base rate:
US manufacturing share of GDP has declined almost every year since 1980, from ~20% to ~9.5% in 2025. The average annual change is -0.121 pp/year. No peacetime period in modern US history shows a 2+ pp increase in manufacturing share over 3-4 years. The only historical precedents for significant increases are WWII-era mobilization.
evidence updates:
Every piece of evidence pushes DOWN from already low base rates: (1) Current share is 9.5%, down 0.7 pp YoY — actively declining, not rising; (2) SCOTUS struck down emergency tariffs, weakening primary policy lever; (3) Replacement 15% tariff is less aggressive and legally uncertain; (4) Manufacturers still shedding jobs; (5) Services growing faster than manufacturing; (6) Factory construction takes years; (7) No historical precedent. The only slight upward consideration is the theoretical possibility of tariff-driven nominal price inflation in goods relative to services mechanically increasing the nominal share, but BEA uses value-added which partially controls for this, and the effect would likely be small (maybe 0.5 pp at most).
combination method:
Weighted average of three sub-questions with strong positive correlation. All three sub-questions converge on extremely low probability (~1-2%), and their conjunction is required for YES. The weighted average gives 1.6%, adjusted to 2% for tail risks.
final:
The probability of US manufacturing reaching 13.1% of GDP by Q4 2028 is approximately 2%. This would require reversing a 45-year structural trend by 3.6 percentage points in under 3 years, with no historical precedent, weakened policy instruments (SCOTUS ruling), and current data showing the trend is moving in the wrong direction. Only an extreme scenario (major war, complete services sector collapse) could plausibly achieve this, and even those scenarios are unlikely to manifest in BEA data by Q4 2028.
ensemble:
{'ensemble_count': 3, 'perspectives': {'base_rate': {'p_yes': 0.02, 'confidence': 0.95}, 'evidence_driven': {'p_yes': 0.02, 'confidence': 0.95}, 'contrarian': {'p_yes': 0.02, 'confidence': 0.95}}, 'spread': 0.0, 'individual_p_yes': {'base_rate': 0.02, 'evidence_driven': 0.02, 'contrarian': 0.02}}