## wp1903

## Source details

**Canonical URL:** [wp1903](https://www.imf.org/-/media/files/publications/wp/2019/wp1903.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2019/wp1903.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2019/wp1903.pdf.json)

---

### Introduction and Research Question
- Focus: How citizenship laws (jus soli vs jus sanguinis vs mixed) affect economic development, particularly in developing countries.
- Hypothesis: Jus soli (inclusive) laws facilitate integration and are associated with higher GDP per capita; jus sanguinis (exclusive) laws hinder integration and growth, especially where institutions are weak.
- Main analysis sample: 105 developing countries.

### Citizenship Law Concepts, History, and Mechanisms
- Definitions and regimes:
  - Jus soli: citizenship by birth on the country’s soil (inclusive).
  - Jus sanguinis: citizenship by blood/parents (exclusive).
  - Mixed regime: elements of both; increasing convergence over time.
- Historical patterns:
  - Western Hemisphere: prevalent jus soli (examples include United States, Canada historically).
  - Continental Europe, parts of Asia and Africa: historically jus sanguinis; documented shifts (example: Germany reform in 1999).
  - Colonial legacies: many colonies inherited colonizer citizenship regimes; major post-independence switches documented.
- Mechanisms linking citizenship regimes to development:
  - Distorting investment: lack of citizenship leads to short investment horizons, asset diversion overseas, avoidance of specialized capital investment.
  - Political instability and corruption: marginalized non-nationals may induce instability, higher military/public security spending, or encourage rent-seeking.
  - Public sector efficiency: ethnic/identity fragmentation linked to lower public good provision and weaker institutions.
  - Labor market distortions: exclusion from professions, restrictions on ownership (e.g., land) reduce labor market efficiency.
  - Revenue effects: Economic Citizenship Programs — inclusive citizenship policies can directly raise revenues in some cases.

### Data, Model, and Econometric Approach
- Data compilation and main variables:
  - Citizenship law dataset coded as of 2014 from multiple sources.
  - Main binary variable: Jusoli = 1 if country has a jus soli regime, 0 otherwise.
  - Categorical variable Citizenship law: 1 = jus sanguinis; 2 = mixed; 3 = jus soli.
  - Additional features recorded: citizenship by marriage, gender discrimination in transmission, residency requirements, renunciation requirement, language/income thresholds.
- Sample and controls:
  - Main cross-country sample: 105 developing countries.
  - Core controls: latitude; rule of law index; education proxied by secondary school enrolment rate; trade openness = (imports + exports)/GDP; financial depth = credit to private sector / GDP.
- Econometric strategies:
  - Cross-section OLS baseline: y_i = δ0 + δ1 Jusoli_i + θ X_i + ε_i ; dependent variable: level of GDP per capita at end of period (1970–2014).
  - Panel strategy: seven sub-periods of five years; random effects and Hausman-Taylor estimators used because Jusoli is largely time-invariant (less than 5 percent of sample changed law during 1975–2014).
  - Instrumental variables: legal origin (British/French) and colonial power citizenship regime used to instrument Jusoli.
  - Counterfactual: Synthetic Control Method (SCM) and pooled SCM applied to countries with colonial-era jus soli that later switched; pre-treatment covariates include latitude, Penn World Tables human capital proxy, population size, real capital stock, real income per capita (levels in 1960, 1964, 1969).
  - Treated-year assumption for SCM: change in citizenship law assumed in 1970 for all treated units (sensitivity checks with 1965 and 1975 produced qualitatively unchanged findings).

### Main Results — OLS, Panel, and IV
- Cross-country OLS headline:
  - Shifting from a jus sanguinis to a jus soli law is associated with a 34 log point increase in income per capita (about a 40½ percent increase) in the full specification (Table 1, column 7).
  - The Jusoli coefficient is statistically significant across specifications.
  - Interaction (Rule of law)*Jusoli has expected sign (dampening effect of stronger institutions) but is not conventionally significant; rule of law itself often significant.
- Panel estimates:
  - Pooling and random-effect models show positive and statistically significant effects of Jusoli on log GDP per capita: examples include 0.350 in pooling; 0.575 and 0.674 in random-effect variants; 0.739 in a Hausman-Taylor variant.
  - Panel sample: 97 countries; 331 observations in main panel specifications.
- IV and causality checks:
  - IV specifications instrumenting Jusoli with British and French legal origin and with the colonial power’s citizenship regime produce results consistent with causality: Jusoli remains positive and significant.

