Almost 80 percent of the growth in remittances to developing countries over the past 20 years is an illusion
This page in:
Remittances sent by migrant workers to developing countries have soared in the past two decades. According to the World Development Indicators, workers’ remittances to developing countries were just US$47 billion in 1980 (in constant 2011 dollars). After barely rising by 1990 ($49 billion), they doubled by 2000 ($102 billion), and from there, tripled by 2010 ($321 billion). The World Bank’s prospects group’s latest estimates are that remittances to developing countries reached $404 billion (in nominal terms) in 2013, and are predicted to grow to $516 billion by 2016.
These amounts now greatly exceed the total sum of all aid flows to developing countries, and such big numbers have brought increasing policy attention to the impact of remittances on sending country economies, and various proposals for policy actions to increase the development impacts of remittances. But it has also asked policymakers and some researchers to ask why have such rapid increases in remittances not resulted in noticeable improvements in economic growth in the recipient countries?
In a new working paper with Michael Clemens of the Center for Global Development, we investigate several reasons for the lack of apparent impact of remittances on economic growth. In today’s post I want to focus on one major reason: our claim that growth rates are vastly overstated.
The large overstating of remittance growth
The solid line in the figure below shows macro data on remittances for Mexico. We see steady growth from the mid-1980s to around 2000, then incredibly rapid growth following this. The dotted line shows remittances for Mexico as measured in Mexican household surveys. The micro and macro data track each other very closely in growth rates until 2000/2001, and then diverge from each other, with much faster growth in the macro data in the 2000s.
![]()
We find the same pattern in several other countries. So which of these series is correct?
We note first that the divergence in the two series in Mexico occurs at the same time as a law change affecting reporting in the macro data: before 2002 only commercial banks reported their remittance transactions to Mexico’s central bank, but then in 2002 a new regulation required all money transfer companies to register at the central bank and report their remittance transactions on a monthly basis. So suddenly measured macro remittances soar as a result. This suggests that in this case at least, much of the growth in remittances in the macro data is due to changes in measurement. Is this a general phenomenon?
Understanding macro data on remittances
Macro data on remittances come from balance of payments data provided by each country to the IMF. The underlying data that feeds into these aggregates comes from central banks in each individual country. They typically require banks, money transfer operators and other institutions to provide reports on the transactions they process. But this coverage is often partial, and separating remittance transactions from other financial flows can be challenging. As a result, it has long been recognized that macro data on remittances are noisy, but we haven’t seen any previous attempts that attempt to quantify what this means for growth.
We can write:
Global remittances to developing countries = The number of migrants from developing countries*Average earnings per migrant*share of income each migrant remits.
As a result, if we assume that migrants haven’t changed the share of income they remit (an assumption we provide some empirical evidence for in the paper, and also discuss robustness to), then we can write:
Growth in remittances = Growth in migrants + Growth in Migrant earnings + Growth in migrants*Growth in migrant earnings
The World Bank, UN, OECD, and a number of researchers have made a concerted effort to use national census data from all around the world to construct bilateral matrices of migration stock. We use UN Population Division data to measure growth in the number of migrants. According to these estimates, there were 123.4 million migrants from developing countries in 1990, growing to 185.0 million in 2010, a 49.9 percent increase. This is shown in the first row of Table 1 below, which shows that this growth in migrant stock is only 9 percent of the recorded growth in remittances over this time period.
Data on the earnings of migrants at a global level over time are not available. We therefore proxy the growth in income earned per migrant with the growth in per capita GDP in the main destination country in which migrants from that country reside (showing in the paper that this appears to be a reasonable assumption based on US data on migrant earnings). We then construct a weighted average income growth rate for all developing country migrants, which we estimate to be 47.3 percent over the 1990–2010 period (column 2 in Table 1). Combining this with the growth in migrant stock, and using the above decomposition, this gives a predicted growth in remittances of 120.7 percent over the 1990–2010 period.
As a result we estimate that only 21.7 percent of recorded remittance growth is due to changes in the fundamental drivers of remittances, with almost 80 percent therefore potentially due to changes in measurement. We do a similar exercise for the 15 countries that have had the largest absolute growth in remittances between 1990 and 2010, and likewise show that for most of them, very little of the growth is accounted for by fundamentals.
