. A 2% primary deficit in year 0 costs 14 cases. A 1% surplus
adds 5, and 27 at the survey consumption share. Since Israel ran deficits in 2021–2024, the central case (balanced) is optimistic.
- The relative gap almost closes. The displaced without debt end 6% below the matched control, and
those with debt 1% above. In step 249 the figures were 26% and 21% below. Returned surpluses reach everyone, including the displaced.
The two unexplained results
A 25% gain preserves nothing (0 of 48 in every variant at 0.90). The budget starts with no room. Automation displaces workers and cuts wage-tax revenue, while a small productivity gain adds little output. In all 48 cases net debt reaches the 90% cap and the benefit is rationed. The median room at year 10 is −1.4% of output, and the displaced fall to a few percent of their initial consumption at the worst point. With a 1% primary surplus at the survey consumption share, 25 of the 48 cases survive.
Fewer cases are preserved at a 100% gain than at 50% (40 against 45). This is not a budget effect: none of these cases is rationed, and peak net debt stays near 0.62. The extra failures are all under concentrated ownership. Most are under full pass-through, where the larger gain lowers the price level more. Rent and mortgage payments are nominal, and the benefit is indexed to goods. The minimum consumption index of the failing cases is 64–85. It is the deflation-housing channel of step 267.
Limits
- Housing burdens are the grid's, scaled to reference income, as in steps 230 and 249.
- The 2021–2024 means include COVID-19 and war years.
- With VAT on consumption, the corporate tax base uses the output-based VAT.