### Robustness and Additional Findings
- Additional controls and composite indices (Table 4 and robustness checks):
  - Government consumption ratio to GDP: coefficient -0.167 (associated with lower GDP per capita).
  - Ethno-linguistic fractionalization: coefficient -0.008 (negative association, marginal significance).
  - Protection of property rights: positive association (small but significant coefficient).
  - Composite index (Marriage and Naturalization) constructed by PCA from sub-indicators (citizenship by marriage; gender equality in transmission; years of residency; no-renunciation dummy; number of other conditions) enters positively and significantly.
  - Interaction Jusoli*(Marriage and Naturalization) is negative and significant, indicating substitutability across inclusion channels.
- Appendix OLS details (selected coefficients from Appendix Table A1):
  - Latitude coefficients reported as: 0.002; 0.002; -0.001; 0.002; -0.001; -0.002; -0.002 with corresponding standard errors.
  - Jusoli reported coefficients: 0.678; 0.614; 0.299; 0.680; 0.542; 0.283 with [0.204]***; [0.195]***; [0.160]*; [0.202]***; [0.202]***; [0.158]* standard errors respectively.
  - Rule of law reported coefficients: 0.609; 0.647; 0.162; 0.539; 0.373; 0.051; 0.059 with [0.114]***; [0.131]***; [0.131]; [0.131]***; [0.144]**; [0.128]; [0.127] standard errors respectively.
  - Education (log) reported as: 0.956; 0.892; 0.902 with [0.097]***; [0.108]***; [0.107]*** standard errors.
  - Observations across columns: 105; 105; 98; 100; 99; 91; 91. R-squared values: 0.28; 0.28; 0.60; 0.31; 0.36; 0.65; 0.65.
- First-stage IV details (Appendix Table A2):
  - British legal origin: 0.163 [0.072]** (column (1)).
  - French legal origin: 0.269 [0.056]*** (column (1)).
  - Citizenship law of colonial power — Mixed regime: 0.108 [0.065]* (column (1)).
  - Jus soli (historic): 0.487 [0.064]*** (column (1)).
  - Observations: 242; 238. R-squared: 0.11; 0.26. Sargan-Hansen (p-value): 0.70; 0.38.

### Synthetic Control Method (SCM) Results
- SCM sample and setup:
  - Sample restricted to 26 countries that had jus soli under colonial rule and had sufficient data; 15 of these maintained jus soli to 2014; the rest shifted to jus sanguinis.
  - Pooled SCM applied; treated-year assumed 1970.
- SCM headline:
  - Pooled SCM median treatment effect: moving away from jus soli to jus sanguinis implies a 37.8 log point loss in income per capita by 2014 (about a 46 percent decline).
  - This magnitude is close to the OLS 34 log point estimate.
- Heterogeneity and country-specific SCM findings:
  - Out of 11 treated units analyzed individually, 8 had the expected negative treatment effect.
  - Per-country losses for those 8 ranged from a 137-log point loss (The Gambia) to a 17-log point loss (Nigeria) in 2014.
  - Three treated countries outperformed their synthetic counterfactuals (possible idiosyncratic shocks or imperfect matching).
- Placebo tests:
  - Placebo SCM runs on donor pool units largely show lack of strong treatment effects, supporting statistical significance of actual treated-unit effects (with exceptions aligned to the three counterintuitive treated cases).
- Appendix Figures:
  - Appendix Figure A1 plots Real GDP per Capita (log) for Treated and Control units, 1960–2014 for Bangladesh; Fiji; Gambia; India; Kenya; Malawi; Malaysia; Namibia; Nigeria; Tanzania; Zimbabwe.
  - Appendix Figure A2 shows placebo tests — treatment effect time series for treated units versus donors, 1960–2014.