![]()
If most of the measured growth in remittances at the global level is illusory, resulting from changes in measurement, rather than genuine growth, then we should not expect to see changes in GDP growth coming from it.
Are we now good?
It is frequently claimed that the presence of informal transfers and measurement concerns lead recorded remittances at the macro level to be an undercount of true remittances. The implicit assumption is then that over time we are getting closer to the truth. However, there are also a number of reasons why macro data can overstate remittances, by recording other types of transfers as remittances. For example, the IMF notes that many other types of small transactions such as transfers to family members studying abroad, transfers to travelers undertaking trips, and small trade transactions are often transferred through the same money transfer operators as remittances, and can be difficult to distinguish in the data. Some of these types of transactions are likely to be increasing over time, as the internet facilitates person-to-person trade across borders, and the popularity of study abroad programs grows. Conversations with officials in charge of monitoring remittances also reveal the suspicion that property market booms in some developing countries lead to large surges in measured remittance flows, even though the purpose of these transactions was often for migrants to buy property for themselves. As a result we should not automatically conclude that recorded remittances are necessarily getting closer to the truth in levels over time.
These amounts now greatly exceed the total sum of all aid flows to developing countries, and such big numbers have brought increasing policy attention to the impact of remittances on sending country economies, and various proposals for policy actions to increase the development impacts of remittances. But it has also asked policymakers and some researchers to ask why have such rapid increases in remittances not resulted in noticeable improvements in economic growth in the recipient countries?
In a new working paper with Michael Clemens of the Center for Global Development, we investigate several reasons for the lack of apparent impact of remittances on economic growth. In today’s post I want to focus on one major reason: our claim that growth rates are vastly overstated.
The large overstating of remittance growth
The solid line in the figure below shows macro data on remittances for Mexico. We see steady growth from the mid-1980s to around 2000, then incredibly rapid growth following this. The dotted line shows remittances for Mexico as measured in Mexican household surveys. The micro and macro data track each other very closely in growth rates until 2000/2001, and then diverge from each other, with much faster growth in the macro data in the 2000s.
We find the same pattern in several other countries. So which of these series is correct?
We note first that the divergence in the two series in Mexico occurs at the same time as a law change affecting reporting in the macro data: before 2002 only commercial banks reported their remittance transactions to Mexico’s central bank, but then in 2002 a new regulation required all money transfer companies to register at the central bank and report their remittance transactions on a monthly basis. So suddenly measured macro remittances soar as a result. This suggests that in this case at least, much of the growth in remittances in the macro data is due to changes in measurement. Is this a general phenomenon?
Understanding macro data on remittances
Macro data on remittances come from balance of payments data provided by each country to the IMF. The underlying data that feeds into these aggregates comes from central banks in each individual country. They typically require banks, money transfer operators and other institutions to provide reports on the transactions they process. But this coverage is often partial, and separating remittance transactions from other financial flows can be challenging. As a result, it has long been recognized that macro data on remittances are noisy, but we haven’t seen any previous attempts that attempt to quantify what this means for growth.
We can write:
Global remittances to developing countries = The number of migrants from developing countries*Average earnings per migrant*share of income each migrant remits.
As a result, if we assume that migrants haven’t changed the share of income they remit (an assumption we provide some empirical evidence for in the paper, and also discuss robustness to), then we can write:
Growth in remittances = Growth in migrants + Growth in Migrant earnings + Growth in migrants*Growth in migrant earnings
The World Bank, UN, OECD, and a number of researchers have made a concerted effort to use national census data from all around the world to construct bilateral matrices of migration stock. We use UN Population Division data to measure growth in the number of migrants. According to these estimates, there were 123.4 million migrants from developing countries in 1990, growing to 185.0 million in 2010, a 49.9 percent increase. This is shown in the first row of Table 1 below, which shows that this growth in migrant stock is only 9 percent of the recorded growth in remittances over this time period.
