Pre-registration
India's 1991 balance-of-payments-crisis-driven liberalisation programme (Manmohan Singh's package: rupee devaluation, industrial delicensing, trade liberalisation, FDI opening, partial financial- sector reform) produced a sustained acceleration in per-capita GDP growth. Pre-reform (1965-1990) annualised real GDP-per-capita growth was substantially lower than post-reform (1992-2019, pre-COVID) annualised growth by at least 2 percentage points per year on a sustained basis. The effect is identifiable even at the smaller within-country structural-break scale.
Falsification criterion — what would disprove this
This hypothesis is considered falsified if:
PRIMARY (dispositive): falsified if the annualised log-growth rate 1992-2019 minus the annualised log-growth rate 1965-1990 is less than 0.02 (2 percentage points per year). A reversed sign would be decisive refutation. INFORMATIVE (not gating): the acceleration should be larger (≥0.025) if the pre-window is trimmed to 1975-1990 (excluding earlier commanding-heights experiments that pulled down the pre-reform baseline).
formal test & threshold
test: india_pre_post_1991_annualised_growth_gap threshold: PRIMARY: annualised_log_growth(1992-2019) - annualised_log_growth(1965-1990) >= 0.02 INFORMATIVE: trimmed-window robustness (1975-1990 pre vs 1992-2019 post) yields a similar or larger gap. METHOD_VALID: WDI NY.GDP.PCAP.KD available for IND across 1965-2019.
Method
- Template
descriptive- Clustering
none- Sample
- 1 countries · 1965 – 2019
- Evidence type
- descriptive
Pre-post annualised log-growth comparison. Primary statistic is the difference between mean annualised log-growth 1992-2019 and 1965-1990. Visualise as a kinked trend line with a break at 1991. No DiD — the 1991 package is abrupt and well-documented; this is a within-country structural break.
Data
| Variable | Source | Transform |
|---|---|---|
log_gdp_pc_constant_usd outcome | world_bank_wdi:NY.GDP.PCAP.KDtier 2 | log |
● ready · ● pending · ● reconstruct-needed
Detailed result card
Result card — India 1991 liberalisation growth acceleration
Verdict: SUPPORTED — post-1991 annualised log-growth +4.67%/yr vs pre-1991 +1.96%/yr; acceleration +2.70pp/yr (threshold +2.00pp/yr).
Headline numbers
- Series: WDI
NY.GDP.PCAP.KD(constant 2015 USD). - Pre-reform window 1965-1990 annualised log-growth: +1.961%/yr (cumulative +0.490 log-points, ~+63%).
- Post-reform window 1992-2019 annualised log-growth: +4.666%/yr (cumulative +1.260 log-points, ~+252%).
- Acceleration: +2.705pp/yr (post − pre).
- Trimmed-pre robustness (1975-1990 vs 1992-2019): +2.353pp/yr.
Threshold applied
- PRIMARY:
annualised_log_growth(1992-2019) − annualised_log_growth(1965-1990) >= 0.02(2pp/yr). - INFORMATIVE: trimmed-pre window (1975-1990) yields a similar or larger acceleration.
| Component | Threshold | Realised | Pass | |---|---:|---:|:---:| | Annualised acceleration | >= +2.00pp/yr | +2.705pp/yr | yes | | Trimmed-pre robustness | >= primary | +2.353pp/yr | no |
Interpretation
This is a within-country structural-break descriptive comparison; results are a pattern match, not causal identification. The 1991 BoP-crisis liberalisation package is bundled with subsequent reforms (1990s telecoms, 2000s services-export boom) and global tailwinds (China-driven commodity supercycle, IT-services offshoring). The descriptive estimator documents the break; it does not isolate the marginal contribution of 1991-specific instruments.
Sources
- World Bank WDI
NY.GDP.PCAP.KD(vintage NY.GDP.PCAP.KD@2026-04-28T125340Z.parquet).
Steelman live concerns
See hypotheses/steelman/india_1991_liberalisation_growth_acceleration.md.
Strongest opposing argument
Every hypothesis ships with its charitable opposing argument. The framework earns credibility by handling objections at their strongest, not weakest.
Notes
Canonical natural-experiment tier (HYPOTHESIS_FRAMEWORK_AUDIT.md §E6 case #10). Runnable today with WDI data. Primary spec deliberately simple (structural-break descriptive comparison) — in a framework that can handle magnitudes smaller than the China 1978 case, this 2-pp annualised acceleration is still a first-order signal.