Printer Friendly

Remittances and Economic Development: Evidence from the Caribbean.

Whether and how remittances impact long-run economic development in the Caribbean region are important and controversial questions. Remittances were the second largest inflow of foreign capital to the Caribbean over the last two decades, behind FDI (foreign direct investment) and ahead of foreign aid. For the Dominican Republic, Haiti, and Jamaica, remittances were the largest inflow, constituting 8 percent, 20 percent, and 14 percent of GDP on average over the period 1998-2010. Remittances have also become less volatile, and in the case of the recent global financial crisis, remittances were less volatile than FDI and aid. The magnitude and relative stability of remittances are important given that access to external financing is a significant constraint to the economic development of Caribbean countries (ECLAC 2013; Li 2013; Hurley 2015). At the same time, however, remittances are associated with migration. Migration rates, particularly among skilled workers, are high. Approximately 12 percent of the Caribbean labour force migrated to OECD countries during the period 1965-2000 (Mishra 2006). In 2010, approximately 16 percent of the Caribbean population resided in developed countries, increasing to 17 percent in 2015. (1) Migration is significant given that Caribbean countries, like other small island developing states (SIDS), face greater human capital constraints to growth owing to their small size. If remittances are not sufficiently contributing to economic development, the long-term economic and welfare costs of migration are likely to be onerous for these small economies.

There have been recent efforts by policymakers in the Caribbean to channel remittance inflows towards economic development. Dominican Republic, Guyana, Haiti, and Jamaica have been more active in the policy dimension. Efforts have included microfinance projects for the channelling of remittances from the diaspora to specific domestic investment initiatives, greater Diaspora engagement through fostering closer ties and engaging multilateral organizations for technical assistance (International Organization for Migration 2014; World Bank 2016a). While important, these attempts have only occurred in a handful of countries and they remain limited in scope. The policy discussion tends to revolve around how to engage the Diaspora, but too few policies explicitly focus on increasing remittance inflows and, more explicitly, linking them to broader development plans (Kirton 2011). Given the potential for external development financing presented by remittances and the risks associated with migration, it is imperative to know whether long-term remittance policies are required and what types of policies are required.

Empirically assessing the relationship between remittances and economic development is important in guiding the policy discussion for Caribbean countries. However, the empirical literature on this relationship in the Caribbean is sparse and consequently provides limited guidance to policymakers. At a theoretical level, remittances and economic development may be related in different ways, each with its own distinct policy implications. Several debates exist. The first is whether there is a causal link from remittances to development in the long run. If there is no causal link, or a causal link only exists in the short-run, policies promoting greater remittances for instance are not likely to promote long-run economic growth, and policies encouraging migration may harm economic development. The latter would therefore need to be more carefully considered. If there is such a long-run causal link, policies promoting remittances inflows are likely to be more effective if aligned to policies enhancing the domestic channels through which remittances promote growth. The second debate centres on the nature of the relationship between remittances and development in the long term. One hypothesis is that remittances positively affect long-term development. In this case, remittances may be beneficial to the home country and policies can aim to simultaneously encourage greater inflows and create a development strategy promoting the benefits of remittances. Another hypothesis is that remittances negatively affect development. In this scenario, policy might aim to reduce this negative impact. Finally, it may be that economic development itself impacts remittances. In this case for instance, policies that foster greater engagement of the Diaspora may be more fruitful in attracting remittances and channelling them to more productive uses.

In this article, we examine whether and how remittances relate to long-run economic development in Caribbean countries. More specifically, we address two questions: are remittances and economic development linked in the long run; and if they share a long-run relationship, how are they linked? In doing so, we provide support as to whether policymakers should pursue policies on remittances and if so, what types of policies. This is important given the access to external financing that remittances provide and the challenges that exist due to migration.

This article is most closely related to Lim and Simmons (2015) who study the relationship in general, using panel data for Caribbean countries. In contrast to the implicit assumption in that study that Caribbean countries are a homogenous group, we adopt a country-specific approach that allows the relationship to differ among countries. Additionally, we allow for different directions of causality between remittances and economic development. Our work extends on single-country studies, such as Kumar (2013) for Guyana, by studying the relationship in a wider group of countries. As such, we are able to compare and contrast findings across countries, and assess to what extent one-size-fits-all policies may be relevant to the region. Given our focus on policy support, these extensions are relevant to crafting more precise policies for each country.

This article examines the relationship between remittances and economic development in eleven Caribbean countries. We employ the bounds test approach to cointegration testing and the related ARDL framework outlined by Pesaran and Shin (1999) and Pesaran, Shin, and Smith (2001) to shed light on long-run causality between the variables. Overall, our results suggest that remittances relate to long-run economic development in several countries and the nature of these linkages is more heterogeneous than previously concluded. These have implications for nuanced policy approaches aimed at promoting long-run remittance-development relationships.

The remainder of this article is as follows: we discuss the background of capital flows to the Caribbean, provide a comprehensive review of the related literature, describe the data used and empirical techniques adopted, and discuss the results obtained. In the last section, we conclude and provide policy recommendations.


How do remittances compare to other inflows such as FDI and aid? Do remittances differ across countries, and if so, how? How have remittances evolved? This section discusses the properties of remittances and other capital flows to the Caribbean region.

Figure 1 shows the average size of remittances, FDI and ODA as a fraction of GDP for 79 developing countries, as well as the average remittance-to-GDP ratio for 14 Caribbean countries over the period 1998-2010.

The graph illustrates that, for developing countries, remittances have a greater share than FDI and, since the mid-2000s, a greater share than ODA, which has been on the decline. This reflects the growing importance of remittances in the development financing literature (UNCTAD 2013). The graph also shows remittances to the Caribbean in context by comparing its relative size in the region to other developing countries. Remittances represent a larger share of GDP to countries over time in the Caribbean than elsewhere. Additionally, remittances to the region have increased after 2009 in comparison to the decline for other developing countries, which shows a quick recovery following the 2008 global financial crisis. Given that remittances represent a significant inflow to developing countries, the evidence that the Caribbean average is higher, provides further justification for trying to understanding their impacts to these countries. Feeny, Iamsiraroj, and Mcgillivray (2014) mention that the size of remittances relative to the economies of small islands are larger than their size to developing countries and that this is one of the reasons that warrants a separate investigation of impacts in SIDS.

Figure 2 shows the average size of remittances, FDI, and ODA (official development assistance) as a fraction of GDP for fourteen Caribbean countries, over the period 1998-2010.