### Policy Implications and Recommendations
- Core implication:
  - In developing countries, especially where institutions are weak, more inclusive citizenship laws (jus soli) are associated with higher income per capita; switching from jus sanguinis to jus soli can positively contribute to development, ceteris paribus.
- Nuances for policymakers:
  - Institutional context matters: stronger rule of law can dampen or mediate effects of citizenship regimes.
  - Mixed regimes: may not deliver higher income per capita relative to jus sanguinis if acquisition by birth is subject to stringent requirements.
  - Complementary reforms: easing citizenship via marriage and naturalization can offset some income loss from restrictive birthright regimes.
  - Country specificity: historical legacies, ethnic mixes, and institutional quality vary—policy change should account for local context and potential transitional risks.
- Suggested further research:
  - Disaggregate channels: impact on investment, public sector efficiency, labor market segmentation.
  - Institutional interactions: interplay between citizenship law and family law, commercial law, labor regulation, and how these condition economic outcomes.

### Key Quantitative Facts and Statistics (preserved exactly)
- Time coverage used in many analyses: 1960–2014; dependent-level period 1970–2014 for main OLS.
- Main cross-country sample size: 105 developing countries.
- Panel sample and observations: Number of countries 97; observations 331 in panel analyses.
- OLS headline: 34 log point increase in income per capita associated with shift to jus soli (about a 40½ percent increase).
- SCM headline: 37.8 log point loss by 2014 for median treated unit that moved from jus soli to jus sanguinis (about a 46 percent decline).
- SCM country-specific loss range (for 8 treated units with expected sign): from a 137-log point loss (The Gambia) to a 17-log point loss (Nigeria) in 2014.
- Appendix Table A1 reported observations and R-squareds: Observations: 105; 105; 98; 100; 99; 91; 91. R-squared: 0.28; 0.28; 0.60; 0.31; 0.36; 0.65; 0.65.
- Appendix Table A2 reported observations and diagnostics: Observations: 242; 238. R-squared: 0.11; 0.26. Sargan-Hansen (p-value): 0.70; 0.38.

*Source: wp1903*

### References .............................................................................................................

### References

### Figures
- 1. Citizenship Laws, 2014 ............................................................................................................................................. 10
- 2. Real GDP Per Capita in Jus Soli and Non-Jus Soli Developing Countries, 1970–2014 ................... 15
- 3. Synthetic Control Method: Treatment Effect of Moving from a Jus Soli Law to a Jus Sanguinis       
Law. ...................................................................................................................................................................................... 26

### Tables
- 1. Citizenship and Development: OLS Results ..................................................................................................... 18
- 2. Citizenship and Development: Panel and Instrumental Variable Estimations. .................................. 19
- 3. Citizenship and Development:  Alternative Definition of Citizenship Law .......................................... 21
- 4. Citizenship and Development: Additional Control Variables ................................................................... 23
- 5. Dynamics of Citizenship Laws in Developing Countries, 1948 vs. 2014 ............................................... 24

### Appendix Tables
- A1. Citizenship and Development: OLS Results .................................................................................................. 32
- A2. First-Stage Regressions to Instrument the Jus Soli Variable .................................................................. 33

*Source: wp1903 - References*

### Appendix Figures

### Appendix Figures

### Introduction and Research Question
- Focus: How citizenship laws (jus soli vs jus sanguinis vs mixed) affect economic development, particularly in developing countries.
- Hypothesis: Jus soli (inclusive) laws facilitate integration and are associated with higher GDP per capita; jus sanguinis (exclusive) laws hinder integration and growth, especially where institutions are weak.
- Sample focal point: Main analysis restricted to a sample of 105 developing countries (data on citizenship laws and covariates available).