Data on the earnings of migrants at a global level over time are not available. We therefore proxy the growth in income earned per migrant with the growth in per capita GDP in the main destination country in which migrants from that country reside (showing in the paper that this appears to be a reasonable assumption based on US data on migrant earnings). We then construct a weighted average income growth rate for all developing country migrants, which we estimate to be 47.3 percent over the 1990–2010 period (column 2 in Table 1). Combining this with the growth in migrant stock, and using the above decomposition, this gives a predicted growth in remittances of 120.7 percent over the 1990–2010 period.
As a result we estimate that only 21.7 percent of recorded remittance growth is due to changes in the fundamental drivers of remittances, with almost 80 percent therefore potentially due to changes in measurement. We do a similar exercise for the 15 countries that have had the largest absolute growth in remittances between 1990 and 2010, and likewise show that for most of them, very little of the growth is accounted for by fundamentals.
If most of the measured growth in remittances at the global level is illusory, resulting from changes in measurement, rather than genuine growth, then we should not expect to see changes in GDP growth coming from it.
Are we now good?
It is frequently claimed that the presence of informal transfers and measurement concerns lead recorded remittances at the macro level to be an undercount of true remittances. The implicit assumption is then that over time we are getting closer to the truth. However, there are also a number of reasons why macro data can overstate remittances, by recording other types of transfers as remittances. For example, the IMF notes that many other types of small transactions such as transfers to family members studying abroad, transfers to travelers undertaking trips, and small trade transactions are often transferred through the same money transfer operators as remittances, and can be difficult to distinguish in the data. Some of these types of transactions are likely to be increasing over time, as the internet facilitates person-to-person trade across borders, and the popularity of study abroad programs grows. Conversations with officials in charge of monitoring remittances also reveal the suspicion that property market booms in some developing countries lead to large surges in measured remittance flows, even though the purpose of these transactions was often for migrants to buy property for themselves. As a result we should not automatically conclude that recorded remittances are necessarily getting closer to the truth in levels over time.
This paper states the obvious for all of us who deal with remittance numbers on a daily basis. The sudden growth in remittances is indeed explained only by the change in the methodology remittances are recorded. However, due to the lack of the data prior to this change, we still need to deal with what is available. Indeed, with reservation, and no connection to any other factors which affect remittances.
The question is: where do we go from here? Remittance data is still poor, most of the developing countries do not publish any records and we have to rely on estimates and qualitative research. As a market researcher working for money transfer company, I struggle with figuring out the right estimates very frequently. World Bank is doing a great job in recording remittances, however, all countries which do not publish remittance data in their Balance of Payments are excluded. And in some countries, as rightfully pointed by David, remittance numbers are overstated.
Looking at the remittance numbers as a basis for quantifying economic growth or decrease in poverty is approach that should be avoided. Helping countries in figuring out how to better record their remittance numbers should come first.
David,
I'm not clear on why migrants investing in property in their country of origin should be distinct from other transfers. As Michael and I implicitly argue in our paper on rethinking remittances (http://www.financialaccess.org/sites/default/files/publications/migrati…) the idea that there is a fundamental change in how households behave when an income earner crosses a national border may be wrong.
Information asymmetries may get larger and preferences may further diverge, but those may be marginal changes. Therefore, the balance of transfers for consumption and investment are all part of the same set of decisions and intrahousehold dynamics and shouldn't be thought of separately at all.
This is a useful paper for sure. But it has been known for a long time that the recorded rate of growth of remittances is distorted by under-recording. Indeed recorded Sub Saharan African remittances may still be half of the actual flows, even less. For example the Bank of Ghana has been recording remittances at only 15% of the actual level shown in micro-surveys. The IMF has recently introduced new Balance of Payments classifications to rectify this. So the Sub Sahara African remittances ‘surge’ may continue for a while.
But how far does this data affect our understanding of economic development prospects? As I emphasize in my book “Development Without Aid: The decline of development aid and the rise of the diaspora” (Anthem Press) the answer is “not much.” Even if Sub Sahara Africa’s remittance flow turned out to be double that which is recorded the probable amount of investment out of those remittances would contribute less than one half of one percent to economic growth. On the other hand the potential annual investment impact of the return of flight capital (which two IMF researchers have estimated at the best part of one trillion dollars) combined with the redeployment to Africa of the savings accumulated by migrants in host country bank accounts could far exceed the investment out of remittances if Africa can make its business environment more attractive.