It illustrates the relative importance of remittances among other international flows to the region. Remittances to these countries have been greater than ODA as a share of GDP. They have also rivalled FDI inflows in some years, even surpassing them as a share of GDP in 2004. In 2010, remittances approached FDI in size, given the recovery of remittances and the continual decline of FDI after 2009. This information goes beyond the magnitude of remittances, in dollar terms, mentioned earlier. Remittances not only represent a large inflow of foreign exchange to the region, but they also carry a lot of weight. FDI is likely the major source of development financing for small island states (Hurley 2015), so these figures highlight just how important remittances can be in context. This gives some reasoning to the need to understand the potential significance of remittances to the region.

The potential importance of remittances to the region does not arise from its size alone but also by its sustainability. Figure 3 shows the trend in the volatility of remittances, FDI and ODA as a fraction of GDP over the period 2002-2010, using a five-year rolling window of the standard deviation.

It illustrates how stable these inflows are over time. Since 2004, the volatility of remittances has been generally on the decline, whereas that of FDI has been on the rise. ODA, often considered a stable source of finance for development (Chami et al. 2008), has remained relatively steady. In 2010, the volatility of remittances is lower than that of even ODA. Despitethe consistency of aid inflows, we saw in Figure 2 that it is now less than both remittances and FDI in the region. The growing magnitude of remittances coupled with increasing stability lends evidence to the view that remittances as development financing can have potentially sustained impacts.

Finally, we look at some of the differences among Caribbean countries. Figure 4 shows the size of remittances, FDI, and ODA as a fraction of GDP for fourteen countries in the region, averaged over the period 1998-2010.

It also shows the standard deviation of remittances over the period, as a measure of volatility. From Figure 3 we can see that seven countries have averaged a remittance-to-GDP ratio of five percent or more, with three of these over ten percent, while seven countries have a ratio of less than five percent. Additionally, we see considerable variation in the structure of international inflows to these countries. For many, FDI surpasses remittances. However, for some, such as Antigua & Barbuda and Trinidad & Tobago, FDI is clearly the dominant inflow, whereas in others such as Barbados and Dominica, remittances are close in size to FDI. Aid plays a large role in three countries: in Suriname, where it plays a much larger role than remittances, (2) and in Guyana and Haiti, where, despite the significance of aid, remittances are almost as much or even more. In countries like the Dominican Republic and Jamaica, remittances are the largest of the three inflows. The standard deviation of remittances ranges from 0.16 in Trinidad & Tobago to 7.12 in Guyana. We also see that the three highest remittance recipients have greater volatility in their remittance flows. These differences can mean, for example, that there may be (and should be) a relevant policy focus by the countries pertaining to each of the international flows. Thus, there is the need for country-specific focus.

This discussion highlights some points about remittances to the region. They are higher when compared to remittances to other developing countries; higher than aid; and, at times, rival FDI flows within the region. They are increasing in stability and there is variation among the Caribbean countries with respect to these inflows. However, while the properties of the remittance data highlight the possibility of large and sustained impacts for the Caribbean, it does not guarantee that such impacts exist. The properties of the remittance data show that a one-size-fits-all policy is not feasible for the region without further empirical investigation.


Theoretical Background

The current discussion on the long-run relationship between remittances and economic development is controversial. Several hypotheses explain the existence and nature of this relationship. First, there is the causal link from remittances to economic development. One view suggests that this interaction can be positive. For example, remittances can promote economic growth by financing capital accumulation--both physical and human. This is likely if remittances act to ease existing credit constraints, leading to greater expenditure on education and health or the accumulation of physical assets. Remittances may also promote growth by increasing the efficiency of investment (Chami et al. 2008). In theory, this occurs when remittances act as collateral for households, thus lowering the cost of capital and increasing the provision of credit, which leads to financial development (Chami et al. 2008; Barajas et al. 2009).

Another view, however, suggests that remittances may affect economic development negatively. Appreciation of the real exchange rate caused by greater inflows of remittances can reduce the size of the tradable goods sector. If this sector contributes to an economy's technological capacity, then total factor productivity falls (Chami et al. 2008). It is also possible for remittances to reduce labour force participation as the income it provides substitutes for labour income (Itzigsohn 1995; Amuedo-Dorantes and Pozo 2006). A third view is that there is no causality from remittances to economic development. For instance, remittances are unlikely to promote economic growth when used for consumption or unproductive investment. For example, remittances may largely finance the consumption of demerit or luxury goods, or housing expenditures. Thus, removing restrictive credit constraints would not lead to greater capital accumulation. Alternatively, remittances may face constraints to impacting development. For example, strong institutions are likely to channel this income into growth-enhancing purposes (Barajas et al. 2009). In the absence of these, we may not notice such effects.

There is also the causal link from economic development to remittances. Three main viewpoints address the nature of this reverse causality. The first relates to the altruistic motive, which suggests that migrants send remittances back home out of concern for the welfare of recipients. The utility of migrants therefore depends on the utility of the recipients at home. This view implies that as economic development increases, remittances will decline as the recipients become better off. If growth or development declines, however, remittances will increase to counteract this reduction in well-being in the home country.

The second view is that remittances follow an investment motive. Remittances are thus sent back home for the accumulation of assets. This may occur if the migrant expects inheritance in the future. Migrants may also expect to return home and favourable investment opportunities represent an avenue to increase wealth. The implication is that remittances increase in response to favourable conditions in the home country, rather than decline, according to the altruistic motive.

Alternatively, remittances may represent an informal arrangement for migration (Poirine 1997). The idea is that remittances may begin as repayment of an implicit loan taken out to migrate, which then becomes an implicit loan for future generations. This suggests that remittances are unlikely to respond to economic development in the home country and are generally stable over time. In this case, there is no causality from economic development to remittances.

Caribbean Literature

The empirical literature on the relationship between remittances and economic development in the Caribbean is mixed. A handful of studies investigate the causal link from development to remittances by examining the motivations to remit. Agarwal and Horowitz (2002) investigate these reasons in the case of Guyana, using household data. The authors found evidence to support the altruism motive, given that remittances decline as the number of migrants increase.

Campbell (2003) investigated the determinants of total foreign transfers, including remittances, into Barbados. Using the Johansen procedure, and the Engle-Granger cointegrating regression, the study found that total transfers increased as real income in the home country increased. The author suggests that this can encourage policy to attract remittances. Alleyne (2006) explicitly investigated the determinants of remittances for nine CARICOM countries using a GMM estimation approach, and found evidence for the investment motive. Both domestic GDP and the interest rate differential had a positive impact on remittances. Alleyne et al. (2008) examined the determinants of remittances to Jamaica, but with a focus on short-run determinants only. The authors used a time-varying parameter model and found evidence of the investment motive.