### Citizenship Law Concepts, History, and Mechanisms
- Definitions and regimes:
  - Jus soli: citizenship by birth on the country’s soil (inclusive).
  - Jus sanguinis: citizenship by blood/parents (exclusive).
  - Mixed regime: elements of both; increasing convergence over time.
- Historical patterns:
  - Western Hemisphere (e.g., United States, Canada): prevalent jus soli.
  - Continental Europe, parts of Asia and Africa: historically jus sanguinis; shifts observed (e.g., Germany reform in 1999).
  - Colonial legacies: many colonies inherited colonizer citizenship regimes; major post-independence switches documented.
- Mechanisms by which citizenship regimes affect development:
  - Distorting investment: lack of citizenship leads to short investment horizons, asset diversion overseas, avoidance of specialized capital investment.
  - Political instability and corruption: marginalized non-nationals may induce instability, higher military/public security spending, or encourage rent-seeking.
  - Public sector efficiency: ethnic/identity fragmentation linked to lower public good provision and weaker institutions.
  - Labor market distortions: exclusion from professions, restrictions on ownership (e.g., land) reduce labor market efficiency.
  - Revenue effects: some states sell citizenship (Economic Citizenship Programs) — noted as a counterexample where inclusiveness can directly raise revenues.

### Data, Model, and Econometric Approach
- Data compilation:
  - Citizenship law dataset assembled from multiple sources (including US government report, UNHCR, Manby (2016), Bertocchi and Strozzi (2010)); characteristics coded as of 2014.
  - Main variable: dummy Jusoli = 1 if country has a jus soli regime, 0 otherwise. Also constructed categorical Citizenship law variable: 1 = jus sanguinis; 2 = mixed; 3 = jus soli.
  - Additional features recorded: citizenship by marriage, gender discrimination in transmission, residency requirements, renunciation requirement, language/income thresholds.
- Sample and covariates:
  - Main empirical sample: 105 developing countries.
  - Core control variables: latitude; rule of law index; education proxied by secondary school enrolment rate; trade openness = (imports + exports)/GDP; financial depth = credit to private sector / GDP.
- Baseline specification (cross-country OLS):
  - y_i = δ0 + δ1 Jusoli_i + θ X_i + ε_i ; dependent variable: level of GDP per capita at end of period (1970–2014).
- Panel strategy:
  - Time split into seven sub-periods of five years; random effects and Hausman-Taylor estimators used because Jusoli is time-invariant (less than 5 percent of sample changed citizenship law during 1975–2014).
  - Instruments: legal origin (British/French) and colonial power citizenship regime used to instrument Jusoli for causal inference.
- Counterfactual approach:
  - Synthetic Control Method (SCM), pooled SCM (Cavallo et al., 2013) applied to countries that had jus soli in the colonial period and later switched to jus sanguinis after independence.
  - Pre-treatment covariates for SCM: latitude, Penn World Tables human capital proxy, population size, real capital stock, real income per capita (levels in 1960, 1964, 1969).
  - Treated-year assumption for SCM: change in citizenship law assumed in 1970 for all treated units (sensitivity checks with 1965 and 1975 produced qualitatively unchanged findings).

### Main Results — OLS, Panel, IV, and Robustness
- Cross-country (OLS) headline result:
  - In the full specification (Table 1, column 7), shifting from a jus sanguinis to a jus soli law is associated with a 34 log point increase in income per capita (about a 40½ percent increase). This represents close to three-fifth of the observed gap between jus soli and non-jus soli countries in 2014.
  - The Jusoli coefficient is statistically significant across specifications. Interaction of rule of law and Jusoli has expected sign (dampening effect of stronger institutions) but lacks conventional significance; rule of law itself often significant.
- Panel estimates:
  - Pooling and random-effect models (Table 2) continue to show a positive and statistically significant effect of Jusoli on log GDP per capita (examples of coefficients: 0.350 in pooling; 0.575 and 0.674 in random-effect variants; 0.739 in a Hausman-Taylor variant). Sample sizes: 331 observations, 97 countries in main panel specifications.
  - Mundlak test flagged correlation of some covariates with country effects; Hausman-Taylor estimation used to relax random-effect assumptions.
- Instrumental variables (IV) and causality checks:
  - IV specifications instrumenting Jusoli with British and French legal origin and with the colonial power’s citizenship regime produce results consistent with causality: Jusoli remains positive and significant (e.g., Hausman-Taylor IV estimates reported).
- Additional robustness checks (Table 4):
  - Inclusion of government consumption ratio to GDP: higher government consumption associated with lower GDP per capita (coefficient -0.167, significant).
  - Ethno-linguistic fractionalization: negative association with GDP per capita (example coefficient -0.008, marginal significance).
  - Protection of property rights: positive association (small but significant coefficient reported).
  - Composite index (Marriage and Naturalization): constructed by PCA from sub-indicators (citizenship by marriage; gender equality in transmission; years of residency; no-renunciation dummy; number of other conditions). The composite index enters positively and significantly; interaction Jusoli*(Marriage and Naturalization) is negative and significant, indicating substitutability across inclusion channels.
- SCM counterfactual and pooled SCM results:
  - Sample restricted to 26 countries that had jus soli under colonial rule and had sufficient data; 15 of these maintained jus soli to 2014, the rest shifted to jus sanguinis.
  - Pooled SCM median treatment effect: moving away from jus soli to jus sanguinis implies a 37.8 log point loss in income per capita by 2014 (about a 46 percent decline). This is close in magnitude to the OLS 34 log point estimate.
  - Country heterogeneity: out of 11 treated units analyzed individually, 8 had the expected negative treatment effect; per-country losses for those 8 ranged from a 137-log point loss (The Gambia) to a 17-log point loss (Nigeria) in 2014. Three treated countries outperformed their synthetic counterfactuals (possible idiosyncratic shocks or imperfect matching).
  - Placebo tests: placebo SCM runs on donor pool units largely show lack of strong treatment effects, supporting statistical significance of actual treated-unit effects (with exceptions aligned to the three counterintuitive treated cases).

