מסמך טכני מיומן המחקר · באנגלית, כפי שנכתב בשלב העבודה
Step 297: levy-ceiling sensitivity and distance to the fiscal thresholds
Compare no support, 50% gap support and 75% gap support at three shared annual levy ceilings (10%, 15%, 20% of positive profits), four investment decay settings and two interest sensitivities. Each comparison uses the same 432 cases and applies the same shared ceiling even without support. The benefit levy has first claim on the shared ceiling. The individual levy rates and the support spending cap are held fixed. The joint pass definition is the same as v296.
Counts below are no support / 50% / 75%.
| Investment decay | Rate sensitivity | Ceiling 10% | Ceiling 15% | Ceiling 20% |
|---|---|---|---|---|
| 0 | 0.6 | 259 / 259 / 259 | 297 / 297 / 296 | 297 / 297 / 296 |
| 0 | 0.15 | 255 / 258 / 258 | 265 / 268 / 267 | 265 / 268 / 267 |
| 0.518 | 0.6 | 242 / 253 / 260 | 269 / 280 / 285 | 269 / 280 / 285 |
| 0.518 | 0.15 | 223 / 243 / 249 | 261 / 285 / 291 | 261 / 285 / 291 |
| 1.27 | 0.6 | 60 / 122 / 128 | 60 / 143 / 144 | 60 / 176 / 175 |
| 1.27 | 0.15 | 27 / 81 / 82 | 27 / 88 / 90 | 27 / 109 / 115 |
| 1.5 | 0.6 | 22 / 59 / 59 | 22 / 66 / 64 | 22 / 81 / 82 |
| 1.5 | 0.15 | 8 / 37 / 38 | 8 / 38 / 38 | 8 / 46 / 45 |
The 50% rule retains every joint pass from no support across the 24 settings. At 75%, five losses summed across settings occur at ceiling 15% and the same number at ceiling 20%. All occur at the lowest productivity setting. At each of these two ceilings, two losses fail only consumption at zero investment decay, and three fail only debt at decay 0.518 and interest sensitivity 0.60. No losses occur at ceiling 10%. All 24 joint counts at the 15% ceiling reproduce v296.
The consumption margin is the minimum consumption index across both displaced groups and all years minus 90. Debt headroom is 90 minus peak net public debt in percentage points of output. Negative values are shortfalls. The table reports medians across all 432 cases at the historical stress decay of 1.5 and rate sensitivity 0.60. Rationing counts cases with more years of less-than-full payment of calculated benefits than under no support at the same ceiling.
| Ceiling | Support | Median consumption margin | Median debt headroom (pp) | Cases with more benefit rationing years |
|---|---|---|---|---|
| 10% | none | −90.00 | −67.00 | 0 |
| 10% | 50% | −87.35 | −36.66 | 29 |
| 10% | 75% | −81.63 | −36.46 | 35 |
| 15% | none | −90.00 | −41.51 | 0 |
| 15% | 50% | −70.61 | −22.09 | 116 |
| 15% | 75% | −70.51 | −19.64 | 128 |
| 20% | none | −90.00 | −41.51 | 0 |
| 20% | 50% | −36.91 | −13.55 | 3 |
| 20% | 75% | −41.98 | −11.72 | 9 |
At 20% and 50% support in this stress case, the median within-case improvements are 12.60 consumption index points and 38.36 debt headroom points. Differences between separate medians need not equal these paired medians. Both absolute medians remain below their thresholds. Preservation of baseline joint passes can coexist with additional benefit rationing in other cases. Rationing years are an adequacy diagnostic for the model's calculated benefit, not a comparison of nominal benefit amounts or a decomposition of causal effects.
Run python scripts/israel_investment_financing_v297.py for all 31,104 runs and their matched changes. The upper ceiling is a sensitivity assumption. The model does not estimate how firms would change behavior under higher levies.