Moore and Greenidge (2008) and Alleyne, Kirton, and Figueroa (2008) both investigated the determinants of remittance inflows in the wider Caribbean, using data on fifteen and eight countries, respectively. Moore and Greenidge (2008) used a panel feasible GLS approach and found that remittances increased in response to a widening interest rate gap, but also increased when domestic income was lower relative to the host country. These results implied that investment and altruistic motives existed. Alleyne, Kirton, and Figueroa (2008) came to a similar conclusion, using panel cointegration and FMOLS estimation.

Several studies also examined the impact of remittances on economic development in the Caribbean region. A few household-level studies shed some light on the uses of remittances. Clarke and Wallsten (2003) and Beuermann, Ruprah and Sierra (2014) found that remittances acted as an insurance to natural disaster and health shocks, respectively, while Amuedo-Dorantes, Georges and Pozo (2010) and Bredl (2011) concluded that remittances eased constraints for poor households in Haiti which allowed for investment in education. These findings suggest that remittances can benefit economic development at the household level. On the other hand, Kim (2007) and Bussolo and Medvedev (2008) found that remittances had a negative impact on labour supply in Jamaica, as did Jadotte and Ramos (2016) in Haiti. This implies that remittances can hamper the functioning of the labour market. Stephenson and Wilsker (2016) showed that remittances had an effect on consumption decisions in Jamaica. In particular, they increased the likelihood of spending on education, housing, and luxuries. While this suggests conspicuous consumption, spending on education represents investment in human capital.

The mixed evidence on the causal relationship from remittances to development is also evident at the macroeconomic level. A few of these studies include Caribbean countries as part of a broader sample. Mundaca (2009) included seven regional states in a panel of twenty-five from the Latin American and the Caribbean (LAC) countries. Using GMM estimation, the study concluded that remittances did promote growth. Ramirez (2013) and Nsiah and Fayissa (2013) conducted similar analyses, including eight Caribbean countries and using panel cointegration techniques. Both studies concluded that remittances promoted growth in the LAC region. Garcia-Fuentes and Kennedy (2009) included four Caribbean countries in their panel of LAC countries and found that remittances only promote growth when there is a minimum level of human capital. The authors used a random effects model for estimation. Jackman, Craigwell and Moore (2009) and Amuedo-Dorantes, Pozo and Vargas-Silva (2010) conducted panel data analyses for SIDS, including Caribbean countries. These studies found that remittances reduced volatility and the real exchange rate, respectively, in small island states. On the other hand, Amuedo-Dorantes and Pozo (2004) found evidence that remittances reduced competitiveness in the export sector and led to Dutch disease effects in the LAC region. The authors included four Caribbean countries in their panel and used fixed-effects and IV regressions. Feeny, Iamsiraroj and McGillivray (2014) conducted a panel study for SIDS using GMM estimation. The authors found no evidence of causality from remittances to economic growth in the region, although they did find evidence of a positive relationship in other groups of SIDS.

Only a handful of studies explicitly investigated the macroeconomic impact of remittances on development in the Caribbean region. Campbell (2009) looked at this relationship in Barbados using the Johansen approach and an impulse response function in a model that includes FDI. The results suggest that remittances led to economic growth. Kumar (2013) provides similar evidence for Guyana, using the ARDL approach in an augmented Solow framework. The author found that this result held in both the long and short run. Evidence from causality tests suggested that remittances were responsive to certain variables in the short run. Lim and Simmons (2015), however, found no evidence of a long-run relationship between remittances and growth in CARICOM countries. In the study, which used panel cointegration techniques, the authors do find some evidence that remittances promote consumption in the long term.

Overall, the literature for Caribbean countries is inconclusive. The majority of studies have tested the remittance-development relationship in one direction, and the results are not consistent.

International Literature

The literature on the remittances-economic development relationship encompasses many studies. However, there is little agreement on the nature of this interaction. Is there a long-run relationship between the two? If so, how do they relate? Given the nature of remittance inflows, this interaction is likely to be heterogeneous across countries and time spans (Chami et al. 2008). Early microeconomic evidence, though often unable to provide long-run or aggregate implications, suggests that remittances contribute to greater spending on education, healthcare, and asset accumulation, as well as higher levels of conspicuous consumption (Taylor 1992; Adams 1998; Inter-American Development Bank 2004).

Chami, Fullenkamp, and Jahjah (2005) conducted one of the earlier studies looking at the impact of remittances on longer-term growth. The authors investigated this relationship for a panel of countries and found a negative relationship between the two, suggesting that remittances were not likely contributing to greater capital accumulation, but to consumption. However, in the literature that followed, one point emerged--the effect of remittances on growth might be conditional on other factors. Giuliano and Ruiz-Arranz (2009) found that remittances promoted growth in countries with less developed financial sectors. Catrinescu et al. (2009) discovered that remittances were more likely to contribute to growth in countries with stronger institutions. Garcia-Fuentes and Kennedy (2009) suggest that the effect is dependent on having a minimum level of human capital, while Feeny, Iamsiraroj, and McGillivray (2014) found that a relationship was more likely to exist for SIDS than for other developing countries. Despite this, more recent literature is contradictory and the nature of the relationship remains controversial (Rao and Hassan 2011; Senbeta 2013; Lim and Basnet 2017).

A number of studies examined the broader remittance-economic development relationship, but the empirical evidence is also mixed. Bjuggren, Dzansi, and Shukur (2010) provide evidence for the remittance-investment relationship in developing countries but found that it depends on the strength of local institutions. On the other hand, Yiheyis and Woldemariam (2016) looked at the remittance-investment relationship across four African countries and found differing impacts, with significant negative effects in two cases. Similarly, for financial development, Aggarwal, Demirguc-Kunt, and Martinez Peria (2011) found that remittances promoted finance across developing countries, while Brown, Carmignani, and Fayad (2013) showed that remittances may actually reduce domestic credit. Other studies unearthed evidence of remittances promoting government corruption (Berdiev, Kim, and Chang 2013), yet improving the quality of democratic institutions (Deonanan and Williams 2016); increasing human capital formation (Ngoma and Ismail 2013), and reducing public expenditure on health and education in the context of moral hazard (Ebeke 2012).

Several studies also provide evidence for reverse causality--from economic development to remittances. They do so in the context of determining the motives to remit and it is often in isolation from the impacts of remittances. Lueth and Ruiz-Arranz (2006) constructed a gravity model for remittances which showed they may align with the business cycle in the home country The authors found less evidence of altruism than might be expected. Chami, Fullenkamp, and Jahjah (2005), on the other hand, detected that remittances are compensatory in nature, given that they increased in response to lower home country incomes. Recent evidence suggests that this relationship is still complex. The World Bank (2015) found that remittances were not responsive to home country conditions, while McCracken, Ramlogan-Dobson, and Stack (2016) constructed a gravity model for the LAC region and found evidence of altruism and self-interest.


