Results Explorer

Comprehensive analysis results, data export capabilities, and figure gallery

4.2-5.2pp

AI Causal Impact

Agglomeration Effect

8.4pp vs 1.2pp

Industry Heterogeneity

High-AI vs Low-AI

60-80%

Aging Offset

2050 Projection

5 Methods

Causal Identification

Robust Results

Comprehensive Results Summary

Main Treatment Effect
4.5pp ±0.5

AI implementation increases spatial agglomeration by 4.2-5.2 percentage points across all identification methods.

DiD: 4.5pp (SE: 0.18, p < 0.01)
Event Study: 4.2pp (95% CI: [3.8, 4.6])
Synthetic Control: 4.8pp (p < 0.05)
IV: 5.2pp (SE: 0.22, p < 0.01)
Industry Heterogeneity
7.2pp gap

Significant variation in AI effects across industry sectors, with tech-intensive industries showing larger responses.

High-AI Industries: 8.4pp (p < 0.001)
Low-AI Industries: 1.2pp (p = 0.08)
Finance & Tech: 9.8pp (p < 0.001)
Manufacturing: 3.1pp (p < 0.05)
Temporal Dynamics
3-year buildup

Effects build gradually over 3-year period following AI implementation, with peak impact in year 3.

Year 1: 1.2pp (initial adoption)
Year 2: 2.8pp (learning effects)
Year 3: 4.5pp (full integration)
Year 4+: 4.6pp (steady state)
Geographic Variation
Core-periphery

Effects concentrated in central Tokyo wards with strong network connections and digital infrastructure.

Central Wards: 6.8pp (Chiyoda, Shibuya, Shinjuku)
Peripheral Wards: 2.1pp (outer Tokyo)
Tech Clusters: 8.9pp (concentration effects)
Traditional Areas: 1.8pp (slower adoption)
Download Resources
Results Tables (CSV) All Figures (PNG/PDF) Raw Data (ZIP) Complete Repository
Robustness Tests
  • Placebo Tests: No effects in non-treatment periods
  • Sensitivity Analysis: Results stable across specifications
  • Alternative Controls: Consistent with different control groups
  • Subsample Analysis: Effects hold across subsamples
Publication Status

Under Review

Journal of Urban Economics

Interactive results explorer with filtering and visualization tools coming soon.

View Full Repository Back to Home