The Lucas paradox is the empirical regularity, posed by Robert E. Lucas, Jr. in 1990, that international capital does not flow from rich to poor countries as the neoclassical growth model predicts: in cross-country data, the capital a country attracts is strongly and robustly positively associated with its initial GDP per capita. Muhammad Akhtaruzzaman's monograph International Capital Flows and the Lucas Paradox (2019), replicates the leading empirical studies on data extended to 1984-2015 and concludes that none of the standard candidate determinants - institutional quality, capital account openness, human capital, financial development, trade openness, macroeconomic stability, or their interactions - robustly eliminates the paradox, and that even corrected marginal-product-of-capital measures fail to reconcile observed flows with measured returns. The available literature confirms that the problem remains open. Institutions-based, return-measurement, financial-openness, and public-flow accounts each explain a part of the pattern or a subsample, but no account is consensual or quantitatively complete, and work published through 2021-2025 still treats the paradox as an unresolved motivating fact. The open problem is a theoretically grounded, causally identified, and quantitatively complete account of the direction of international capital flows - robust across specifications and flow components - or a definitive demonstration that the puzzle's premise is a measurement artifact.

Background

In the one-sector neoclassical growth model, the marginal product of capital (MPK) - the additional output obtained from one more unit of capital - falls as a country accumulates capital. Poor countries, being capital-scarce, should therefore offer higher returns, and frictionless international lending should move capital 'downhill' from rich to poor countries until returns are equalized. Lucas (1990) observed that the magnitude of the implied reallocation is enormous relative to observed flows: given the income gap between the United States and India, even modest frictions should have produced capital movements many times larger than anything in the data. The question he posed - why doesn't capital flow from rich to poor countries? - became known as the Lucas paradox.

In the empirical literature the paradox is operationalized as the strong positive cross-country association between a country's initial GDP (gross domestic product) per capita, usually purchasing-power-parity (PPP) adjusted, and the capital it subsequently attracts, measured as average inflows per capita of foreign direct investment (FDI, cross-border investment entailing control), portfolio equity, and debt (Alfaro, Kalemli-Ozcan and Volosovych 2008). An explanation of the paradox is conventionally judged by whether controlling for a candidate determinant eliminates the positive coefficient on initial income in these regressions.

Several candidate determinants have been proposed. Institutional quality (rule of law, expropriation risk, contract enforcement) is the leading one: Alfaro, Kalemli-Ozcan and Volosovych (2008), using an instrument constructed from historical European settler mortality among former colonies, report that low institutional quality was the leading explanation of the paradox over 1970-2000. A second route questions the premise of the puzzle: Caselli and Feyrer (2007) compute MPKs using the income share of reproducible (produced) capital rather than total capital, together with price-level corrections, and find that returns are much more similar across rich and poor countries than naive calculations suggest, implying little unexploited return differential. This measurement route is contested: Ferreira (2011) interprets the residual cross-country differences in returns as risk premia, and Akhtaruzzaman (2019) reports that revised MPK estimates are in fact negatively correlated with observed net flows (52 countries, 1995-2007), the opposite of the neoclassical prediction. A third strand documents closely related facts: Prasad, Rajan and Subramanian (2007) document 'uphill' flows from developing to rich countries after the late 1990s; Gourinchas and Jeanne (2013) establish the 'allocation puzzle' - capital tends to flow toward developing countries that invest and grow less - and show that it is driven by public flows, especially international reserve accumulation by governments, while private inflows still go to higher-productivity-growth countries; Alfaro, Kalemli-Ozcan and Volosovych (2014) and Kim and Zhang (2022) further decompose net flows into public and private components and rationalize their different cyclical behavior. Proposed theoretical mechanisms include financial frictions and firm heterogeneity (Ju and Wei 2014, whose non-neoclassical model generates simultaneous financial-capital outflows and FDI inflows for countries with weak financial institutions), and global shortages of safe stores of value (Caballero, Farhi and Gourinchas 2017). Ohanian, Restrepo-Echavarria and Wright (2018) analyze postwar capital flows to Asia and Latin America in a related welfare assessment of global capital misallocation.