We are interested in assessing the long-run relationship between remittances and economic development in eleven Caribbean countries, keeping in mind that it may differ among them. In particular, remittances and economic development may be related in the long run or short run, and four types of causal linkages may exist in the long run--causality from remittances to economic development; causality from economic development to remittances; bi-directional causality; or no causal links. Accordingly, we use a country-specific approach to analyse each country and allow for heterogeneity in the remittance-economic development relationship. Following the economic development literature related to foreign capital inflows, we use real income per capita as our measure of economic development (see Ang 2010). We describe the model, data, and estimation strategy in detail below.


Based on the theoretical relationships described previously, we employ the endogenous growth framework of Rebelo (1991), where economic development ([ED.sub.t]) depends on total factor productivity and capital investment ([I.sub.t]). In turn, total factor productivity is a function of remittances ([R.sub.t]) and financial development ([F.sub.t]). Correspondingly, the long-run relationship of interest is specified as follows:

[ED.sub.t] = f([R.sub.t], [F.sub.t], [I.sub.t]) (1)

In this model, financial development and investment are included as controls. As such, remittances impact economic development through total factor productivity, in its role as a source of private financing.


We use annual data for eleven Caribbean countries covering differing periods (3) for each country, dictated by availability. We use Gross Domestic Product per capita as our economic development variable ([ED.sub.t]), personal remittances received per capita as our remittance variable ([R.sub.t]), and gross fixed capital formation per capita as our measure of investment ([I.sub.t]). All of these measures are expressed in constant 2005 US$ and we use the natural logarithms of all the variables for the analysis. For our financial development variable ([F.sub.t]), a composite measure is used, with its construction outlined in the following section. Table 1 gives the sources of data.

Measurement of financial development

Financial development is a multi-faceted process. Correspondingly, there is little consensus in the finance-growth literature on which proxy best reflects the process of financial development. For our analysis, we use principal component analysis (PCA) to construct an index of financial development from four commonly used indicators: liquid liabilities to GDP (M3-to-GDP ratio), private sector credit to GDP, deposit bank assets to total banking system assets, and bank deposits to GDP. (4) Liquid liabilities and bank deposits are both measures of the absolute size of the financial sector. However, the former reflects the degree to which the financial sector provides transaction services, while the latter indicates the ability of banks to attract savings. Deposit bank assets is a measure of the relative size of the financial sector and private sector credit is a measure of activity of financial intermediaries (Beck, Demirguc-Kunt, and Levine 2000). The use of PCA to construct an index of financial development follows recent examples in the financial development literature (Ang and McKibbin 2007; Ang 2009; Gries, Kraft, and Meierrieks 2011). Table 2 gives the results of the PCA analysis.

Our index of financial development is constructed from common components extracted from the original data. Each of these components explains a significant proportion of the standardised variance of the original series. For the majority of countries, only one component is needed (that is, one component explains most of the variance). The financial development index is calculated from these components as a linear combination of the four series, using weights from the respective eigenvector. These weights are obtained by dividing each value in the eigenvector by the sum of the absolute values in the eigenvector, so that they sum to unity. In two cases, however, we make use of two components to capture a greater proportion of variance of the underlying series (the first component alone did not sufficiently account for total variation). This resulted in the creation of two indices in each case. To obtain a single measure, each index was then assigned its own weight based on its portion of explained variance. The weights were obtained by dividing variance explained by each component by the total variance explained by the two components.


We use the ARDL approach of Pesaran, Shin, and Smith (2001) to test for cointegration and estimate long-run and short-run relationships. This method has several advantages. For one, it can be applied regardless of whether the variables are endogenous. Secondly, it can be applied even if the variables are fractionally integrated. Thirdly, it has desirable small sample properties relative to other cointegration methods. Lastly, long-run and short-run dynamics may be assessed simultaneously through an error-correction model.

Testing entails three steps for each country. The first step is to test for the existence of unit roots. While the ARDL procedure has the advantage of allowing for both I(0) and I(1) variables, the critical values involved in the bounds testing procedure become invalid in the presence of variables having an order of integration greater than one. We employ the Augmented Dickey-Fuller (ADF) test and the Phillips-Perron (PP) test to assess the integration properties of each series. The second step is to test for the presence of a long-run relationship among the variables in the model (that is, assess whether there is cointegration among the variables). We do so by employing the ARDL bounds test. Accordingly, we estimate the two equations below via OLS. The first equation corresponds to [ED.sub.t] being the dependent variable, whereas, the second equation corresponds to [R.sub.t] being the dependent variable. As such, we allow for different causal direction in assessing the linkages between [R.sub.t] and [ED.sub.t].

[mathematical expression not reproducible] (2)

[mathematical expression not reproducible] (3)

A is the difference operator, m - 1 is the lag length and the [[epsilon].sub.t] terms are the normally distributed error terms. The lag lengths for each equation are chosen to ensure that the residuals are normally distributed and lag structures minimise serial correlation. For each equation, we perform an F-test of the coefficients of the lagged level terms and compare against critical bounds values to determine whether they are jointly significant. For example, for Eq. (2) the null hypothesis is [H.sub.0]: [[alpha].sub.x] = [[alpha].sub.2] = [[alpha].sub.3] = [[alpha].sub.4] = 0. The F-statistic is then compared to the upper and lower bounds critical values corresponding to the number of observations used. If the F-statistic is less than the lower bound, the null hypothesis cannot be rejected and we conclude that no cointegration exists. If the F-statistic is greater than the upper bound, the null is rejected and we conclude that there is sufficient evidence that cointegration exists among the variables. If the statistic falls within the two bounds, the test is inconclusive. Given the relatively small samples used in this study, we use the Narayan (2005) critical values to perform the bounds test comparison.

If cointegration is found, the third step is to assess long-run causality. The ARDL approach allows reformulation of the equations where cointegration is found in restricted error correction form. For example, if cointegration is found when [ED.sub.t] is the dependent variable, we reformulate equation (2) as follows:

[mathematical expression not reproducible] (4)

where [ECT.sub.t-1] is the lagged error correction term, and, p, q, r and s are optimal lag lengths. These lag lengths are found by comparing across [(m + 1).sup.k+1] models and employing an information criterion to select the model that is optimal. m is as defined previously and k is the number of independent variables. The error correction term contains the long-run information. A significant coefficient on [ECT.sub.t-1] provides evidence that the independent variables in levels are 'forcing' the dependent variable (that is, evidence of long-run causality). We perform a likelihood ratio (LR) test to assess the significance of the coefficient on [ECT.sub.t-1].


