Migrants as transnational actors

Migration is one of the defining characteristics of our current period of globalization.Footnote ⁸ The United Nations estimates that as of 2010 over 213 million people, or 3 per cent of the world's population, live outside of their country of origin.Footnote ⁹ The lion's share of this migration has flowed from poorer to richer countries. While migrants represent less than 2 per cent of the population in less developed regions, for more developed regions, migrants represent over 10 per cent of the population. For most developing countries, a significant proportion of the country's citizens reside in another country, often a developed country.

Inexpensive and readily available international transport and communications networks make it possible for migrants to remain connected with their home countries.Footnote ¹⁰ International telecommunications traffic was estimated at 413 billion minutes in 2010.Footnote ¹¹

In a cross-national analysis, Perkins and Neumayer find that migrants generate a large share of this traffic.Footnote ¹² Single-country studies support these findings. A survey of Mexicans living in the United States found that over half of all respondents talk to family in Mexico at least once a week,Footnote ¹³ and ethnographic research documents the integral role international calls play in the lives of Filipino,Footnote ¹⁴ Jamaican,Footnote ¹⁵ SalvadoranFootnote ¹⁶ and SomaliFootnote ¹⁷ migrants. Additionally, many migrants frequently return to their countries of origin for visits. In a series of surveys of Latin American migrants living in the United States, Guarnizo, Portes and Haller find that one-fifth of respondents regularly visit their country of origin,Footnote ¹⁸ while Waldinger finds that two-thirds of respondents return home at least once.Footnote ¹⁹ Ethnographic studies highlight the importance of return trips home for Caribbean,Footnote ²⁰ Filipino,Footnote ²¹ PolishFootnote ²² and SalvadoranFootnote ²³ migrants.

A growing body of migration research has found that many migrants maintain close ties with their home communities, documenting migrants’ familial, social, economic, religious, political and cultural embeddedness in both their host and home countries.Footnote ²⁴ As Levitt and Jaworski note, ‘while the numbers who engage in regular transnational practices may be fairly small, those who engage in occasional, informal transnational activities, including social, cultural, and religious practices, in response to elections, economic downturns, lifecycle events, and climactic disasters are much greater’.Footnote ²⁵

Migrants maintain ties with their homelands for a variety of reasons, ranging from instrumental motives, such as supporting family members who stay behind, to non-instrumental motives, such as maintaining a sense of belonging to their home communities. Moreover, these ties appear to persist over time. In contrast to early models of migrant behaviour, in which migrants were expected to assimilate into their host country's society and abandon ties with their country of origin, recent analyses find that migrants neither fully assimilate into their host country's society nor sever their ties with their home country.Footnote ²⁶ In the light of these findings many migration scholars view migrants as transnational actors who maintain stable, active relationships with both their home and host countries.Footnote ²⁷

Migrants’ ties with their countries of origin can be seen tangibly through the money they send home. The World Bank estimates that developing countries received US$325 billion in remittances in 2010.Footnote ²⁸ These remittances represent a significant contribution to migrants’ home countries’ economies. As of 2005, remittances were almost 90 per cent as large as foreign direct investment flows and outstripped official development assistance, making remittances the second largest source of external funding for developing countries. For the top twenty remittance-receiving countries, remittances exceed 10 per cent of gross domestic product (GDP).Footnote ²⁹ Remittances help reduce poverty and income volatility, increase health and education expenditure and provide investment capital for small-scale entrepreneurs.Footnote ³⁰ For many developing countries, remittances have become a vital support for the economy and a source of economic security for remittance recipients.

Migrants’ decisions to send money home reflect both instrumental and non-instrumental motives.Footnote ³¹ Economically, remittances can finance consumption back home,Footnote ³² serve as insurance against adverse economic shocksFootnote ³³ and underwrite investments.Footnote ³⁴ Socially, remittances provide a mechanism for increasing social status,Footnote ³⁵ maintaining community ties,Footnote ³⁶ strengthening familial bondsFootnote ³⁷ and solidifying clientelistic power bases.Footnote ³⁸

Cross-national analyses of remittance flows have found support for migrants’ multifaceted economic and social motives for sending money home.Footnote ³⁹ However, such analyses have ignored migrants’ political motivations. Implicitly, this rests on the assumption that migrants are either not politically involved in their home countries, or to the extent that they are politically involved, their political involvement is not financial. These assumptions contrast with many country-level and micro-level studies that emphasize that migrants are often important political actors in their home countries and have identified specific cases of politically-motivated remittances.

