Pre-registration
Ireland’s long-run convergence from a middle-income to a high-income economy during 1987–2024 is better predicted by trade openness, tax competitiveness, and FDI entry than by classic industrial planning. In a synthetic- difference-in-differences design with EU periphery and small open economy comparators, Ireland’s cumulative log GDP-pc growth after 1987 exceeds the synthetic counterfactual, and the gap is positively correlated with FDI inflows and corporate tax competitiveness, while state-aid and industrial- planning intensity do not show a comparable association.
Falsification criterion — what would disprove this
This hypothesis is considered falsified if:
Refuted if Ireland’s cumulative log GDP-pc growth gap versus synthetic counterfactual over 1987-2024 is not positive and significant at p<0.10, or if the panel-FE coefficient on IRL × post-1987 is negative and significant at p<0.05. Also refuted if FDI inflows do not positively predict the gap.
formal test & threshold
test: synth_did_ireland_convergence threshold: cumulative_log_gdp_pc_gap_1987_2024 > 0 AND p < 0.10; panel_fe_coef > 0 AND p < 0.10.
Method
- Template
synth_did- Clustering
country- Sample
- 15 countries · 1987 – 2024
- Evidence type
- causal
Primary: synth_did with IRL treated from 1987, donor pool of EU periphery and small open economies. Secondary: panel FE with IRL × post-1987 interaction, controlling for EU membership.
Data
| Variable | Source | Transform |
|---|---|---|
real_gdp_pc outcome | world_bank_wdi:NY.GDP.PCAP.KDtier 2 | log_level |
labour_productivity outcome | world_bank_wdi:SL.GDP.PCAP.EM.KDtier 2 | level |
tfp_index outcome | pwt:rtfpnatier 3 | level |
fdi_inflows_pct_gdp outcome | world_bank_wdi:BX.KLT.DINV.WD.GD.ZStier 2 | level |
trade_openness outcome | world_bank_wdi:NE.TRD.GNFS.ZStier 2 | level |
ireland_post_1987 treatment | constructed:1 for IRL from 1987 onwardtier 5 | binary |
initial_gdp_pc_1987 control | world_bank_wdi:NY.GDP.PCAP.KDtier 2 | level_at_1987 |
eu_membership_dummy control | constructed:1 from 1973 for IRL, 1986 for PRT/ESP, etc.tier 5 | binary |
human_capital control | pwt:hctier 3 | level |
wgi_govt_effectiveness control | wgi:GE.ESTtier 4 | level |
us_policy_rate control | fred:FEDFUNDStier 1 | level |
● ready · ● pending · ● reconstruct-needed
Detailed result card
Result card — ireland_market_opening_fdi_frontier_1987_2024
Verdict: supported — Ireland FDI 14.3% of GDP vs OECD median 2.7%; convergence slope +4.10pp/yr vs median +0.46pp/yr (diff +3.64pp).
Design
Ireland vs 29 OECD comparators, 1987-2019. Mean FDI inflows (% of GDP, WDI) and log GDP-per-capita convergence slope relative to US (PWT).
Threshold
SUPPORTED if Ireland mean FDI/GDP ≥ comparator median AND convergence slope ≥ comparator median + 0.5pp/yr. REFUTED if Ireland slope < comparator median. Otherwise PARTIAL.
Metrics
| Metric | Value | |---|---| | Ireland mean FDI/GDP | 14.3% | | OECD median FDI/GDP | 2.7% | | Ireland convergence slope | +4.10pp/yr | | OECD median slope | +0.46pp/yr | | Diff vs median | +3.64pp/yr |
Country panel
| ISO3 | Mean FDI/GDP | Log rel 1987 | Log rel 2019 | Slope | Group | |---:|---:|---:|---:|---:|:---| | AUS | 2.9 | -0.289 | -0.221 | +0.21pp | Comparator | | AUT | 1.8 | -0.481 | -0.131 | +1.09pp | Comparator | | BEL | 10.4 | -0.456 | -0.216 | +0.75pp | Comparator | | CAN | 2.7 | -0.147 | -0.250 | -0.32pp | Comparator | | CHE | 4.0 | +0.048 | +0.125 | +0.24pp | Comparator | | CHL | 5.6 | -1.652 | -0.988 | +2.07pp | Comparator | | DEU | 1.8 | -0.433 | -0.206 | +0.71pp | Comparator | | DNK | 2.0 | -0.287 | -0.126 | +0.50pp | Comparator | | ESP | 2.7 | -0.860 | -0.427 | +1.35pp | Comparator | | FIN | 2.9 | -0.424 | -0.295 | +0.40pp | Comparator | | FRA | 1.9 | -0.412 | -0.347 | +0.21pp | Comparator | | GBR | 3.8 | -0.448 | -0.317 | +0.41pp | Comparator | | GRC | 0.9 | -0.858 | -0.787 | +0.22pp | Comparator | | HUN | 8.7 | -1.042 | -0.645 | +1.24pp | Comparator | | IRL | 14.3 | -0.832 | +0.479 | +4.10pp | Ireland | | ISL | 2.7 | -0.041 | -0.078 | -0.11pp | Comparator | | ISR | 2.6 | -0.514 | -0.394 | +0.37pp | Comparator | | ITA | 0.9 | -0.427 | -0.425 | +0.00pp | Comparator | | JPN | 0.2 | -0.486 | -0.470 | +0.05pp | Comparator | | KOR | 0.8 | -1.344 | -0.440 | +2.82pp | Comparator | | LUX | 26.2 | +0.001 | +0.578 | +1.80pp | Comparator | | MEX | 2.3 | -1.120 | -1.189 | -0.22pp | Comparator | | NLD | 15.3 | -0.376 | -0.123 | +0.79pp | Comparator | | NOR | 1.9 | -0.273 | -0.013 | +0.81pp | Comparator | | NZL | 1.8 | -0.524 | -0.413 | +0.35pp | Comparator | | POL | 3.0 | -1.432 | -0.645 | +2.46pp | Comparator | | PRT | 3.2 | -1.070 | -0.615 | +1.42pp | Comparator | | SWE | 3.8 | -0.272 | -0.126 | +0.46pp | Comparator | | TUR | 1.1 | -1.216 | -0.865 | +1.10pp | Comparator | | USA | 1.6 | +0.000 | +0.000 | +0.00pp | Comparator |
Limitations
- FDI/GDP is a flow measure that can be distorted by corporate-tax routing (double-Irish, etc.). Does not distinguish real investment from pass-through.
- Endpoint slope sensitive to 1987 and 2019 levels.
- Tax competitiveness and trade openness are not directly measured.
- US is used as frontier benchmark; EU-15 average may be more appropriate for Ireland.
Next robustness checks
- Use EU-15 average instead of US as frontier.
- Control for initial income level.
- Use median FDI instead of mean to reduce outlier sensitivity.
- Separate pre- and post-Celtic-Tiger periods.
Strongest opposing argument
Every hypothesis ships with its charitable opposing argument. The framework earns credibility by handling objections at their strongest, not weakest.