On April 22, 2020, the World Bank forecast the steepest decline in global remittances in modern history: a 20% collapse in flows to low- and middle-income countries. Remittances to these countries, which had just hit a record $554 billion, would fall to $445 billion. Governments in Manila, Dhaka, Dushanbe and San Salvador were told to prepare for a fiscal shock without precedent.
It didn’t happen. Officially recorded flows to those countries in 2020 came in at $540 billion, just 1.6% below a revised 2019 total of $548 billion.
The forecast missed the actual outcome by more than 18 percentage points. By 2021, remittances had recovered fully and were growing again, by 8.6%. Remittances had held up under conditions that the models said should have crushed them.
Most economists made the same forecast for the same reasons, so the miss was no lapse of competence at the World Bank. The gap between prediction and outcome says something about what remittances are and how they work. Investment capital reprices when sentiment shifts, aid can be redirected when political conditions change, and foreign direct investment stops when expected returns fall. Remittances express an obligation between a worker abroad and a family at home, carried by nothing more than wire-transfer infrastructure and the desire to help, and an obligation to family is harder to model than any of those.
In 2024, migrants sent home about $905 billion worldwide, up from $865 billion a year earlier.
Of that, $685 billion went to low- and middle-income countries: more than those countries received in foreign direct investment and official aid put together, and roughly three times everything the rich donor countries spend on aid in a year.
No Bretton Woods conference designed this system, and no treaty or multilateral body decides where it goes. The private decisions of some 200 million migrant workers add up to the largest source of outside money that low- and middle-income countries receive. A worker in Riyadh decides how much of this month’s wages to send to Lahore; a construction worker in Moscow calculates what his family in Dushanbe needs before rent is due in two weeks.
That a system this large and this resilient grew up without any central design explains both why it works and why it can’t, by itself, do what its scale implies it should.
The flow nobody designed
The numbers describing remittance growth since 1990 come with a caveat. Global flows rose from under $70 billion in 1990 to $905 billion in 2024, roughly thirteen-fold in nominal terms. That trajectory is real but hard to read: much of the early growth was old money becoming visible rather than new money being sent. Better banking infrastructure, mobile money platforms and reporting requirements formalised billions of dollars that used to move through hawala networks, cash in envelopes and trusted intermediaries, and were never counted. The 1990 figure is an undercount, perhaps severely so; the 2024 figure is more nearly complete.
So the growth story is partly an accounting story, and mixing the two hides what is happening. The flow is large and has grown; it has also always been larger than the historical data suggested.
What remains is still extraordinary. The biggest recipients in 2024 were India, with about $129 billion, and Mexico, with about $65 billion, followed by China, the Philippines and Pakistan. India’s remittances exceed the foreign direct investment it attracts; Mexico’s dwarf what it earns from foreign tourism. For this argument, though, the more instructive cases are small economies where remittances make up a huge share of national income, places where the economy is built around the money.
In Tajikistan, remittances were worth about 48% of GDP in 2024, the highest ratio of any country on earth and up from 39% a year earlier. In El Salvador they were 24% of GDP, with roughly 92% of the money sent from the United States. Lebanon received $5.8 billion in 2024, 17.7% of nominal GDP, much of it sent by Lebanese working in the Gulf.
In these economies remittances are among the largest sources of outside income, and household consumption, the exchange rate, the trade balance and, through the taxes on what households spend, government finances all organise around them.
The institutional vacuum around this system is built into it. Because each transfer is a private transaction between two individuals, remittances sit outside the mechanisms through which states allocate capital. Governments license and supervise the companies that move the money, and international principles set out how those services should work, but none of that decides where a transfer goes or what it is spent on. There is no remittance equivalent of the OECD’s aid committee, no body setting conditions and no minister who can redirect the money toward priority sectors. That absence explains the effectiveness: the money goes directly to the person who needs it, without passing through the ministries where, in many of these countries, it would shrink. It also sets the limits: a system with no central allocator cannot be steered toward national priorities, or made to do things its participants don’t choose to do.