Akhtaruzzaman's 2019 monograph replicates the influential cross-sectional, panel fixed-effects, and dynamic panel specifications of the paradox on data extended to 1984-2015, adds an analysis of political risk for FDI, and reviews the marginal-product and net-versus-gross-flow issues. Its conclusion is that explanations in the existing literature are not robust to alternative model specifications and econometric approaches: the coefficient on initial income remains positive and significant in the preferred cross-sections (using a symmetric log-type, inverse-hyperbolic-sine transformation of inflows) with institutional quality (measured by the International Country Risk Guide, ICRG, index of political and economic country risk), capital account openness (measured by the Chinn-Ito index, a standard index of legal restrictions on cross-border financial transactions), human capital, distantness (a country's average distance from world markets), trade openness, financial development, and inflation as controls, and likewise in panel fixed-effects and dynamic panel specifications (the latter estimated by system GMM, a generalized-method-of-moments estimator for dynamic panels), while the book's instrumental-variables (IV) cross-sectional designs themselves suffer weak-instrument problems. Akhtaruzzaman, Hajzler and Owen (2018) reached the same substantive conclusion for institutional quality. The book further flags limitations that delimit what would count as a resolution: reverse causality and omitted time-varying factors in panel models, weaker and less consistent results for portfolio and private debt flows than for FDI, unexplored non-neoclassical mechanisms (increasing returns to scale, knowledge spillovers, economies of scale), geopolitical developments (trade wars, capital-flow reversals) discussed only briefly and without empirical evidence, and the absence of firm-level microfoundations for multinational investment decisions.

Problem Statement

The problem is the one posed by Lucas (1990) and operationalized in the subsequent empirical literature: account for the direction of international capital flows. Specifically, explain why, in cross-country data, the capital a country attracts over subsequent decades - foreign direct investment, portfolio equity, and debt inflows per capita - is strongly and robustly positively associated with its initial GDP per capita (the Lucas paradox), instead of flowing predominantly to capital-scarce countries with higher returns as the neoclassical model predicts.

The problem can be resolved in either of two ways:

(1) A causal, quantitatively complete account: an identified set of determinants - which may include institutional quality, capital account openness, human capital, financial development, risk premia, public-sector saving and reserve accumulation, or mechanisms outside the neoclassical framework such as increasing returns to scale, knowledge spillovers, or financial frictions - whose estimated causal contribution eliminates the positive association between initial income and capital inflows in the standard cross-country and panel specifications. This elimination must be robust across econometric approaches rather than confined to a single estimator or specification; it must either hold across the main components of flows (FDI, portfolio equity, private debt, and official or public flows) or be accompanied by a principled account of why the components differ; and the account's implied return differentials must reproduce the observed pattern of flows, matching not only the direction but also the order of magnitude of the observed association.

(2) A definitive measurement refutation: a demonstration that the puzzle's premise - that there exist large unexploited return differentials favoring poor countries - is an artifact of how returns are measured (for example, of capital-share, price-level, or public-versus-private capital corrections), such that observed flows are in fact consistent with properly measured returns. Because corrected return measures have been reported to correlate negatively with observed flows, such a refutation must reconcile the competing corrections - accounting for why their measured return differentials diverge - rather than re-deriving only one of them.

To eliminate the association, an explanation must be robust across flow types, periods, and specifications, and its identifying assumptions must survive scrutiny. The account may also draw on determinants outside the neoclassical framework or on firm-level evidence on multinational investment decisions.