Migrants as political actors

Migrants’ political ties range from following political news through home country media, to participating in domestic elections (as voters, candidates and financial underwriters), and shaping political debate in their countries of origin. With the large number of migrants and migrants’ enduring ties to their home countries, ‘immigrant communities turn into an unexpected, but increasingly visible actor in the politics of their home towns and countries’.Footnote ⁴⁰

Many migrants remain informed about current events back home and increase their attention to home country news around elections.Footnote ⁴¹ Migrants may participate in home country politics more actively, as well. Migrants have lobbied for the opportunity to vote and to run as candidates in home country elections.Footnote ⁴² For countries that do not allow for absentee voting, migrants may even return home to vote, with trips sometimes subsidized by political parties.Footnote ⁴³

As international communications networks make transnational communications easier, migrant diasporas are increasingly playing an active role in shaping the political agenda in their home countries.Footnote ⁴⁴ As transnational social networks, diasporas provide a forum for political debate, organization and co-ordination free from constraints the state might impose on political freedoms. As such, diasporas have been particularly important in less democratic, newly independent and post-civil conflict countries. The political support and endorsements lent by diasporas have helped exiled political leaders return home (for example, Ahmad Chalabi in Iraq, Ellen Johnson-Sirleaf in Liberia and Ashraf Ghani in Afghanistan), and have garnered international recognition for new regimes.Footnote ⁴⁵ Political debate within diasporas has shaped the formulation of national constitutions and the terms of ceasefire agreements in civil conflicts.Footnote ⁴⁶ A few examples help to illustrate the importance of diasporas as political actors. Support from the overseas Indian community has been crucial in the resurgence of the Hindu right in India,Footnote ⁴⁷ while ‘diaspora Eritreans have mobilized demonstrators, amassed funds for war, debated the formulation of the constitution and influenced the government of Eritrea’,Footnote ⁴⁸ and the need for diaspora support has led most Liberian presidential hopefuls to announce their candidacies in the United States.Footnote ⁴⁹

Of particular interest for this analysis, many migrants also contribute financially to their favoured political causes. As countries’ expatriate communities expand, political parties are increasingly turning to migrants for campaign finance. Many parties have opened overseas chapters in migrant communities, and political candidates campaign and raise funds abroad.Footnote ⁵⁰ Examples abound. Elections in the Dominican Republic are contested both at home and in the United States, and campaign contributions from the United States account for 10 to 15 per cent of major Dominican parties’ revenues.Footnote ⁵¹ In Ethiopia, both major political coalitions are dependent on diaspora organizations for their funding.Footnote ⁵² Turkish opposition parties use overseas campaign trips to fill the party coffers.Footnote ⁵³ The Indian Bharatiya Janata Party relies on funds raised in the United States, Great Britain and Australia.Footnote ⁵⁴ With rising migration, political campaigns are becoming increasingly global, as parties fight for migrants’ funding and support.

As with all dimensions of migrants’ transnational ties, migrants’ decisions to engage in politically-motivated remittances reflect both instrumental and non-instrumental motives. For migrants wishing to influence political outcomes in their countries of origin, direct financial contributions to politicians or their parties may increase migrants’ political leverage, either by increasing the likelihood that their preferred candidates attain office or by increasing their access to the political process. Political contributions may be given directly to parties by migrants or passed on to politicians by migrants’ families.