The measurement problem
A Ghanaian nurse in London who taps a phone app to top up her mother's mobile wallet in Accra shows up in today's statistics; the same money handed to a trusted courier in 1995 often showed up nowhere. Nobody can say exactly how much of the growth since 1990 is new money and how much is money that was always sent but never counted. The trend is real, but less steep than the raw numbers suggest.How the money moves — and what it costs
Between the worker in Moscow and the family in Dushanbe sits a financial industry that takes a cut of every transfer. In the third quarter of 2025, the global average cost of sending $200 home was 6.36%. The channel makes a big difference: banks charge an average of 14.99%, while digital-only money transfer operators come in at 3.54%, roughly a quarter as much.
Apply that rate to the $685 billion sent to low- and middle-income countries and the fees run to tens of billions of dollars a year. That is a rough figure, since bigger transfers cost less per dollar. It is also money that, in the absence of policy failure, should not exist.
Regional variation in these costs is stark. South Asia, long the cheapest region to send money to, saw prices creep back up during 2025 and lost that place to the region spanning the Middle East, North Africa, Afghanistan and Pakistan. Sub-Saharan Africa remains the most expensive, and most of the handful of corridors worldwide where fees swallow more than a fifth of the transfer start there.
For the families at the receiving end of those corridors, the fee comes straight out of what arrives. Hawala, the informal, trust-based transfer network that has operated in corridors from the Gulf to Pakistan and Egypt for centuries, shows what those costs push people toward.
A worker in Dubai hands cash to a hawaladar, a money broker, who contacts a counterpart in Karachi, who delivers the same sum minus a small fee to the recipient family. The brokers settle up through coded ledgers, reverse trades and the network’s reputational economy, offsetting transactions against each other over time. No money physically crosses a border.
Hawala is not criminal by nature. It fills a gap where formal banking costs too much, moves too slowly or doesn’t reach the rural areas where recipient families live. After September 11, anti-money-laundering and counter-terrorist-financing rules drove it underground in many jurisdictions, treating informal transfer networks as presumptively suspect. Hawala survived, more opaque and harder to monitor than before, because the formal alternative still costs too much.
SDG 10.c: the commitment nobody met
Sustainable Development Goal target 10.c, adopted at the UN General Assembly in September 2015, reads: "By 2030, reduce to less than 3 per cent the transaction costs of migrant remittances and eliminate remittance corridors with costs higher than 5 per cent." The global average as of Q3 2025 is 6.36% — with the deadline still ahead, even the average sits above the 5% line that every corridor is meant to be under by 2030.
Digital-only providers operate at 3.54% globally, and some corridors already cost less than 3%, so cheap transfers are technically possible. Prices stay high because of market structure: banking incumbents with remittance operations keep their pricing power in corridors where they face limited competition, and they lobby effectively against regulatory changes that would let digital challengers in faster. Recipient-country governments, the parties with the clearest interest in lower costs, have the least leverage at the table: they don't host the sending institutions and have no meaningful jurisdiction over their pricing. The G8 promised in 2009 to bring the average down to 5% within five years. The average missed that deadline by roughly three percentage points and has never reached 5% since; the 3% goal for 2030 is heading for a wider miss.What the evidence shows
Remittances are countercyclical to the receiving economy: when drought, a hurricane or a currency crisis hits the family at home, the money tends to rise. They are also remarkably stable. In the global financial crisis of 2008-09, foreign direct investment fell by more than 40% from peak to trough, while remittances to developing countries dipped by about 5%. When COVID-19 hit and every institutional forecast predicted collapse, remittances fell by 1.6%. Part of the mechanism is simple: migrants increase transfers when their families face hardship, drawing on savings or cutting their own consumption to do so. In 2020 other forces helped too: stimulus spending left host economies in better shape than expected, and money that once moved as cash or through informal channels shifted into recorded ones.
Countercyclicality is real, and it has a limit. It applies when shocks originate in the receiving country. When shocks originate in the sending country, the mechanism inverts. Tajikistan shows how: when the Russian economy contracts, its labour market tightens, or the ruble loses purchasing power against major currencies, the real value of what Tajik workers can send home falls with it. The risk recurs, because so much of the economy’s income depends on what migrants earn in Russia. The insurance that remittances provide runs in one direction only, and it weakens when both economies are stressed at once. More broadly, across twenty economies, about two-thirds of the swings in remittances follow a single global cycle. When the world economy slows and credit tightens, remittance growth tends to slow in many countries at once.