### Key Quantitative Facts and Statistics (preserved exactly as in source)
- Time coverage used in many analyses: 1960–2014; dependent-level period 1970–2014 for main OLS.
- Main cross-country sample size: 105 developing countries.
- Panel sample and observations: Number of countries 97; observations 331 in panel analyses.
- OLS headline: 34 log point increase in income per capita associated with shift to jus soli (about a 40½ percent increase).
- SCM headline: 37.8 log point loss by 2014 for median treated unit that moved from jus soli to jus sanguinis (about a 46 percent decline).
- SCM country-specific loss range (for 8 treated units with expected sign): from a 137-log point loss (The Gambia) to a 17-log point loss (Nigeria) in 2014.
- Table reporting: Table 1 (OLS), Table 2 (Panel and IV), Table 3 (alternative citizenship law definition), Table 4 (additional controls). Note: robustness checks include use of 1990–99 averages and variations in lagging covariates.

### Policy Implications and Recommendations
- Core policy implication:
  - In developing countries, especially where institutions are weak, more inclusive citizenship laws (jus soli) are associated with higher income per capita; switching from jus sanguinis to jus soli can positively contribute to development, ceteris paribus.
- Nuances and considerations for policymakers:
  - Institutional context matters: stronger rule of law can dampen or mediate effects; interaction effects show partial substitutability across components of citizenship law (birthright vs marriage/naturalization).
  - Mixed regimes: many mixed regimes do not necessarily deliver higher income per capita relative to jus sanguinis if acquisition by birth is subject to stringent requirements.
  - Complementary reforms: easing citizenship via marriage and naturalization (composite index) can offset some income loss from restrictive birthright regimes.
  - Country specificity: historical legacies, ethnic mixes, and institutional quality vary—policy change should account for local context and potential transitional risks.
- Suggested avenues for further research:
  - Disaggregate channels: analyze impact on investment, public sector efficiency, labor market segmentation separately.
  - Institutional interactions: study interplay between citizenship law and family law, commercial law, labor regulation, and how these condition economic outcomes.

*Source: wp1903 - Appendix Figures (IMF working paper content extracted from the source PDF).*