Remittances can indirectly serve to meet migrants’ instrumental political goals as well. Sending money home to family members can provide migrants with political leverage. As Ricardo Monreal, a former governor of the state of Zacatecas, Mexico, notes, migrants’ ‘economic influence is huge and their political clout as a consequence of that is huge too’.Footnote ⁵⁵ In assessing the impact of Mexican migrants’ political preferences and remittances on voters back home, Meyer notes that remitters ‘can influence the votes of those who stayed … as they send money, they also send messages’.Footnote ⁵⁶ Reflecting the patronage networks embedded in many remittance-receiving households,Footnote ⁵⁷ scholars have found that many migrants view their family's pursuit of their political goals as an explicit quid-pro-quo for their economic support.Footnote ⁵⁸ Sacchetti reports this viewpoint explicitly in an interview with a Dominican migrant living in Boston. ‘He isn't shy about telling his relatives whom to vote for to get results. ‘What I tell the family to do, they do it … We are the ones who support them.’Footnote ⁵⁹ This view is mirrored by remittance recipients who often feel obliged to represent their remitter's political preferences.Footnote ⁶⁰

Not all of migrants’ politically-motivated remittances reflect instrumental motives. Contributing to political campaigns can also serve as a non-instrumental form of political engagement. In an analysis of campaign finance in the United States, Ansolabehere, de Figuieredo and Snyder argue that most campaign contributors donate for non-instrumental reasons. ‘Individuals give because they are ideologically motivated, because they are excited by the politics of particular elections, because they are asked by their friends or colleagues and because they have the resources necessary to engage in this particular form of participation.’Footnote ⁶¹ Added to these motivations for migrants, contributing to political campaigns also strengthens their sense of attachment to their homeland. Contributing to political campaigns allows migrants to be involved in politics back home. In particular, for migrants from countries currently undergoing democratization, participating in the political process is exciting and allows them to be a part of their country's political transformation. This motivation is especially strong for migrants supporting opposition candidates. As one Mexican migrant in the United States responded when asked why he campaigned for Vicente Fox, Mexico's first opposition candidate to win the presidency, ‘He is the hope for our community here and in our homeland’.Footnote ⁶² Migrants’ decisions to engage in politically-motivated remittances, as with all forms of migrants’ transnational behaviour, reflects migrants’ multidimensional motivations.

The research discussed above has identified individual examples in which migrants have used their economic resources to participate politically back home. However, while this literature provides strong evidence that migrants have engaged in politically-motivated remittances, it cannot assess the prevalence of political remittances more broadly or the degree to which such remittances are systematically influenced by political and economic factors. This article analyses aggregate remittance flows for evidence that remittances reflect migrants’ political interests. It is the first study to provide cross-national evidence of political remittance cycles in developing countries, and moreover, to demonstrate that political remittance cycles vary systematically based on countries’ political and economic environments.

Elections and Politically-Motivated Remittances

Elections provide one of the most important opportunities for migrants to engage in politics back home. If migrants send more money home as a non-instrumental form of political participation, then remittances should increase when politics is most accessible and engaging. For most survey respondents who express interest in politics, their interest waxes and wanes in lockstep with the electoral calendar.Footnote ⁶³ In response to the public's greater engagement, political media coverage also increases around elections. As a result, I expect remittances as a form of political participation to increase in election years.

Instrumentally, with elections looming political elites are most open to influence, both directly through processes of selection and replacement, and indirectly as politicians become more responsive to public opinion. Politicians are also most in need of money at this time, providing potential donors such as remitters with greater leverage. As a result, direct financial contributions to politicians or their parties intended to increase migrants’ political leverage, either by increasing the likelihood that their preferred candidates attain office or by increasing their access to the political process, should increase prior to elections. Similarly, if migrants’ remit more to increase their influence over remittance recipients’ voting behaviour, then remittances should also increase prior to elections.

Thus, if migrants’ decision to remit is sensitive to the electoral environment in their country of origin, remittances should be greater in election years than in non-election years, all else equal. This leads to Hypothesis 1.

Moreover, politically-motivated remittances should depend on the political and economic context in which the election occurs. Migrants are likely to send more money home prior to elections in which their remittances have greater political leverage or generate greater enthusiasm. If the outcome of an election is considered a foregone conclusion, or if migrants’ remittances are unlikely to have a significant impact on politicians’, parties’ or voters’ behaviour, then the political leverage of migrants’ money should be no greater in an election year than in a non-election year.