References

  1. Robert E. Lucas, Jr., 'Why Doesn't Capital Flow from Rich to Poor Countries?', American Economic Review 80(2), 1990, pp. 92-96.
  2. Laura Alfaro, Sebnem Kalemli-Ozcan and Vadym Volosovych, 'Why Doesn't Capital Flow from Rich to Poor Countries? An Empirical Investigation', Review of Economics and Statistics 90(2), 2008, pp. 347-368. DOI: 10.1162/rest.90.2.347, https://doi.org/10.1162/rest.90.2.347
  3. F. Caselli and J. Feyrer, 'The Marginal Product of Capital', Quarterly Journal of Economics 122(2), 2007, pp. 535-568. DOI: 10.1162/qjec.122.2.535, https://doi.org/10.1162/qjec.122.2.535
  4. Eswar Prasad, Raghuram Rajan and Arvind Subramanian, 'Foreign Capital and Economic Growth', Brookings Papers on Economic Activity, 2007, pp. 153-230. DOI: 10.1353/eca.2007.0016, https://doi.org/10.1353/eca.2007.0016
  5. Pierre-Olivier Gourinchas and Olivier Jeanne, 'Capital Flows to Developing Countries: The Allocation Puzzle', Review of Economic Studies 80, 2013, pp. 1484-1515. DOI: 10.1093/restud/rdt004, https://doi.org/10.1093/restud/rdt004
  6. Jiandong Ju and Shang-Jin Wei, 'A Solution to Two Paradoxes of International Capital Flows', Economic and Political Studies 2(1), 2014, pp. 3-43. DOI: 10.1080/20954816.2014.11673850, https://doi.org/10.1080/20954816.2014.11673850
  7. Dennis Reinhardt, Luca Antonio Ricci and Thierry Tressel, 'International Capital Flows and Development: Financial Openness Matters', Journal of International Economics 91(2), 2013, pp. 235-251. DOI: 10.1016/j.jinteco.2013.07.006, https://doi.org/10.1016/j.jinteco.2013.07.006
  8. Laura Alfaro, Sebnem Kalemli-Ozcan and Vadym Volosovych, 'Sovereigns, Upstream Capital Flows, and Global Imbalances', Journal of the European Economic Association 12(5), 2014, pp. 1240-1284. DOI: 10.1111/jeea.12106, https://doi.org/10.1111/jeea.12106
  9. Ricardo J. Caballero, Emmanuel Farhi and Pierre-Olivier Gourinchas, 'The Safe Assets Shortage Conundrum', Journal of Economic Perspectives 31(3), 2017, pp. 29-46. DOI: 10.1257/jep.31.3.29, https://doi.org/10.1257/jep.31.3.29
  10. Sirsha Chatterjee and Kanda Naknoi, 'The Marginal Product of Capital, Capital Flows, and Convergence', American Economic Review 100(2), 2010, pp. 73-77. DOI: 10.1257/aer.100.2.73, https://doi.org/10.1257/aer.100.2.73
  11. Alex Luiz Ferreira, 'On the Differences between the Marginal Product of Capital across Countries', The Manchester School 79, 2011, pp. 455-479. DOI: 10.1111/j.1467-9957.2009.02163.x, https://doi.org/10.1111/j.1467-9957.2009.02163.x
  12. Lee E. Ohanian, Paulina Restrepo-Echavarria and Mark L. J. Wright, 'Bad Investments and Missed Opportunities? Postwar Capital Flows to Asia and Latin America', American Economic Review 108(12), 2018, pp. 3541-3582. DOI: 10.1257/aer.20151510, https://doi.org/10.1257/aer.20151510
  13. Muhammad Akhtaruzzaman, Christopher Hajzler and P. Dorian Owen, 'Does Institutional Quality Resolve the Lucas Paradox?', Applied Economics 50(5), pp. 455-474 (online 2017, print 2018). DOI: 10.1080/00036846.2017.1321840, https://doi.org/10.1080/00036846.2017.1321840
  14. Muhammad Akhtaruzzaman, 'International Capital Flows and the Lucas Paradox', Springer Singapore, 2019. DOI: 10.1007/978-981-13-9069-2, https://doi.org/10.1007/978-981-13-9069-2
  15. Nataliia Osina, 'Global Governance and Gross Capital Flows Dynamics', Review of World Economics 157, 2021, pp. 463-493. DOI: 10.1007/s10290-020-00404-z, https://doi.org/10.1007/s10290-020-00404-z
  16. Hyojin Kim and Hannah Jun, 'Can Blended Finance Be a Game Changer in Sustainable Development? An Empirical Investigation of the Lucas Paradox', Sustainability 14(4), 2022, article 2186. DOI: 10.3390/su14042186, https://doi.org/10.3390/su14042186
  17. Yun Jung Kim and Jing Zhang, 'International Capital Flows: Private Versus Public Flows in Developing and Developed Countries', International Economic Review 64, pp. 225-260 (online 2022, print 2023). DOI: 10.1111/iere.12599, https://doi.org/10.1111/iere.12599
  18. Nombulelo Braiton and Nicholas M. Odhiambo, 'Capital Flows and Institutional Quality: A Systematic Literature Review', Oradea Journal of Business and Economics 9, 2024, pp. 113-123. DOI: 10.47535/1991ojbe186, https://doi.org/10.47535/1991ojbe186
  19. Cathrin Mohr and Christoph Trebesch, 'Geoeconomics', Annual Review of Economics 17, 2025, pp. 563-587. DOI: 10.1146/annurev-economics-092424-111952, https://doi.org/10.1146/annurev-economics-092424-111952