And this is not all. Even more important probably are the social capital assets – skills, business knowhow, and entrepreneurship – which returnees may bring back with them. Classic examples of this dynamic are probably Koreans from Japan in the 1940s and Indians from the US in the 2000s. In Africa now the Somali diaspora is engaged in a recovery effort in their homeland which started years ahead of the arrival of the development aid agencies, consisting of remittances and entrepreneurs.
So the main issue is not remittances. To some extent they are a sideshow which has consumed a disproportionate amount of aid donor time and attention. It is what those remittances mean in terms of the much wider dynamic of diaspora engagement that is critical to future development and aid policy.
Hi David,
Its a great article and gives insights on a macroeconomic factor which is widely accepted to be a major driver of economic growth in many developing countries. In Bangladesh, remittance growth is generally cited as a pivotal factor in bringing our above 6% annual GDP growth rate.
But I agree with Milica (previous comment) that it is known in Bangladesh that measurement methodologies (along with money sent back through informal channels) distort actual growth rates. Citing a former central bank governor in Bangladesh, our remittance figures would probably be twice of what they are if remittance through informal channels could be measured. We believe what's more important is how best to make the public data more accurate and reflective of the true picture. With that regard, Bangladesh Bank has taken great strides in updating remittance info
(as reported by all the commercial banks) on a bi-weekly basis. But , there is still a lot to be done in areas such as accurately recording number of yearly manpower exports, consistency in methodologies used to calculate remittance inflow by the banks etc. These changes are happening gradually in Bangladesh and it will be very interesting to see if the "share of recorded remittance growth accounted for" of 27.6 for Bangladesh improves in the next few years.
Hi David
It is an interesting article which argues that why remittance data might be overestimated. However, there are several reasons as to why remittances figures might be underestimated rather than overestimated. First, as mentioned in the paper, still a substantial proportion of remittances are transferred through informal channels such as hundi, hawala, chit funds due to various reasons such as hosting countries’ macroeconomic environment and government policies and regulations such as currency controls and overvalued currencies, high tariffs and taxes and slow and expensive licensing process for the financial institutions. For example, in the context of Nepal, the living standard measurement has shown that only about 1 percent of remittances inflow from India comes through formal channel! Second, due to the high cost, slow transfer service and lack of access to financial services render the informal money transfer more attractive to migrants (Buencamino and Gorbunov, 2002).
Second the data on the stock of migrants and hence the growth of migrants is scant and not reliable. For example, in a recent report on improving the database on migrants, it states: “Our patchy statistics on international migration amount to an enormous blind spot (Tomas et al. 2009, Preface, p. v). The report further states that “the data on international migration that countries now collect and publish are so limited, however, that we know much less about how much and what kind of migration is happening in today’s world than we know about international trade and investment flows (p.1)”. Since the data on migration is politically sensitive and not many countries report the true values, we can speculate the true figures would be much higher than the scant official data. In that context, the growth of migration (gL*) in equation (5), (p. 12) would be much smaller and the consequent estimate of the remittances would be much lower.
Third, in contrast to settlement migration, contract migration and seasonal migration have been a feature of many developing countries. The seasonal or short term workers, who stay less than one year in the destination countries, by definition, would not be counted as ‘residents’ of the destination countries according to the IMF’s Balance of Payments manual (IMF 1993; Rienke 2007). The contract migration has been a feature specially for the workers working in the Gulf countries from primarily South Asian countries. Thus, the true remittances would fall short of the official figures.
Fourth, as argued in the paper there are discrepancies in the micro data estimated from the household surveys and the macro data in the balance of payments statistics. However, household surveys are only representative surveys and people tend to under report their true income (Shonkwiler et al. 2014). Moreover, since the balance of payments statistics is based on double entry book keeping system, a large and persistent errors and omissions might provide useful check into the accuracy of the remittances data (though errors and omissions can come from different sources and sometimes mutually offsetting).