Political leaders in developing countries enjoy a large incumbency advantage.Footnote ⁶⁴ In an analysis of presidential elections in developing countries from 1975 to 2004, Baturo finds that 80 per cent of presidents in democratic or partially-democratic regimes who ran for re-election won.Footnote ⁶⁵ This incumbency advantage extends to incumbents’ chosen successors, as well. Many executives in developing countries face term limits, which restrict incumbents’ ability to run in successive elections. When the incumbent cannot stand for re-election, a successor is often chosen to run. While successors are not as successful as incumbents in elections, they still have a considerable electoral advantage.Footnote ⁶⁶ As a result, elections in which incumbents or chosen successors do not run generally generate greater interest and attention and may provide greater opportunities for migrants to influence politics. This effect should apply regardless of whether migrants support the incumbent, the opposition or both. This leads to Hypothesis 2.

Finally, to the extent that migrants are motivated by their ability to influence politics back home, migrants’ political leverage should be greater when remittances represent a greater relative contribution either to politicians’ coffers or recipients’ income. Thus, the extent to which remittances influence political outcomes may depend upon how wealthy the country is. For poorer countries, because there is less money available domestically for political purposes, politically-motivated remittances may provide more ‘bang for the buck’. As a result, the less wealthy the country, the more influence politically-motivated remittances may provide for migrants. This leads to Hypothesis 3.Footnote ⁶⁷

Cross-National Empirical Analysis

The discussion presented in the previous section suggests that migrants’ decisions to remit may reflect political considerations, and in particular their home countries’ electoral calendars. Before turning to the statistical analyses, Figure 1 presents a brief plausibility assessment of the three hypotheses analysed in this section.

Fig. 1

Comparison of mean values of remittances by elections

Note: Number of observations in each category in parentheses.

Hypothesis 1 posits that remittance flows are larger in election years than in non-election years. As a preliminary diagnostic, Panel A presents the mean value of remittances for election and non-election years. On average, remittances are 30 per cent larger in election years than in non-election years, providing initial support for Hypothesis 1. Hypothesis 2 suggests that migrants remit fewer politically-motivated remittances when the incumbent or chosen successor runs. Panel B presents the mean value of remittances in election years depending on whether or not the incumbent or a chosen successor ran in the election. On average, remittances are 58 per cent larger in election years in which the incumbent or a chosen successor did not run than in election years in which the incumbent or chosen successor did run. This difference provides preliminary support for Hypothesis 2. Finally, Hypothesis 3 posits that elections have a greater impact on remittance flows in poorer countries. As an initial cut, Panels C and D present the mean value of remittances in election and non-election years for above median income and below median income observations. In support of this hypothesis, on average, remittances in poorer countries are 36 per cent larger in election years than in non-election years. For richer countries, remittances are only 19 per cent larger in election years than in non-election years. Taken together, these descriptive statistics provide initial support for the argument developed in this article.

These hypotheses are tested more rigorously below through a time-series cross-sectional analysis of annual remittance flows to eighty-one nominally democratic developing countries, from 1990 to 2005.Footnote ⁶⁸ The estimation technique employed is linear regression with panel-corrected standard errors.Footnote ⁶⁹ Given the unbalanced and heterogeneous nature of the sample, linear regression with panel-corrected standard errors represents a valuable combination that gains explanatory leverage from cross-national and cross-temporal data, while correcting for possible spatial and temporal correlation of errors.Footnote ⁷⁰

Ideally, the dependent variable for this analysis would explicitly identify politically-motivated remittances, discriminating between migrants’ political motivations and other motivations such as migrants’ economic or social motivations. Unfortunately, such data do not exist. The research strategy adopted here is to examine aggregate remittance flows, testing for evidence that they are partly driven by political motives, while controlling for other factors known to influence remittance flows. This research strategy is a commonly adopted approach when analysing aggregate economic data, as is the case, for example, in analyses of government spending,Footnote ⁷¹ political business cycles,Footnote ⁷² trade flowsFootnote ⁷³ and foreign direct investment.Footnote ⁷⁴