The poverty evidence is substantial. Across 33 African countries between 1990 and 2005, a 10% rise in remittances as a share of GDP went with a drop of almost 3% in both the share of people living in poverty and how deep that poverty runs. The same pattern held among 38 of the biggest recipients in Africa, Asia and Latin America between 1990 and 2021. This body of work suggests that remittances are one of the more powerful poverty-reduction tools available to households in high-remittance economies, more direct than most aid and free of the conditionality that often makes aid less effective than its volume implies.
Education effects are just as well documented. In El Salvador, children in households receiving money from abroad were far less likely to drop out of school, and each remittance dollar did much more to keep them there than a dollar of other income: in the cities, at least ten times as much. That fits the idea that money sent with the explicit purpose of supporting the family gets spent differently from ordinary income. Pooling dozens of studies from 30 countries confirms the pattern: households that receive remittances spend markedly more on schooling, though not everywhere. In Eastern Europe and East Asia the effect disappears, and elsewhere its size varies by country, household structure and migration pattern.
The likeliest common thread in these findings is directness. Much official aid passes through donor-country bureaucracies, recipient-country ministries, implementing partners and local contractors before it reaches households, with room for inefficiency and capture at each stage. Remittances skip all of that. They arrive as cash or mobile money to a specific person, who spends it on her family’s specific needs. The same directness limits them, because money that goes straight to households passes through nothing designed to turn it into public institutions.
The person you cannot replace
Remittances work because a worker has left, and the same departure creates the brain drain: every transfer home depends on someone who is no longer there to work.
Sub-Saharan Africa has some of the world’s lowest physician-to-population ratios. It also trains doctors at public expense, at institutions whose operating costs are partly funded by foreign aid, and then exports those doctors to the NHS, to Canadian hospitals and to American health systems. More than 3,000 Ghanaian health professionals left for the United Kingdom between 2018 and 2021 alone, and the drain is old: of the doctors who graduated in Ghana between 1985 and 1994, around three in five went on to work abroad. Ghana still has roughly one doctor for every 4,000 people, despite sustained public investment in medical education.
Consider a doctor Ghana trains over seven or more years, at the expense of Ghanaian taxpayers. She emigrates and works in a British NHS hospital that would otherwise have had to pay to train a domestic replacement. She sends home $500 a month. That is real money and it helps a real family, but it does not, by itself, train another doctor, staff the clinic she left or lighten the load on the colleagues who stayed. For the household, the calculation is rational and beneficial. For the country, it is a subsidy from one of the world’s poorest countries to one of the wealthiest, paid in human capital and repaid, fractionally, in wire transfers.
Whether skilled emigrants make up the difference in cash is disputed. Across a large panel of developing countries, a bigger brain drain goes with smaller remittance flows, a pattern put down to skilled migrants staying abroad longer and bringing their families with them. Surveys of migrants in eleven rich destination countries point the other way: the more educated send more money home, mainly because they earn more.
There is a serious counterargument: the prospect of emigration can induce more people to acquire qualifications than would otherwise do so. If some of those people subsequently don’t emigrate, or return after building skills abroad, the sending country ends up with a larger educated workforce than it would have had without migration as an option. This is the “brain gain” hypothesis, and it has some empirical support.
Brain gain: the real cases and their limits
The Philippines deliberately expanded nursing training for international export from the 1970s onward. The result was a nursing workforce far larger than domestic healthcare needs required, with a pipeline of trained professionals available both for export and for domestic service. Remittances from Filipino nurses abroad fund household consumption; the domestic healthcare system benefits from a larger remaining pool than it would have had otherwise. India's IT sector presents a similar dynamic: emigration of software engineers to Silicon Valley in the 1990s and 2000s created networks that later facilitated return migration and the growth of India's domestic technology industry.
These are real cases of brain gain, but they do not generalise. The Philippines case depended on deliberate state industrial policy and a specific feature of nursing training (rapid expansion is possible). India's IT case depended on the existence of a rapidly growing domestic sector capable of absorbing returnees. Both countries are middle-income economies with the institutional infrastructure to capture some of the gains from skilled emigration.