### REFERENCES

### wp1903 - REFERENCES

### Key literature and thematic coverage
- Citations span topics including:
  - Synthetic control methods and comparative case studies (Abadie, Diamond, and Hainmueller, 2010; Abadie and Gardeazabal, 2003; Galiani and Quistorff, 2016).
  - Institutions, colonial origins, and law (Acemoglu, Johnson and Robinson, 2001; Acemoglu and Robinson, 2010; La Porta et al., 1998; McArthur and Sachs, 2001).
  - Ethnic diversity, fractionalization, and conflict (Alesina et al., 2003; Chandra, 2001; Collier et al., 2003; Montalvo and Reynal-Querol, 2005; Humphreys, Posner and Weinstein, 2002).
  - Citizenship, migration, and nationality law (Aleinikoff and Klusmeyer, 2000; DeVoretz and Pivnenko, 2008; Manby, 2016; Weil, 2001; Vink and de Groot, 2010).
  - Growth empirics, convergence, and determinants of development (Barro, 1997; Durlauf, 2009; Durlauf, Johnson and Temple, 2005; Henderson, Parameter and Papageorgiou, 2012; Masanjala and Papageorgiou, 2008).
  - Policy-relevant topics: corruption and growth (Mauro, 1995); natural disasters and growth (Cavallo et al., 2013); gender equality investment (Buvinic, Lunde and Sinha, 2010); immigration policy history (Timmer and Williamson, 1998).

### Appendix Table A1 — OLS results: Log of real GDP per capita in 2014
- Dependent variable: Log of real GDP per capita in 2014.
- Columns (1) through (7) report coefficients with robust standard errors in brackets. Selected reported coefficients and statistics:
  - Latitude: 0.002; 0.002; -0.001; 0.002; -0.001; -0.002; -0.002 [standard errors: 0.005; 0.005; 0.004; 0.006; 0.005; 0.004; 0.004].
  - Jusoli: 0.678; 0.614; 0.299; 0.680; 0.542; 0.283 [standard errors: 0.204]***; [0.195]***; [0.160]*; [0.202]***; [0.202]***; [0.158]*.
  - Citizenship law (column with single coefficient): 0.179 [0.082]**.
  - Rule of law: 0.609; 0.647; 0.162; 0.539; 0.373; 0.051; 0.059 [standard errors: 0.114]***; [0.131]***; [0.131]; [0.131]***; [0.144]**; [0.128]; [0.127].
  - (Rule of law)*Jusoli: -0.148 [0.283].
  - Education (log): 0.956; 0.892; 0.902 [standard errors: 0.097]***; [0.108]***; [0.107]***.
  - Trade openness (log): 0.328; 0.409; 0.441 [standard errors: 0.216; 0.183]**; 0.181]**.
  - Private sector credit ratio (log): 0.409; 0.043; 0.046 [standard errors: 0.125]***; [0.107]; [0.108].
  - Constant terms by column: 7.813; 7.836; 4.356; 6.444; 6.599; 2.729; 2.301 [all with reported standard errors and significance indicated].
  - Observations: 105; 105; 98; 100; 99; 91; 91.
  - R-squared: 0.28; 0.28; 0.60; 0.31; 0.36; 0.65; 0.65.
- Notes:
  - Robust standard errors in brackets; significant at 10%; ** significant at 5%; *** significant at 1%.
  - Education proxied by the secondary school enrolment rate.
  - Jusoli is a dummy variable taking 1 when a jus soli law is applicable, and zero otherwise.
  - Citizenship law categorical: 1 if jus sanguinis, 2 if mixed, 3 if jus soli.
  - Rule of law, education, private sector credit ratio and trade openness are averaged over 1990-1999.

### Appendix Table A2 — First-stage regressions to instrument the Jus Soli variable
- Dependent variable: Jus Soli Dummy variable. Columns (1) and (2) shown (Pooling).
- Reported coefficients and statistics:
  - British legal origin: 0.163 [0.072]** (column (1)).
  - French legal origin: 0.269 [0.056]*** (column (1)).
  - Citizenship law of the colonial power before the country’s independence — Mixed regime: 0.108 [0.065]* (column (1)).
  - Jus soli (historic): 0.487 [0.064]*** (column (1)).
  - Latitude: -0.002; 0.003 [0.002]; [0.002].
  - Rule of law: -0.036; -0.085 [0.061]; [0.060].
  - Education (log): 0.176; 0.121 [0.051]***; [0.050]**.
  - Trade openness (log): 0.022; 0.089 [0.035]; [0.040]**.
  - Private sector credit ratio (log): 0.028; 0.024 [0.042]; [0.041].
  - Constant: -0.749; -0.963 [0.245]***; [0.236]***.
  - Observations: 242; 238.
  - R-squared: 0.11; 0.26.
  - Sargan-Hansen (p-value): 0.70; 0.38.
- Notes:
  - Robust standard errors in brackets; significant at 10%; ** significant at 5%; *** significant at 1%.
  - Education proxied by secondary school enrolment rate.
  - Jusoli is a dummy variable taking 1 when a jus soli law is applicable, and zero otherwise.
  - Rule of law, education, private sector credit ratio and trade openness are lagged one period.