Remittance data are from the World Bank, and are a composite measure of officially-recorded flows of workers’ remittances, compensation of employees and migrant transfers.Footnote ⁷⁵ While each of these categories is analytically distinct, they are practically indistinguishable for most countries.Footnote ⁷⁶ Although the World Bank's remittance data are the most comparable and comprehensive cross-national remittance data available, a degree of caution is required when using these data due to the difficulties that exist in collecting these data.Footnote ⁷⁷ In particular, there are two types of data ‘missingness’ to consider with respect to the World Bank's remittances dataset. The first concerns country–year coverage. These data are available for 85 per cent of the country–year observations for which election data are available for nominally democratic developing countries, 1990–2005. Unsurprisingly, there is a systematic bias in which observations are missing data.Footnote ⁷⁸ Most of the missing observations occur in the first few years after a regime transition in newly independent and post-communist countries, and correspond more broadly with an absence of economic data for these countries, thus the cross-national missingness of remittance data is comparable to that of other cross-national analyses using economic data. To the extent that remittance data are more likely to be missing for those post-transition elections that one would expect to matter greatly to migrants, this type of missing data is likely to result in an underestimation of an election-year effect on migrants’ remittances.

Secondly, and more fundamentally, because remittances are often transferred unofficially, official measures underestimate remittance flows, with analysts estimating that informal flows range from 50 to 250 per cent of official flows.Footnote ⁷⁹ As a result, official remittance flows do not capture the universe of remittances and, thus, unofficial remittances remain outside the scope of this dataset. With respect to politically-motivated remittances, the inability to assess informal remittance flows means that if migrants are more likely to transmit politically-motivated remittances unofficially, then the results reported below will underestimate the level of political remittances. However, it is unclear whether politically-motivated remittances are systematically more likely to be remitted unofficially than other types of remittances.Footnote ⁸⁰

For cross-national comparisons, remittances are analysed either as a percentage of GDP,Footnote ⁸¹ or as absolute levels of remittances.Footnote ⁸² While similar conceptually, the two measures can be quite different in practice. For example, a $100 million increase in remittances to Mexico or Albania would look identical when analysing absolute remittance flows, but would represent a 0.2 per cent increase in terms of GDP in Mexico and a 2 per cent increase for Albania. There is no theoretical reason why either measure is preferred for the analysis of politically-motivated remittances. As a result, I use natural log transformations of both measures as dependent variables for this project.Footnote ⁸³ The two series correlate at 0.53. Because the results from the two series are almost identical, I present those for Remittances, ln in Table 1, and those for Remittances as a percentage of GDP, ln in the online appendix.

Table 1

Elections and Remittances

Note: Number of observations, 926; ^p < 0.1, *p < 0.05, **p < 0.01; standard errors in parentheses.

In order to be confident in the effect of elections on remittance flows, it is necessary to consider migrants’ other motivations and other factors that may influence remittance flows. To do so, I build on the empirical results of Niimi and Ozden and Freund and Spatafora, two articles which examine the determinants of remittance flows cross-nationally.Footnote ⁸⁴ Remittance flows tend to reflect economic conditions in the home country. Remittances are most often viewed as an income supplement for a migrant's family, and as a form of insurance against bad economic times.Footnote ⁸⁵ In keeping with these two motives, remittances are expected to be higher for poorer countries, and for countries undergoing economic downturns. Interestingly, while these articles reveal strong support for the negative correlation between home country income and remittance flows, there is no clear impact for economic growth. One explanation for the lack of a relationship between remittances and economic growth is that while migrants may remit more in bad times to cushion their families from the poor economy, they might also remit more in good times as an economic investment.Footnote ⁸⁶ To control for the effects of the home country's economy on remittance inflows, I include measures for the size of the home country's economy (GDP, ln), its economic income (GDP per capita, ln) and its growth rate (GDP growth).Footnote ⁸⁷