The countries where brain drain is most acute, among them sub-Saharan African economies with very few doctors per person and small island states in the Pacific, are the countries least able to run the kind of industrial policy that turned emigration into a strategy in Manila or Bangalore. In those countries, a mechanism that reliably benefits households does so partly by removing many of the people best placed to build the institutions those households need.The currency of dependency
Where Dutch disease takes hold, it compounds the brain drain. A country simultaneously losing its most capable workers and pricing its productive sectors out of competition sees its human capital decline at the same time as the economic environment in which that capital might have been deployed deteriorates.
Dutch disease is the mechanism by which large, persistent inflows of foreign currency hollow out the tradeable sector. It was first documented for resource windfalls in the Netherlands in the 1970s, hence the name, but it applies to any large, persistent foreign-currency inflow: oil revenues, aid or remittances. The pathway runs as follows. Foreign-currency inflows increase domestic demand, particularly for non-tradeable goods and services such as housing, construction and local services. Prices in the non-tradeable sector rise. The real exchange rate appreciates. Export sectors such as agriculture and manufacturing face higher costs in domestic currency while earning revenues in foreign currency at a fixed international price. Competitiveness erodes, firms exit or don’t enter, and the tradeable sector shrinks relative to the economy.
The evidence for a remittance-driven version of this mechanism is real but messier than the theory. Across 32 Asian developing countries, more remittances per person went with a stronger real exchange rate, the textbook Dutch disease signature. But the effect showed up only in countries where remittances were small relative to the economy. Where they were large, the currency tended to weaken instead. Dutch disease remains a real risk for remittance economies, but the heaviest recipients are not guaranteed to catch it.
Tajikistan, among the heaviest recipients, shows a different exposure in concentrated form, one that arrives from abroad rather than through a rising currency. With remittances worth nearly half of GDP, sent overwhelmingly by Tajik labour migrants in Russia, the somoni’s real purchasing power is in significant part a function of wage levels on Russian construction sites. When the ruble collapsed in 2014-15, after the oil-price crash and Western sanctions, the money migrants sent home fell by a third in dollar terms in a single year. The official growth figures barely flinched, because the economy kept growing in real terms, but the external shock travelled straight through the remittance channel into household purchasing power.
Where remittances do push the real exchange rate up, the trap reinforces itself. Tradeable sectors face competitiveness pressure that firms in lower-remittance economies don’t face. Agricultural exporters compete at a disadvantage, and manufacturers struggle to reach the cost structures needed to enter export markets. The domestic economy reorients toward consumption and non-tradeable services, the sectors that benefit from remittance spending, and away from the tradeable activities that historically drive structural transformation. Facing few productive jobs at home, more people emigrate. More emigration generates more remittances, which push the exchange rate up further.
Policy responses are constrained in ways unusual for fiscal policy. A government can devalue, but the remittance flow is one of the main drivers of the exchange rate it’s trying to adjust. It can pursue industrial policy to support tradeable sectors, but where the currency has been pushed up, firms in those sectors carry a competitiveness disadvantage that subsidies must more than offset. It can try to steer remittances toward productive investment through diaspora bonds or matched savings programmes, but migrants decide what to do with their money, and the evidence that financial incentives substantially redirect consumption-driven flows toward capital formation is limited. The state is trying to manage the consequences of a flow it didn’t create and can’t stop.
The legislature you cannot vote out
Where Dutch disease takes hold, it narrows the toolkit of an already-constrained state. The sovereignty trap is a separate argument: that the toolkit was never adequate to begin with, and that years of accumulating remittance dependency entrenched its inadequacy. It asks what happens when the most consequential variable in a national economy can be reshaped by a foreign legislature.
Development, as a practical project rather than a rhetorical commitment, requires institutional agency: the capacity to make decisions about allocation, regulation and policy at the national level and have those decisions matter. Countries whose primary economic variable is a private flow determined by wage levels in another country’s economy, subject to that country’s immigration policy and carried by transfer services priced in markets they don’t regulate, have surrendered institutional agency over their core economic function. They surrendered it gradually, through the accumulated logic of millions of rational individual choices and their states’ adaptation to conditions they could not change.
