# Step 279: a symmetric demand closure and investment that responds to profitability

**Why.** The third referee report found that the demand test of steps 273–274 was weaker than the text
claimed:

- When demand exceeded capacity, investment was rationed without any cost, and no prices responded.
- Planned investment was calibrated so that demand rarely fell short, so "210 against 211" was close to
  built in.
- Wages saved through the Okun lay-offs went to firms untaxed.

Generated by `scripts/demand_symmetric_v279.py`, with the model in `scripts/closure_model_v279.py` (a
copy of step 278; archived files untouched). Archived in `data/processed/demand_symmetric_v279.json`.
All 432 distinct cases. Base: the revised specification of step 276 (present-value equity, 15% profit
levy when the cap binds), with the demand closure of steps 273–274 (calibrated autonomous investment,
Okun 0.5, 60% unemployment insurance) and profit tax on saved wages. Illustrative; not a forecast.

## New options

- **Phillips.** The price level, and every nominal flow with it, moves with last year's demand gap, in
  both directions: m_t = m_(t−1) × (1 + φ × (demand − capacity)/capacity). Excess demand raises prices
  and erodes nominal contracts; a shortfall lowers prices.
- **q-investment.** The autonomous part of planned investment scales with Tobin's q relative to year 0.
  Investment rises when equity is valued above the capital stock.
- **Accelerator.** On capacity, or on realized output.

With the defaults, the model reproduces step 278 exactly (tested). Money is conserved.

## Results

| Variant | Preserved (of 432) | Fiscal crises | Use of capacity, median / 10th pct | Peak excess demand, median | Extra inflation per year, median | Retained workers keep 90% |
|---|---:|---:|---|---:|---:|---:|
| Revised, no demand constraint | 277 (64%) | 10 | 100% / 100% | 0 | 0 | 432 |
| Demand, prices fixed | 277 | 10 | 100% / 99% | 7.3% | 0 | 432 |
| Phillips 0.25 | 322 | 0 | 100% / 98% | 7.5% | +0.9% | 432 |
| Phillips 0.5 | 356 | 0 | 100% / 97% | 8.0% | +1.6% | 432 |
| **Symmetric (Phillips 0.5 + q-investment)** | **378 (88%)** | 0 | 100% / 99% | 12.9% | +2.9% | 432 |
| Symmetric, accelerator on realized output | 375 | 0 | 100% / 96% | 13.7% | +2.8% | 432 |
| Demand, investment slump, prices fixed | 250 (58%) | 23 | 95% / 90% | 4.5% | 0 | 432 |
| **Symmetric, investment slump** | **150 (35%)** | 204 | 94% / 89% | 5.8% | −2.1% | 345 |
| Symmetric, slump, accelerator on realized output | 165 | 202 | 95% / 90% | 8.2% | −1.6% | 354 |

## What it shows

**The referee was right.**
- The calibrated investment rule does not keep demand at capacity. Most cases see excess demand at some
  point, with a median peak of 7% of output.
- The step-273/274 closure absorbed it silently by cutting investment. Once prices respond, that excess
  demand becomes inflation of about 1.5–3% a year.
- That inflation erodes nominal public debt, which is why preservation rises to 82%–88%.

**The demand side is a first-order uncertainty, in both directions.**
- **If firms keep investing as automation raises capacity,** demand runs ahead of capacity. The
  resulting inflation helps the budget, and the displaced are preserved in most cases.
- **If investment slumps as displacement rises,** demand falls short. Prices fall and the real burden
  of nominal debt rises. Lay-offs spread to the retained workers, fiscal crises occur in about half of
  the cases, and preservation falls to 35%.
- The no-demand result (64%) lies between the two.

**The thesis cannot say which world is more likely.** It can say what each requires. The headline should
present the demand side as a range that depends on investment, not as a robustness check that leaves
the result unchanged.

## Limits

- The Phillips coefficient and the q elasticity are assumed.
- There is no monetary policy. A central bank would lean against both the inflation and the deflation,
  narrowing the range.
- Nominal wages move with the price level one-for-one, so inflation here is neutral for wage earners.