Global economic conditions matter as well. First, migrants’ economic fortunes affect how much money they can send home – higher migrant income is associated with higher remittances. This finding is well-established at the level of the individual migrant, but has not been the focus of cross-national work due to the difficulty in identifying the destination of each country's migrants.Footnote ⁸⁸ Parsons, Skeldon, Walmsley and Winters have developed an international bilateral migration stock database for 226 by 226 countries that helps address this problem.Footnote ⁸⁹ This dataset, which is based on country census data for the year 2000, reports for each home country the number of migrants residing in each of the other 225 countries. I transform these dyadic data into the percentage of the home country's migrants residing in each country. I then use these percentages to weight migrant-receiving countries’ GDP per capita and GDP growth to create measures of Migrant GDP per capita and Migrant GDP Growth. These measures constitute cross-national proxies for migrants’ economic conditions.Footnote ⁹⁰ In keeping with individual-level results, I expect that remittances will be positively correlated with both Migrant GDP per capita and Migrant GDP Growth.

Secondly, as an international capital flow, remittances may respond to global investment opportunities. In particular, remittances to developing countries may be negatively correlated with investment opportunities in developed countries. Such investment opportunities are more likely the better economic performance is in developed economies. To control for this, High Income Country GDP Growth is included in the analyses. Thirdly, as global migration increases, remittance flows may also reflect the higher number of foreign nationals living in the home country. As the number of migrants living in the home country increases, the amount of remittances going to these households may increase proportionately. To capture this effect, I include a measure for the home country's international migrant stock (Migrant Stock, ln).Footnote ⁹¹

Finally, as with most time-series variables, the best indicator of remittance flows in the current period is remittance flows in the previous period. Therefore, a one-period lag of the dependent variable is included as an explanatory variable.

Turning to Model 1 in Table 1, as expected, the lagged dependent variable is a strong predictor of remittances in the current period. Similarly, an increase in the country's international migrant stock (Migrant Stock, ln) is correlated with an increase in remittances. Surprisingly, home country economic conditions do not appear highly correlated with remittance flows (neither GDP per capita nor GDP Growth is statistically significant). With respect to migrants’ economic characteristics, both Migrant GDP per capita and Migrant GDP Growth are positively correlated with remittance flows, but only Migrant GDP per capita is statistically significant. This suggests that while migrants’ long-run wealth may affect their decision to remit, short-run fluctuations may have relatively little effect. The recent decline in the growth rate of remittances from the United States, which respondents in an Inter-American Development Bank survey attribute to the slowdown in the American economy, belies this implication.Footnote ⁹² An alternative explanation for the lack of statistical significance for Migrant GDP Growth may be that this measure is a relatively noisy proxy for migrants’ economic performance. To the extent that migrants’ economic performance does not fully reflect that of their country of residence, then Migrant GDP Growth will not fully capture migrants’ economic context.Footnote ⁹³

Finally, remittances appear to respond less to investment opportunities than do other international capital flows. Although remittances are negatively associated with High Income Country GDP Growth, the relationship is not statistically significant in Model 1 (and only marginally significant in Models 2–8). This result is in line with previous studies that find that migrants’ decisions to remit are more likely to reflect altruistic and familial concerns than market-based investment opportunities.Footnote ⁹⁴

Turning to the electoral impetus for remittances, election data are from Hyde and Marinov, who have developed a database that includes descriptive information on all elections for national office in developing countries.Footnote ⁹⁵ The argument presented in this article presupposes that holding an election implies some risk that the incumbent will give up power. In some elections, however, opposition is banned or otherwise restricted. To exclude elections that are a priori uncompetitive, Election is coded from three questions in the Hyde and Marinov data: Was opposition allowed? Was more than one party legal? Was there a choice of candidates on the ballot? Election is coded 1 if the answer to all three questions is ‘Yes’ and 0 otherwise, generating a list of elections in which competition is possible. By this measure, there are 281 elections in the sample, 30 per cent of the observations.