### Appendix Figures A1 and A2 — Synthetic Control Method (SCM) results and placebo tests
- Appendix Figure A1: SCM Results — Real GDP per Capita (log) in Treated and Control Units, 1960–2014.
  - Treated units plotted with their synthetic control counterparts for countries: Bangladesh; Fiji; Gambia; India; Kenya; Malawi; Malaysia; Namibia; Nigeria; Tanzania; Zimbabwe.
  - Vertical axis labeled "Real income per capita (log)" with tick labels "6 7 8 9 1 0" consistent with figure presentation; horizontal axis years 1960–2014.
  - Source: Authors’ calculations.
- Appendix Figure A2: Placebo Tests — Treatment effect for Treated Units and Donor Pool, 1960–2014.
  - Panels show treatment effect time series for each treated unit versus donors for the same country set as Figure A1.
  - Vertical axis tick labels include negative and positive values shown as "- 1 - .5 0 .5 1" consistent with figure presentation.
  - Source: Authors’ calculations.

### Country samples
- Country sample for the Cross-Section and Panel Regressions (grouped by citizenship regime):
  - Jus soli (selected list): Argentina; Belize; Bolivia; Brazil; Costa Rica; Cuba; Dominican Republic; Ecuador; El Salvador; Equatorial Guinea; Grenada; Guatemala; Guinea-Bissau; Guyana; Honduras; Jamaica; Lesotho; Mauritius; Mexico; Nepal; Nicaragua; Niger; Pakistan; Panama; Paraguay; Samoa; St. Lucia; St Vincent and the Grenadines; Venezuela, RB; Zambia.
  - Mixed regime (selected list): Benin; Burkina Faso; Cabo Verde; Cambodia; Cameroon; Central African Republic; Chad; Colombia; Congo, Rep.; Egypt, Arab Rep.; Gabon; Ghana; Guinea; Iran, Islamic Rep.; Mali; Mauritania; Peru; Rwanda; Sao Tome and Principe; Senegal; Sierra Leone; South Africa; Togo; Tunisia; Uganda; Vanuatu.
  - Jus sanguinis (selected list): Algeria; Angola; Bangladesh; Bhutan; Botswana; Burundi; Comoros; Cote d'Ivoire; Djibouti; Dominica; Eritrea; Ethiopia; Fiji; Gambia, The; Haiti; India; Indonesia; Iraq; Jordan; Kiribati; Lao PDR; Lebanon; Liberia; Libya; Madagascar; Malawi; Malaysia; Maldives; Micronesia, Fed. Sts.; Mongolia; Myanmar; Namibia; Nigeria; Papua New Guinea; Philippines; Solomon Islands; Sri Lanka; Sudan; Suriname; Swaziland; Tanzania; Thailand; Tonga; Vietnam; Yemen, Rep.; Zimbabwe.
- Country sample for the Synthetic Control Method (Citizenship Law in 2014 and Citizenship law in 1948 (Jus soli) groups):
  - Listed under Citizenship law in 1948 (Jus soli): Argentina; Brazil; Costa Rica; Dominican Republic; Ecuador; Guatemala; Honduras; Lesotho; Mexico; Nepal; Nicaragua; Pakistan; Panama; Paraguay; Zambia.
  - Listed under Citizenship Law in 2014 with Jus soli and Jus sanguinis treated units: Bangladesh; Fiji; Gambia; India; Kenya; Malawi; Malaysia; Namibia; Nigeria; Tanzania; Zimbabwe.

*Source: wp1903 - REFERENCES (IMF working paper appendix and reference list).*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2019/wp1903.pdf_