The results of both unit root tests presented in Table 3 indicate that the majority of variables are either I(0) or I(1). In two cases, however, both tests suggest variables are I(2). In these two cases, the variables are transformed to I(1) variables by first-differencing prior to expressing in logarithmic form. Accordingly, all variables used in this analysis meet the integration criteria required to conduct the bounds test.

In the estimation of equations (2)-(3) for each country, models corresponding to lags of m = 2, m = 3 and m = 4 are estimated. The optimal models chosen correspond to the lag structure supported by Schwarz Bayesian Criterion (SBC) and meeting diagnostic tests criteria (minimizing normality and serial correlation). (5) Given the relatively short span of the data, models where m > 4 are not considered to conserve degrees of freedom. Table 4 presents the results of the bounds test corresponding to equations (2)-(3) along with critical values. In cases where we find cointegration, we estimate parsimonious error-correction models corresponding to equation (4). We present long-run causality results in Table 5 and long-run estimates in Table 6 corresponding to these error-correction models.

The cointegration results in Table 4 indicate that the variables are cointegrated in seven of the eleven countries examined in this study. Of the seven countries with cointegration, we find evidence of a single cointegrating relationship in five (Antigua & Barbuda, Dominica, St. Lucia, St. Vincent & the Grenadines and Jamaica), and two cointegrating relationships in two (St. Kitts & Nevis and Trinidad & Tobago). These results suggest the variables share a long-run relationship. In the remaining countries (Belize, Barbados, Grenada and Suriname), we find no evidence of cointegration for the models considered. It is possible, however, that remittances and economic growth may be related in the short-run. (6) The finding of cointegration for seven of the eleven countries examined presents new evidence on the existence of a long-run remittance-economic development relationship for the region. Our findings contradict recent studies that consider the region as homogenous group and whose analyses suggest an absence of any long-run relationship between remittances and growth (Feeny, Iamsiraroj, and Mcgillivray 2014; Lim and Simmons 2015). Our results complement those of Kumar (2013) who found evidence of a long-run relationship in Guyana.

The long-run results presented in Table 5 indicate different patterns of causality for the seven countries with cointegration. In four countries (Antigua & Barbuda, Dominica, St. Lucia and St. Vincent & the Grenadines), we find evidence of unidirectional causality to remittances. This suggests that economic development, financial development, and investment together influence remittance inflows. In Jamaica, we find evidence of unidirectional causality to economic development. This suggests that remittances, financial development, and investment together influence economic development. In the remaining two countries (St. Kitts & Nevis and Trinidad & Tobago), we find evidence of bi-directional causality. For these two, the results suggest that economic development, financial development, and investment together influence remittance inflows; and that remittances, financial development, and investment together influence economic development. Overall, the causality findings suggest that only in three countries (Jamaica, St. Kitts & Nevis and Trinidad & Tobago) are policies affecting remittances, financial development, and investment likely to influence economic development. On the other hand, in the majority of countries (Antigua & Barbuda, Dominica, St. Lucia, St. Vincent & the Grenadines, St. Kitts & Nevis and Trinidad & Tobago), policies affecting economic development, financial development, and investment are likely to influence remittances inflows.

While we find evidence of long-run relationships among the variables in seven countries and different patterns of causal linkages, the long-run estimates in Table 6 suggest even greater heterogeneity in the remittance-economic development relationship among the seven countries. For the three countries where we find causality to economic development, while the remittance variable is positive for each, only in one country is the remittance variable significant--Trinidad & Tobago. This suggests that changes in remittances promote economic development through a productivity channel in Trinidad & Tobago. However, while remittances may be promoting economic development in Jamaica and St. Kitts & Nevis, the mechanism may not be through improving total factor productivity. On the other hand, in the six countries where we find causality to remittances, the underlying motives behind remittance inflows appear to be different. In two countries (Dominica and St. Lucia), remittances and variables proxying for economic activity ([G.sub.t] or [I.sub.t]) share a significant negative relationship, suggesting that remittances may be primarily motivated by insurance motives. In three countries (St. Vincent & the Grenadines, St. Kitts & Nevis, and Trinidad & Tobago), remittances and variables proxying for economic activity share a significant positive relationship, suggesting that remittances may be primarily motivated by investment motives. In the case of Antigua & Barbuda, the coefficient on [G.sub.t] is negative and significant whereas the coefficient on [I.sub.t] is positive and significant. Collectively, the evidence for Antigua & Barbuda suggests both an insurance and investment motive behind remittances inflows.


In this article, we examine whether and how remittances may be linked to long-run economic development in eleven Caribbean countries. More specifically, we address two questions: Are remittances and economic development linked in the long run? And if they share a long-run relationship, how are they linked? These questions are relevant to providing insight into whether policies are required, and what types. These questions are important to the Caribbean region for at least two reasons. Firstly, remittances represent the second largest inflow of capital to the Caribbean region. Secondly, access to external financing is often cited as a significant constraint to economic development in Caribbean countries. We examine the remittance-economic development relationship within an endogenous growth model, where we control for financial development and investment. We use the recent ARDL approach to assess cointegration, estimate long-run relationships and determine long-run Granger causality in each country over the period 1975-2011.

We contribute new insights into the long-run remittance-economic development relationship in the Caribbean. Our results are fourfold. Firstly, we find evidence of a long-run relationship between economic development, remittances, financial development, and investment in seven of the eleven countries examined. Secondly, we find evidence of unidirectional causality to remittances in four countries (Antigua & Barbuda, Dominica, St. Lucia, and St. Vincent & the Grenadines); unidirectional causality to economic development in one country (Jamaica); and bi-directional causality to remittances and economic development in two countries (St. Kitts & Nevis and Trinidad & Tobago). Thirdly, for the three countries (Jamaica, St. Kitts & Nevis, and Trinidad & Tobago) where we find causality to economic development, the remittance variable is positive for each, but only significant in one (Trinidad & Tobago), and in this latter case, economically small, suggesting negligible impact through increasing total factor productivity. Lastly, in the six countries where we find causality to remittances, the underlying motives behind remittance inflows appear to be different. In two countries (Dominica and St. Lucia), we find remittances may be primarily motivated by altruistic motives; in three (St. Vincent & the Grenadines, St. Kitts & Nevis, and Trinidad & Tobago), we find remittances may be primarily motivated by investment motives; and in one (Antigua & Barbuda), the evidence suggests both altruistic and investment motives behind remittances inflows.