If Hypothesis 1, which posits that remittance flows increase in election years, is correct, then elections should be positively and significantly associated with remittances. Model 2 provides support for this hypothesis. Based on these results, remittances are estimated to increase 6.6 per cent in election years.Footnote ⁹⁶ For Slovakia, which on average receives US$51 million in remittances annually and represents the twenty-fifth percentile of remittance recipients, an election is predicted to increase remittances by US$3 million. For Korea, which receives US$892 million annually in remittances and represents the seventy-fifth percentile of remittance recipients, an election is predicted to increase remittance flows by US$55 million.

Although Election excludes elections that are structurally uncompetitive, some of the included elections may not have been highly competitive in practice. Given the strong incumbency bias in developing country elections, migrants may perceive elections in which the incumbent or the incumbent's chosen successor runs to be of less interest and less open to their influence than elections in which neither the incumbent nor their chosen successor runs. As a result, Hypothesis 2 posits that migrants remit fewer politically-motivated remittances in election years in which the incumbent or successor runs. To test this hypothesis, I bifurcate Election into Incumbent or Successor Runs (140 elections) and Neither Incumbent nor Successor Runs (141 elections).Footnote ⁹⁷ If migrants only engage in politically-motivated remittances when the incumbent or chosen successor does not run, then Neither Incumbent nor Successor Runs should be positive and significant, and Incumbent or Successor Runs should not be significantly different from zero.

Turning to Model 3, the results are supportive of this hypothesis.Footnote ⁹⁸ Remittances are estimated to increase twelve per cent in election years in which neither the incumbent nor a chosen successor runs. This suggests that for Armenia, which on average receives US$210 million in remittances annually and represents the median remittance receiving country, an election in which neither the incumbent nor a chosen successor runs is predicted to increase remittances by US$25 million. In contrast, there is no support for the existence of politically-motivated remittances when the incumbent or chosen successor runs. This suggests that higher levels of electoral uncertainty may generate more remittances, and that when it comes to engaging in politically-motivated remittances, migrants care not only about whether an election occurs, but also how contested the election is.

Migrants’ decision to engage in politically-motivated remittances should also reflect remittances’ economic leverage. For poorer countries, because there is less money available domestically, politically-motivated remittances may provide more ‘bang for the buck’. As a result, Hypothesis 3 posits that election-year increases in remittances are larger the less wealthy the home country. To test this hypothesis, Model 4 includes both Election, and an interaction between Election and GDP per capita, ln. If Hypothesis 3 is correct, I expect the largest increase in election-year remittances to occur at the lowest level of GDP per capita, ln, and for the increase in election-year remittances to taper off as GDP per capita, ln increases.

Because Model 4 tests an interactive effect, the size and significance of an election on remittance flows is difficult to ascertain from the regression results. To gauge support for Hypothesis 3 more effectively, Figure 2 graphically displays the marginal effect of an election as a country's income varies.Footnote ⁹⁹ In keeping with Hypothesis 3, Figure 2 displays a significant increase in election-year remittances when GDP per capita is below $1,677. In contrast, when GDP per capita is above $1,677, politically-motivated remittances do not appear statistically significant.Footnote ¹⁰⁰ To contextualize this result, 53 per cent of the observations in the sample have GDP per capita below $1,677. Two conclusions can be drawn from these results. First, in keeping with Hypothesis 3, the poorer the home country, the more financially active migrants may be in politics. Secondly, although politically-motivated remittances may be more prevalent in poorer developing countries, such remittances appear to be a common feature of developing country elections.

Fig. 2

Marginal effect of an election on remittances as GDP per capita varies

Although the preceding analyses provide support for the argument that migrants send more money home in election years, they cannot demonstrate conclusively that this increase in remittances reflects migrants’ political participation. Elections may be associated with migrants’ non-political motivations for sending money to their countries of origin. In particular, migrants may send more money home in election years as an insurance policy against domestic uncertainty that might arise in response to elections. Alternatively, migrants may send more money home in election years due to a better investment climate around elections. These alternative explanations are explored below.