El Salvador is where this argument became concrete and dated. Remittances represent approximately 24% of El Salvador’s GDP, and roughly 92% of them come from the United States, which ties a large share of Salvadoran household income, and the tax revenue raised on its spending, to a single foreign economy. In July 2025, the US Congress passed the One Big Beautiful Bill Act, which includes a 1% excise tax on remittance transfers funded by cash, money orders or similar physical instruments, effective January 1, 2026. The Center for Global Development estimated that a 1% levy on remittances would cost El Salvador the equivalent of 0.6% of its national income; the final law, which taxes only cash-funded transfers, bites on a narrower base.
Nobody in El Salvador voted on this, and no Salvadoran official had a seat at the table or any leverage over the outcome, though remittances are worth nearly a quarter of national income.
The tax has exemptions: transfers funded by debit or credit cards issued by US financial institutions are excluded, as are transfers from accounts subject to Bank Secrecy Act requirements. The effect will therefore be concentrated on the part of the Salvadoran diaspora that uses cash-funded transfers, which is disproportionately lower-income and less banked, and least likely to have the alternatives that make the exemptions meaningful. Whatever the total, the loss will fall on households already at the margin.
Lebanon’s version of the trap runs through a different but structurally similar dependency. The country received $5.8 billion in remittances in 2024, 17.7% of nominal GDP. A large share of that money is earned in the Gulf, which makes Lebanon’s remittance income sensitive to GCC labour market policies, oil price cycles and the labour-nationalisation drives, such as Saudisation and Emiratisation, through which Gulf governments periodically cut their reliance on foreign workers. Beirut doesn’t set those policies or negotiate them; it adapts to their consequences.
The banking crisis that erupted in 2019-20 reordered Lebanon’s remittance flows in ways that reveal the country’s condition. When the formal banking system collapsed, as depositors lost confidence and the sector ran up more than $72 billion in losses, transfers shifted decisively to informal channels. The Lebanese diaspora effectively became the country’s offshore balance sheet. When the lira lost more than 98% of its value and formal savings were trapped in banks that wouldn’t return them, remittances were how middle-class families kept up their consumption. The money kept households afloat inside the political order whose failures had made it necessary.
Tajikistan’s sovereignty trap is the starkest. With remittances worth 48% of GDP and sent overwhelmingly from Russia, the country’s income statement is a derivative of Russian economic conditions. When the ruble falls, Tajik household incomes fall in real terms; when Russian construction, the sector that employs the most Tajik migrants, slows down, so do remittances. The country has no adequate policy lever in response, and no treaty Dushanbe can invoke when Moscow’s economy changes. Nobody planned the dependency; it grew out of the economy’s structure, which is what makes it durable.
Tajikistan, El Salvador and Lebanon did not design remittance dependency as a development strategy. They are countries where institutional failure made emigration individually rational, where those emigrants sent money home because they were obligated to and because it worked, and where the real household-level effectiveness of the mechanism created a path dependency that now organises economic life.
At each step, the individual decisions were rational. Added up, they produced something that looks like a development strategy without being one, and calling it a strategy implies agency these states don’t have. They are administering the consequences of decisions made by hundreds of thousands of their citizens and by the governments of the countries those citizens moved to.
What the wire transfer cannot do
In 2020, when the forecasts said remittances would fall by 20%, they fell by 1.6%.
In a straightforward sense, that flow of money is everything development economics spent seventy-five years trying to build: direct and targeted, with no conditionality, programme cycle, implementing partner or ministerial signature, and none of the openings for capture that can drain money moving through the formal aid architecture. The money goes from a specific person to a specific family and is spent on what that family needs. At that level, it works: poverty depth falls, school dropout rates fall, households survive shocks that destroy other mechanisms.
That kind of effectiveness has limits. It cannot bring back the doctor who left, restore the exchange-rate competitiveness that Dutch disease may have eroded, or vote on the legislation that taxes the wire transfer itself.
Nor can it substitute for the institutions that would turn remittance-funded household consumption into structural economic transformation: banking systems that mobilise savings, regulatory environments that attract investment, a civil service able to administer industrial policy, a tax base to fund the next generation of public education and healthcare.
