# MacroVision v101: synthesis around the thesis question
29 October 2026. Matched-case synthesis of existing hypothetical simulations, not another macro model or empirical proof.

## Research question
Under what conditions can automation associated with 20% unemployment preserve displaced households' real consumption through lower prices and government transfers, and how do initial debt and productive-asset ownership change the outcome?

The current operational outcome is **nonhousing consumption**, indexed to each cohort's own initial level, not total welfare, an absolute adequate living standard or food security. The diagnostic preservation rule is >=90% in every annual observation over a completed 20-year path. It is a research design choice, not an estimated welfare threshold.

## Direct answer supported by this model
Some specified combinations permit consumption preservation; the mechanism is conditional. Price reductions alone do not guarantee it. Fiscal capture, household liabilities, housing and investment financing matter jointly. This does not establish that the real world with 20% unemployment will be prosperous or that the richest decile will consist of initially debt-free asset owners.

The exercise identifies accounting/financing conditions and vulnerabilities. It neither estimates the effects of automation nor determines the likelihood of the inputs.

## Same cases across existing financing variants
Match all 576 factor combinations by exact parameter tuple across seven variants:
1. No corporate credit.
2. Cash-funded credit limit 20, rate 4%.
3. Cash-funded credit limit 60, rate 4%.
4. Cash-funded credit limit 60, rate 8%.
5–7. Limit-60/rate-4% credit followed by debt-equity exchange after 1, 2 or 3 consecutive missed-service years and no new advances after exchange.

The summary avoids comparing different successful subsets:
- **202** cases complete in all seven variants.
- **18** complete in some variants only: classifications remain unresolved across closures.
- **356** complete in none: no terminal consumption claim can be made for them.
- **75** cases preserve the diagnostic threshold for **both displaced cohorts** in all variants.
- Only **42** cases have completed actual AND matched-policy control paths in all variants.

These 202 shared completed cases are the baseline completed nonborrowers. Corporate credit does not operate in them; the seven-variant invariance is largely mechanical. It is **not independent confirmation, a demanding test of borrowing resilience, or robustness across all economic closures**. Exogenous prices, household behavior, tax rules and technology assumptions remain shared.

| Cohort | Preserved in all seven / 202 shared complete | Below threshold in every variant / 202 | Incomplete in at least one variant / 576 |
|---|---:|---:|---:|
| Retained workers | 202 | 0 | 374 |
| Displaced, no initial consumer debt | 83 | 119 | 374 |
| Displaced, initial consumer debt | 75 | 127 | 374 |
| Productive owners | 202 | 0 | 374 |

The debt-free label refers only to the initial consumer loan; those cohorts can carry mortgages. Initial cohort consumption levels differ, so equal indices do not imply equal living standards.

Among the 42 common complete pairs, terminal nonhousing consumption of both displaced cohorts is below its matched control in every variant. Retained workers are above control in all 42; productive owners are above in 30 and below in 12. Preserving a baseline floor is therefore different from avoiding a loss against a growing counterfactual. Controls remove displacement/productivity but retain policy and housing schedules.

## Conditions: descriptive counts, not sufficient/necessary laws
Counts of cases preserving **both displaced cohorts** among the shared complete set:

| Factor setting | Shared complete cases | Both displaced preserved |
|---|---:|---:|
| Accessible profit share .50 | 72 | 10 |
| Accessible profit share 1.00 | 130 | 65 |
| Profit tax .25 | 106 | 25 |
| Profit tax .35 | 96 | 50 |
| Stable housing/rates | 101 | 41 |
| Housing/rate stress | 101 | 34 |
| Concentrated productive ownership | 120 | 35 |
| Broad productive ownership | 82 | 40 |

These rows condition on completion and aggregate over other factors. They are neither causal estimates nor independent observations. Initial housing wealth differs across tenure alternatives. Counts do not establish that higher tax/capture or broader ownership is universally beneficial.

In particular, price pass-through settings 0/.5/1 have 30/26/19 joint-preservation cases among 64/66/72 shared completed cases. The strongest productivity-gain setting has 24 shared completed cases and zero joint-preservation cases. These nonmonotonic patterns warn against claiming “more productivity/cheaper goods always solves it” under the inherited financing closures. They do not show that productivity or deflation is generally harmful.

## What the last stages contribute
| Thesis link | What has been learned | What remains unproven |
|---|---|---|
| Automation -> employment | A 20% terminal unemployment path is imposed in these scenarios | That automation will cause that rate |
| Productivity -> prices | Pass-through and productive gains are varied as assumptions | Their actual size, timing or general-equilibrium response |
| Profits -> fiscal resources -> transfers | Explicit budget and financing constraints prevent reuse of unavailable money | Sustainable real-world tax collection and benefit levels |
| Debt/ownership -> household outcomes | Heterogeneous cohorts expose unequal consumption paths | Causal protection from liquidity or actual future decile membership |
| Investment -> continuation of production | Corporate cash stops reveal incompatible residual investment/finance in many cases | An empirically credible investment/credit closure |
| Corporate credit/workout | Credit completes some paths, often with unpaid interest; exchanges change ownership without cash | Solvency, cash recovery, optimal restructuring or higher welfare |

## Thesis-ready conclusion
The model supports a conditional proposition: lower prices and transfers can preserve a selected floor of nonhousing consumption for displaced households under some fiscal, ownership, debt and financing configurations. It does not support an unconditional prediction that widespread automation and 20% unemployment preserve purchasing power for everyone. Heterogeneous liabilities and asset ownership create distributional differences, while fiscal and corporate financing constrain feasible paths.

This conclusion remains theoretical. The SHED/SIPP work has not established the required causal liquidity mechanism or authorized macro calibration. Existing transaction-data enquiries concern future feasibility, not acquired evidence. The research must distinguish assumed mechanisms, simulation results and empirical findings throughout.

## Boundaries and priority
The financing extensions were needed to avoid interpreting a programming/model closure stop as economic collapse. Continuing to add them indefinitely would dilute the central question.

Next priority is a **source-and-parameter evidence map** for the three central links: productivity-to-prices, profits-to-collectible revenue, and transfers-to-displaced-household consumption. List which parameters are assumptions, which can be anchored descriptively, and which require identification. Keep the current conditional conclusion fixed while evidence is evaluated. A targeted evidence map serves the thesis better now than another favorable financing variant.

The document is a synthesis for the research chapter, not a finished thesis or a registered protocol. A defensible next empirical test must be chosen based on evidence quality, not the coefficient needed to sustain the narrative.

## Reproduction and validation
Run after generating existing v98/v100 full outputs:
```bash
python scripts/corporate_credit_v98.py
python scripts/corporate_workout_v100.py
python scripts/thesis_synthesis_v101.py
python -m unittest discover -s scripts -p test_thesis_synthesis_v101.py -v
```
Input files are local generated outputs, not participant records. Exact case matching rejects duplicate or missing workout cases; source hashes and generator hash accompany the result. Two tests verify complete-case classification and ensure truncated paths cannot become preservation or failure outcomes.

Output: data/processed/thesis_synthesis_v101.json, including all 576 case classifications and factor-conditioned counts. Original models and economic rules are unchanged. No new empirical calibration or causality approval.

References: [v96 sensitivity](MACRO_SENSITIVITY_V96.md), [v98 corporate credit](CORPORATE_CREDIT_V98.md), [v99 cash-service audit](DEBT_SERVICE_AUDIT_V99.md), [v100 workout](CORPORATE_WORKOUT_V100.md), [v90 empirical uncertainty](SHED_DISJOINT_EXTENSION_V90.md).