Overall, our results suggest that long-run remittance-economic development linkages exist, and the heterogeneity in these linkages suggests a more complex relationship than those found in previous macroeconomic studies for the region. Our findings contradict those of Lim and Simmons (2015), who found limited evidence that remittances and economic development share a long-run relationship. The differences in our findings are likely due to our country-specific approach and accommodation of different causal patterns. As such, including remittances in long-run development planning and encouraging greater inflows from the diaspora are important for policymakers seeking to increase their development impact, but such policies need to be considered on a country-specific basis.

At the same time, while we find that remittances are related to economic development in several countries, it is important to point out that we find little or no evidence that remittances directly increase economic development through greater factor productivity. That is not to say that remittances are not being invested in increasing human capital capacity or contributing to other factors that are conducive to long-run growth through productivity channels. It may be that the productivity losses from migration associated with remittances counteract the gains, or that the size of remittances is too small to provide enough gains. While further research would be required to more precisely determine why there are limited factor productivity gains, our results suggest that migration and remittances are not likely to be effective strategies for sustained economic development by themselves. They need to be considered as part of a more comprehensive development plan and linked to other strategies that can bring about increased technological capacity, the basis for sustained growth. In this regard, engagement of the diaspora may be crucial. For instance, policies encouraging exports to diaspora in developed countries may further stimulate expansion of the domestic export sector through market access and experience (learning-by-doing). Engagement of the diaspora, of course, is likely to be important in determining the types of goods and services desired, where they are desired and how best to market them. Similarly, closer ties with the diaspora may increase tourist arrivals, directly through more frequent visits by the diaspora, or through referrals by the diaspora to agents in the host country. Remittances, if channelled to formal financial sectors, may themselves provide the capital for investment needed to fund different growth strategies. How best to link the diaspora to particular developmental strategies and the mix of measures used are likely to be country-specific and relevant lines of further research.

What do these findings suggest about the role of remittances in long-run economic development and the underlying mechanisms at a more nuanced level? In the case of Jamaica where we find remittances are related to economic development, the effect is not likely through increasing factor productivity. At the same time, remittances do not appear to be motivated by altruistic or investment motives in the long run. Collectively, these findings indicate that remittances may more likely be part of an informal arrangement for migration. Remittances may be functioning as replacement income for family member(s) of the migrated. In turn, remittances stimulate the economy through expansion of domestic demand for goods and services, which is in line with findings by Stephenson and Wilsker (2016). This is very likely in the case of Jamaica, given the high rate of skilled emigration and limited opportunities for domestic employment (Jamaica has a high unemployment rate). An implication of this is that policies geared toward forming deeper bonds with the diaspora as a means of increasing the inflow and impact of remittances are likely to be more effective if coupled with policies targeting recipient households. In this regard, bringing recipients into the formal financial sector, along with providing financial literacy programmes may be helpful in reducing the consumption of recipient households, thereby increasing savings and mobilising remittances for investment.

In the case of St. Kitts & Nevis and Trinidad & Tobago, where we find evidence of causality to economic development, the effect is not likely through increasing factor productivity. In the case of St. Vincent & the Grenadines and Antigua & Barbuda, we find no discernible impact on ED. However, what is common among these countries is that we do find that remittances are motivated by investment. The increase in investment itself may lead to higher growth through diversification and diffusion of experience and expertise from the remitter. However, given that remittances are small in Trinidad & Tobago, we may not observe higher productivity over the period studied, simply due to the size of remittances. Or it may be that the impact of migration loss of skilled workers in St. Kitts & Nevis and St. Vincent & the Grenadines thwarts any growth effects, even though remittances are larger. Understanding why there is a difference in impact of investment-bound remittances will require further research to provide a more detailed policy. At any rate, a different array of policies than those in the Jamaican context is likely to be more effective in attracting more remittances and increasing the growth impact. Firstly, improving the investment climate is important. In this regard, making it easier to do business, improving the legal framework to ensure investor rights, and reducing corruption and crime are imperative. Secondly, marketing investment opportunities explicitly to the diaspora may be important. These policies are likely to be more effective if also coupled with policies that encourage investment by more experienced diaspora members, who not only bring capital, but get directly involved in the investment initiative through ownership. In this case, such diaspora members may be able to increase the knowledge base of the domestic economy through the introduction of new technology, which in turn can increase productive capacity.

Finally, in the case of Antigua & Barbuda, Dominica, and St. Lucia where we find evidence of altruistic remittances, we also find no evidence of a causal impact on growth. The suggestion here is that while a consumption smoothing mechanism is present, and likely to temper growth volatility, the reduction in volatility is not translating to greater economic growth. Perhaps remittances may be more effective if coordinated with other forms of capital aimed at reducing downturns. In particular, remittances may be more effective in reducing volatility if also complemented with and differentiated from other stability measures such as aid. This is important for policymakers to consider as Amuedo-Dorantes, Pozo and Vargas-Silva (2010) found that aid crowds out remittances. For instance, coupling the buffer provided by remittances during downturns, with programmes funded by aid to develop human capacity, may together be more effective at navigating and reducing downturns. In this case, remitters may not reduce remittance inflows in response to greater aid because they perceive aid being channelled to another purpose. Coupling remittances during downturns with financial access may also be more effective in relieving credit constraints to development, which are likely to be more binding during downturns. This would require the financial sector to use remittances as collateral or use remittance receipts as part of credit scoring. To that end, policies encouraging the use of remittances as collateral for investment may be important. To the extent such policy efforts are desirable, future research on the interaction between remittances and other mechanisms that provide stability and the relationship between remittances and finance are important in crafting more precise policies.


Adams, Richard H. Jr. 1998. "Remittances, Investment, and Rural Asset Accumulation in Pakistan." Economic Development and Cultural Change 47 (1): 155-73. doi:10.1086/452390.

Agarwal, Reena, and Andrew W. Horowitz. 2002. "Are International Remittances Altruism or Insurance? Evidence from Guyana Using Multiple-Migrant Households." World Development 30 (11): 2033-44. doi:10.1016/S0305-750X(02)00118-3.

Aggarwal, Reena, Asli Demirguc-Kunt, and Maria Soledad Martinez Peria. 2011. "Do Remittances Promote Financial Development?" Journal of Development Economics 96 (2): 255-64. doi:10.1016/j.jdeveco.2010.10. 005.

Alleyne, Dillon. 2006. "Motivations to Remit in CARICOM: A GMM Approach." Social and Economic Studies 55 (3): 69-87.

Alleyne, Dillon, Claremont Kirton, and Mark Figueroa. 2008. "Macro-economic Determinants of Migrant Remittances to Caribbean Countries: Panel Unit Roots and Cointegration." The Journal of Developing Areas 41 (2): 137-53. doi:10.1353/jda.2008.0015.