The countries most dependent on remittances, among them Tajikistan, El Salvador and a structurally paralysed Lebanon, are the countries where those institutions are weakest.
Many of the people best placed to build those institutions have already left, and they are sending money home. That money does real good at the household level, but it cannot do the building for them.
The question a serious development economics has to sit with is how those institutions get built when the people who would build them are earning their wages abroad.
Gen AI Disclaimer
Some contents of this page were generated and/or edited with the help of a Generative AI.
Media
Photo of Kutapalong Refugee Camp in Bangladesh. – Wikipedia
Key Sources and References
World Bank, “World Bank Predicts Sharpest Decline of Remittances in Recent History,” Press Release, April 22, 2020, https://www.worldbank.org/en/news/press-release/2020/04/22/world-bank-predicts-sharpest-decline-of-remittances-in-recent-history
World Bank, “Defying Predictions, Remittance Flows Remain Strong During COVID-19 Crisis,” Press Release, May 12, 2021, https://www.worldbank.org/en/news/press-release/2021/05/12/defying-predictions-remittance-flows-remain-strong-during-covid-19-crisis
World Bank, “Remittances to Reach $630 billion in 2022 with Record Flows into Ukraine,” Press Release, May 11, 2022 — 2021 growth of 8.6%, https://www.worldbank.org/en/news/press-release/2022/05/11/remittances-to-reach-630-billion-in-2022-with-record-flows-into-ukraine
Ratha, Dilip, Plaza, Sonia, and Kim, Eung Ju, “In 2024, remittance flows to low- and middle-income countries are expected to reach $685 billion, larger than FDI and ODA combined,” World Bank People Move blog, December 18, 2024 — 2024 figures: global remittances $905 billion, LMICs $685 billion, https://blogs.worldbank.org/en/peoplemove/in-2024–remittance-flows-to-low–and-middle-income-countries-ar
Banco de México, “Ingresos y Egresos por Remesas, diciembre de 2024,” press release, February 2025: 2024 remittance inflows to Mexico $64.7 billion, https://www.banxico.org.mx/publicaciones-y-prensa/remesas/%7BD08E80AC-6031-C7CF-D792-28CBEF70658F%7D.pdf
OECD, “International aid rises in 2023 with increased support to Ukraine and humanitarian needs,” April 11, 2024, https://www.oecd.org/en/about/news/press-releases/2024/04/international-aid-rises-in-2023-with-increased-support-to-ukraine-and-humanitarian-needs.html; overview: https://www.oecd.org/en/topics/official-development-assistance-oda.html
World Bank, Remittance Prices Worldwide, Issue 54, Q3 2025 — global average 6.36%, banks 14.99%, digital-only MTO 3.54%, Sub-Saharan Africa highest-cost region, https://remittanceprices.worldbank.org/sites/default/files/2026-04/RPW_main_report_and_annex_Q325.pdf
Anyanwu, John C., and Erhijakpor, Andrew E. O., “Do International Remittances Affect Poverty in Africa?” African Development Review, Vol. 22, No. 1, March 2010, pp. 51–91, https://doi.org/10.1111/j.1467-8268.2009.00228.x
Ojeyinka, Titus Ayobami, and Ibukun, Cleopatra Oluseye, “Do remittances mitigate poverty? Evidence from selected countries in Africa, Asia and Latin America,” Economic Change and Restructuring, Springer Nature, Volume 57(3), June 2024, https://link.springer.com/article/10.1007/s10644-024-09666-1
Cox Edwards, Alejandra, and Ureta, Manuelita, “International Migration, Remittances, and Schooling: Evidence from El Salvador,” Journal of Development Economics, Vol. 72, No. 2, December 2003, pp. 429–461, https://doi.org/10.1016/S0304-3878(03)00115-9
Askarov, Zohid, and Doucouliagos, Hristos, “A meta-analysis of the effects of remittances on household education expenditure,” World Development, Vol. 129, 2020, https://www.sciencedirect.com/science/article/abs/pii/S0305750X19305091