While elections may provide one of the most important opportunities for migrants to engage in politics back home, they may also generate greater domestic uncertainty – socially, politically and economically. As a result, migrants may increase their remittances in election years to provide a financial buffer for their recipients to offset potential adverse consequences arising from the elections. There are strong reasons to believe that elections and political violence are often related, and the connection between elections and political violence is well documented in literature examining both contemporary and historical political violence.Footnote ¹⁰¹ Elections provide focal points around which political conflict can erupt, as electoral competition may not only promote inter-group competition during a particular period, but may also enhance grievances. Thus, elections may represent a period of greater risk for migrants’ families back home. If elections serve as a proxy for domestic turmoil, rather than as an opportunity for migrants to participate in and influence politics in their home country, then remittances should increase more in election years in which there is social unrest.

I test this alternative explanation in Models 5 and 6 in Table 2 using measures of social unrest based on two questions included as variables in the Hyde and Marinov elections dataset: (1) Was there significant violence involving civilian deaths immediately before, during or after the election? (2) Were there riots and protests after the election? Based on these data, there was significant violence against civilians in 28 per cent of the elections under analysis, and riots or protests in 18 per cent of the elections.Footnote ¹⁰² If migrants remit more money in election years to insure against election-related domestic turmoil, then remittances should increase more in election years associated with violence against civilians (Model 5) or with riots and protests (Model 6). Based on the results reported in Models 5 and 6, there is no support for this alternative explanation. There does not appear to be an increase in remittances in election years in which there is violence against civilians or post-electoral riots and protests. In contrast, there is strong evidence of election-year increases in remittances when there is no violence against civilians or post-electoral riots and protests.

Table 2

Elections in Good and Bad Times

Notes: Number of observations, Models 5, 6 and 7, 926; Model 8, 913; ^p < 0.1, *p < 0.05, **p < 0.01. Standard errors in parentheses.

Rather than insuring against bad times, election-year remittances may be spurred by economic good times. Governments in developing countries are more likely to be re-elected when the economy is strong.Footnote ¹⁰³ As a result, governments often attempt to call elections when the economy is doing well,Footnote ¹⁰⁴ or, if the electoral calendar is fixed, to increase the strength of the economy in the run-up to the election.Footnote ¹⁰⁵ Thus, elections may be associated with better investment opportunities for migrants. If so, then election-year remittances may increase in response to the improved investment climate, rather than as an act of political participation.

I test this alternative explanation using two measures to capture the home country's investment climate. Investment opportunities tend to increase as the strength of the economy increases. As a result, Model 7 includes an interaction between GDP Growth and Election in Model 7. If migrants remit more money in election years to take advantage of investment opportunities, then remittances should increase in election years with higher GDP growth. The results reported in Model 7 provide no support for this alternative explanation. There appears to be no relationship between election-year remittances and the strength of the home country's economy.

The lack of support for this alternative explanation may reflect the fact that the strength of the economy can have competing effects on migrants’ decisions to remit. While migrants may remit more in economic good times for investment purposes, they might also remit more in economic bad times to cushion their remittance recipients from adverse economic conditions. However, migrants’ remittances are not the only international financial flow that is affected by changes in countries’ investment climates; foreign direct investment should also be responsive to investment opportunities in developing countries. As a result, I expect foreign direct investment to increase as more investment opportunities are available in migrants’ home countries. To focus more directly on the investment climate, in Model 8 I include an interaction between Change in Foreign Direct Investment as a percentage of GDP and Election as a proxy for election-year investment climate. If migrants remit more money in election years to take advantage of investment opportunities, then remittances should increase in election years with a greater increase in foreign direct investment. In keeping with Model 7, the results reported in Model 8 do not provide support for the alternative explanation that the increase in election-year remittances is driven by greater investment opportunities in the home country.

Taken together the results reported in Tables 1 and 2 provide strong support for the existence of political remittance cycles. Migrants send more money home in election years. This increase in remittances reflects migrants’ political motivations, rather than their insurance or investment motivations. Moreover, these results suggest that political remittance cycles reflect not simply the electoral calendar, but also the degree to which the electoral process provides an opportunity for migrants’ to engage in home-country politics.