Alleyne, Dillon, Claremont Kirton, Georgia McLeod, and Mark Figueroa. 2008. "Short-Run Macroeconomic Determinants of Remittances to Jamaica: A Time Varying Parameter Approach." Applied Economics Letters 15 (8): 629-34. doi:10.1080/13504850600721965.

Amuedo-Dorantes, Catalina, Annie Georges, and Susan Pozo. 2010. "Migration, Remittances, and Children's Schooling in Haiti." The Annals of the American Academy of Political and Social Science 630: 224-44. doi: 10.1177/0002716210368112.

Amuedo-Dorantes, Catalina, and Susan Pozo. 2004. "Workers' Remittances and the Real Exchange Rate: A Paradox of Gifts." World Development 32 (8): 1407-17. doi:10.1016/j.worlddev.2004.02.004.

Amuedo-Dorentes, Catalina, and Susan Pozo. 2006. "Migration, Remittances, and Male and Female Employment Patterns." American Economic Review 96 (2): 222-26.

Amuedo-Dorantes, Catalina, Susan Pozo, and Carlos Vargas-Silva. 2010. "Remittances in Small Island Developing States." The Journal of Development Studies 46 (5): 941-60. doi:10.1080/00220381003623863.

Ang, James B. 2009. "Financial Development and the FDI-Growth Nexus: The Malaysian Experience." Applied Economics 41 (13): 1595-601. doi: 10.1080/00036840701222553.

Ang, James B. 2010. "Does Foreign Aid Promote Growth? Exploring the Role of Financial Liberalization." Review of Development Economics 14 (2): 197-212. doi:10.1111/j.l467-9361.2010.00547.x.

Ang, James B., and Warwick J. McKibbin. 2007. "Financial Liberalization, Financial Sector Development and Growth: Evidence from Malaysia." Journal of Development Economics 84 (1): 215-33. doi:10. 1016/j.jdeveco.2006.11.006.

Barajas, Adolfo, Ralph Chami, Connel Fullenkamp, Michael Gapen, and Peter Montiel. 2009. "Do Workers' Remittances Promote Economic Growth?" 09/153. IMF Working Paper. Washington, DC.

Beck, Thorsten, Asli Demirguc-Kunt, and Ross Levine. 2000. "A New Database on Financial Development and Structure." World Bank Economic Review 14 (3): 597-605. doi:10.1093/wber/14.3.597.

Berdiev, Aziz N., Yoonbai Kim, and Chun-Ping Chang. 2013. "Remittances and corruption." Economics Letters 118 (1): 182-85.

Beuermann, Diether W., Inder Jit Ruprah, and Ricardo E. Sierra. 2014. "Remittances as a Safety Net in Jamaica." Washington, DC: Inter-American Development Bank.

Bjuggren, Per-Olof, James Dzansi, and Ghazi Shukur. 2010. "Remittances and Investment." Paper No. 216. CESIS Electronic Working Paper Series.

Bredl, Sebastian. 2011. "Migration, Remittances and Educational Outcomes: The Case of Haiti." International Journal of Educational Development 31 (2): 162-68. doi:10.1016/j.ijedudev.2010.02.003.

Brown, Richard PC, Fabrizio Carmignani, and Ghada Fayad. 2013. "Migrants' Remittances and Financial Development: Macro- and Micro-Level Evidence of a Perverse Relationship." The World Economy 36 (5): 636-60. doi:10.1111/twec.l2016.

Bussolo, Maurizio, and Denis Medvedev. 2008. "Do Remittances Have a Flip Side? A General Equilibrium Analysis of Remittances, Labor Supply Responses and Policy Options for Jamaica." Journal of Economic Integration 23 (3): 734-64.

Campbell, Trevor. 2003. "Determinants of Foreign Inflows from Current Transfers into Barbados." Savings and Development 27 (2): 119-34.

Campbell, Trevor. 2009. "Workers remittances and their impact on economic growth in Barbados." In Money Affairs, Vol.XXII, Number 2, edited by CEMLA, 181-94. Centre for Latin American Monetary Studies.

Catrinescu, Natalia, Miguel Leon-Ledesma, Matloob Piracha, and Bryce Quillin. 2009. "Remittances, Institutions and Economic Growth." World Development 37 (1): 81-92. doi:10.1016/j.worlddev.2008.02.004.

Chami, Ralph, Adolfo Barajas, Thomas Cosimano, Connel Fullenkamp, Michael Gapen, and Peter Montiel. 2008. "Macroeconomic Consquences of Remittances." Washington, DC: International Monetary Fund.

Chami, Ralph, Connel Fullenkamp, and Samir Jahjah. 2005. "Are Immigrant Remittance Flows a Source of Capital for Development?" IMF Staff Papers 52 (1): 55-81. doi:10.2307/30035948.

Clarke, George R. G., and Scott J. Wallsten. 2003. "Do Remittances Act Like Insurance? Evidence From a Natural Disaster in Jamaica." Rochester, NY: SSRN.

Deonanan, Regan, and Kevin Williams. 2016. "The Effect of Remittances on Democratic Institutions." Applied Economics 49 (5): 403-16. doi:10. 1080/00036846.2016.1200180.

Ebeke, Christian Hubert. 2012. "Do Remittances Lead to a Public Moral Hazard in Developing Countries? An Empirical Investigation." The Journal of Development Studies 48 (8): 1009-25. doi:10.1080/00220388.2011.615918.

Economic Commission for Latin Amercia and the Caribbean (ECLAC). 2013. "Development Paths in the Caribbean." Trinidad and Tobago: ECLAC.

Feeny, Simon, Sasi Iamsiraroj, and Mark Mcgillivray. 2014. "Remittances and Economic Growth: Larger Impacts in Smaller Countries?" The Journal of Development Studies 50 (8): 1-12. doi:10.1080/00220388.2014.895815.

Garcia-Fuentes, Pablo A., and P. Lynn Kennedy. 2009. "Remittances and Economic Growth in Latin America and the Caribbean: The Impact of Human Capital Development." In Southern Agricultural Economics Association Annual Meeting. Atlanta, Georgia.

Giuliano, Paola, and Marta Ruiz-Arranz. 2009. "Remittances, Financial Development, and Growth." Journal of Development Economics 90 (1): 144-52. doi:10.1016/j.jdeveco.2008.10.005.

Gries, Thomas, Manfred Kraft, and Daniel Meierrieks. 2011. "Financial Deepening, Trade Openness and Economic Growth in Latin America and the Caribbean." Applied Economics 43 (30): 4729-39. doi: 10.1080/00036846.2010.498352.

Hurley, Gail. 2015. "Financing for Development and Small Island Developing States: A Snapshot and Ways Forward." New York, NY: United Nations Development Program (UNDP).