Nguyen Phuc Hien, Cao Thi Hong Vinh, Vu Thi Phuong Mai and Le Thi Kim Xuyen, “Remittances, real exchange rate and the Dutch disease in Asian developing countries,” Quarterly Review of Economics and Finance, Vol. 77, pp. 131–143, 2020, https://doi.org/10.1016/j.qref.2019.10.006, https://www.sciencedirect.com/science/article/abs/pii/S1062976919300821
Faini, Riccardo, “Remittances and the Brain Drain,” IZA Discussion Paper No. 2155, June 2006, https://docs.iza.org/dp2155.pdf
Bollard, Albert, McKenzie, David, Morten, Melanie, and Rapoport, Hillel, “Remittances and the Brain Drain Revisited: The Microdata Show That More Educated Migrants Remit More,” World Bank Economic Review, Vol. 25, No. 1, 2011, pp. 132–156; working-paper version, World Bank Policy Research Working Paper 5113, November 2009, https://documents1.worldbank.org/curated/en/676331468170692741/pdf/WPS5113.pdf
Wilson Center, “Africa’s Healthworker Brain Drain,” blog post, https://www.wilsoncenter.org/blog-post/africas-healthworker-brain-drain [secondary source citing UK parliamentary research on Ghanaian health professional migration]
Rural and Remote Health, article 1898 (Ghana medical graduate emigration), https://www.rrh.org.au/journal/article/1898
Monterroso, Oscar, and Vilan, Diego, “Global Remittances Cycle,” FEDS Notes, Board of Governors of the Federal Reserve System, February 27, 2025, https://www.federalreserve.gov/econres/notes/feds-notes/global-remittances-cycle-20250227.html
Internal Revenue Service, “Treasury, IRS issue proposed regulations on the new remittance transfer tax established under the One, Big, Beautiful Bill,” IR-2026-48, April 10, 2026, https://www.irs.gov/newsroom/treasury-irs-issue-proposed-regulations-on-the-new-remittance-transfer-tax-established-under-the-one-big-beautiful-bill
Center for Global Development, “Even at 1 Percent, the US Remittance Tax Hits Poor Countries Hard,” blog post, July 1, 2025, https://www.cgdev.org/blog/even-1-percent-us-remittance-tax-hits-poor-countries-hard
L’Orient Today, citing World Bank estimates, “World Bank: Lebanon received $5.8 billion in remittances in 2024,” February 28, 2025, https://today.lorientlejour.com/article/1449753/world-bank-lebanon-received-58-billion-in-remittances-in-2024.html; the figures above are the World Bank series used for the other country shares in this article. Banque du Liban’s own balance-of-payments count, which measures workers’ remittance inflows differently, is higher: $6.8 billion in 2024 (Banque du Liban, “Balance of Payments and International Investment Position – 2024,” May 2025, pp. 9–10), https://www.bdl.gov.lb/CB%20Com/Publications/Publications/Annual%20Report_2024_En%C2%A710530_3.pdf
World Bank, “Tajikistan Shows Strong Economic Momentum but Lags Regional Peers on Digitalizing Trade,” press release, July 30, 2025, https://www.worldbank.org/en/news/press-release/2025/07/30/tajikistan-economic-update-2025
World Bank, “World Bank discusses the impact of economic slowdown on households in Tajikistan,” press release, December 2, 2015, https://www.worldbank.org/en/news/press-release/2015/12/02/world-bank-discusses-the-impact-of-economic-slowdown-on-households-in-tajikistan
Our World in Data, “Tajikistan’s remittances are worth nearly half the country’s GDP,” 2025, https://ourworldindata.org/data-insights/tajikistans-remittances-are-worth-nearly-half-the-countrys-gdp
UNCTAD, World Investment Report 2010: Investing in a Low-Carbon Economy, United Nations Conference on Trade and Development, 2010, https://unctad.org/publication/world-investment-report-2010
RSM US, “One Big Beautiful Bill Act imposes 1% excise tax on cross-border remittances,” 2025, https://rsmus.com/insights/tax-alerts/2025/excise-tax-on-cross-border-remittances.html
Committee on Payment and Settlement Systems and World Bank, “General principles for international remittance services,” Bank for International Settlements, January 23, 2007, https://www.bis.org/publications/general-principles-international-remittance-services
