base rate:
H-1B cap has not been raised since 2004. In any given presidential term, the probability of H-1B cap expansion is historically very low — perhaps 5-10% per term. Previous attempts under more favorable conditions (bipartisan Gang of Eight, Obama era, Bush era) all failed.
evidence updates:
Every piece of evidence strongly pushes DOWN from the already-low base rate: (1) $100K fee causing 87% drop in applications, (2) prevailing wage increases of 21-33%, (3) overall visa issuance down 250K, (4) no legislation found, (5) Republican agenda entirely restrictionist, (6) Big Tech cutting filings 30-50%, (7) no evidence of Musk/tech lobby actively pushing expansion. The only upward factors are speculative: 2.8 years remaining, possible AI labor shortage narrative, possibility of fee backfire prompting reversal. These are weak and speculative, pushing up maybe 1-2%.
combination method:
Used union probability across the two independent resolution paths (legislative and executive), constrained by the very low enabling conditions (presidential support and congressional support). P(legislative) ≈ 5%, P(executive) ≈ 8%, P(at least one) ≈ 12.6%, adjusted down to 8% given that enabling conditions are also low and partially correlated.
final:
The Trump administration's revealed preference through 18 months of policy is overwhelmingly restrictionist on H-1B. Despite late-2024 rhetoric, every concrete action has contracted access. The resolution criteria is specific (cap raise, elimination, or new category) and nothing in the current pipeline approaches this threshold. The legislative path is essentially dead given no bills and no champions. The executive path is marginally more plausible but would require a dramatic reversal of current trajectory. 8% accounts for the remaining ~2.8 years and tail scenarios (AI crisis, tech lobby breakthrough, budget deal horse-trading).
ensemble:
{'ensemble_count': 3, 'perspectives': {'base_rate': {'p_yes': 0.08, 'confidence': 0.82}, 'evidence_driven': {'p_yes': 0.09, 'confidence': 0.78}, 'contrarian': {'p_yes': 0.08, 'confidence': 0.78}}, 'spread': 0.01, 'individual_p_yes': {'base_rate': 0.08, 'evidence_driven': 0.09, 'contrarian': 0.08}}