Inter-American Development Bank (IDB). 2004. "Sending Money Home: Remittance to Latin America and the Caribbean." Washington, DC: IDB.

International Organization for Migration (IOM). 2014. "Handbook to Develop Projects on Remittances." Brussels, Belgium and Madrid, Spain: IOM.

Itzigsohn, Jose. 1995. "Migrant Remittances, Labor Markets, and Household Strategies: A Comparative Analysis of Low-Income Household Strategies in the Caribbean Basin." Social Forces 74 (2): 633-55. doi:10.2307/2580495.

Jackman, Mahalia, Roland Craigwell, and Winston Moore. 2009. "Economic Volatility and Remittances: Evidence from SIDS." Journal of Economic Studies 36 (2): 135-46. doi:10.1108/01443580910955024.

Jadotte, Evans, and Xavier Ramos. 2016. "The Effect of Remittances on Labour Supply in the Republic of Haiti." The Journal of Development Studies 52 (12): 1810-25. doi:10.1080/00220388.2016.1156089.

Kim, Namsuk. 2007. "The Impact of Remittances on Labor Supply: The Case of Jamaica." WPS4120. World Bank Policy Research Working Paper.

Kirton, Claremont. 2011. "Migration & Remittances Trends: The Caribbean Experience." In Caribbean Remittance Forum on Enhancing the Efficiency and Integrity of Remittance Transfers through Effective Regulatory and Supervisory Systems in the Caribbean. Kingston, Jamaica.

Kumar, Ronald Ravinesh. 2013. "Remittances and Economic Growth: A Study of Guyana." Economic Systems 37 (3): 462-72. doi:10.1016/j.ecosys.2013.01.001.

Li, Meng. 2013. "Financing for Sustainable Development in Small Island Developing States." New York, NY: UN-DESA.

Lim, Sokchea, and Hem C. Basnet. 2017. "International Migration, Workers' Remittances and Permanent Income Hypothesis." World Development 96: 438-50. doi:10.1016/j.worlddev.2017.03.028.

Lim, Sokchea, and Walter O. Simmons. 2015. "Do Remittances Promote Economic Growth in the Caribbean Community and Common Market?" Journal of Economics and Business 77 (C): 42-59. doi:10. 1016/j.jeconbus.2014.09.001.

Lueth, Erik, and Marta Ruiz-Arranz. 2006. "A Gravity Model of Workers' Remittances." 06/290. IMF Working Paper. Washington, DC.

McCracken, Scott, Carlyn Ramlogan-Dobson, and Marie M. Stack. 2016. "A Gravity Model of Remittance Determinants: Evidence from Latin America and the Caribbean." Regional Studies 51 (5): 737-49. doi:10. 1080/00343404.2015.1133904.

Mishra, Prachi. 2006. "Emigration and Brain Drain: Evidence From the Caribbean." WP/06/25. IMF Working Paper.

Moore, Alvon, and Kevin Greenidge. 2008. "Determinants and Volatility of Remittances in the Caribbean." Bridgetown: Central Bank of Barbados.

Mundaca, Gabriela. 2009. "Remittances, Financial Markets Development and Economic Growth: The Case of Latin America and Caribbean." Review of Development Economics 13 (2): 288-303. doi:10.1111/j.l467-9361.2008.00487.x.

Narayan, Paresh Kumar. 2005. "The Saving and Investment Nexus for China: Evidence from Cointegration Tests." Applied Economics 37 (17): 1979-90. doi:10.1080/00036840500278103.

Ngoma, Abubakar Lawan, and Normaz Wana Ismail. 2013. "Do Migrant Remittances Promote Human Capital Formation? Evidence from 89 Developing Countries." Migration and Development 2 (1): 106-16. doi: 10.1080/21632324.2013.786883.

Nsiah, Christian, and Bichaka Fayissa. 2013. "Remittances and Economic Growth in Africa, Asia, and Latin American-Caribbean Countries: A Panel Unit Root and Panel Cointegration Analysis." Journal of Economics and Finance 37 (3): 424-41. doi:10.1007/sl2197-011-9195-6.

Pesaran, M. Hashem, Yongcheol Shin, and Richard J. Smith. 2001. "Bounds Testing Approaches to the Analysis of Level Relationships." Journal of Applied Econometrics 16 (3): 289-326. doi:10.1002/jae.616.

Pesaran, M. Hashem, and Yongcheol Shin. 1999. "An Autoregressive Distributed-Lag Modelling Approach to Cointegration Analysis." In Econometrics and Economic Theory in the 20th Century: The Ragnar Frisch Centennial Symposium, edited by Steinar Strom, 371-413. Cambridge University Press. doi:10.1017/CCOL0521633230.011.

Poirine, Bernard. 1997. "A Theory of Remittances as an Implicit Family Loan Arrangement." World Development 25 (4): 589-611. doi:10.1016/S0305-750X(97)00121-6.

Ramirez, Miguel D. 2013. "Do Financial and Institutional Variables Enhance the Impact of Remittances on Economic Growth in Latin America and the Caribbean? A Panel Cointegration Analysis." International Advances in Economic Research 19 (3): 273-88. doi:10.1007/s11294-013-9407-2.

Rao, B. Bhaskara, and Gazi Mainul Hassan. 2011. "A Panel Data Analysis of the Growth Effects of Remittances." Economic Modelling 28 (1-2): 701-9. doi:10.1016/j.econmod.2010.05.011.

Rebelo, Sergio. 1991. "Long-Run Policy Analysis and Long-Run Growth." Journal of Political Economy 99 (3): 500-521. doi:10.1086/261764.

Senbeta, Aberra. 2013. "Remittances and the Sources of Growth." Applied Economics Letters 20 (6): 572-80. doi:10.1080/13504851.2012.718057.

Stephenson, Andrew V., and Amanda Wilsker. 2016. "Consumption Effects of Foreign Remittances in Jamaica." International Advances in Economic Research 22 (3): 309-20. doi:10.1007/s11294-016-9595-7.

Taylor, J. Edward. 1992. "Remittances and Inequality Reconsidered: Direct, Indirect, and Intertemporal Effects." Journal of Policy Modeling 14 (2): 187-208. doi:10.1016/0161-8938(92)90008-Z.

United Nations Conference on Trade and Development (UNCTAD). 2013. "Maximizing the Developmental Impact of Remittances." Geneva, Switzerland.

United Nations (UN). 2015a. "Trends in International Migrant Stock: Migrants by Destination and Origin."

--. 2015b. "Trends in International Migrant Stock: The 2015 Revision."

--. 2016. "National Accounts Main Aggregates Database."

World Bank. 2015. "Global Economic Prospects, January 2015: Having Fiscal Space and Using It." Washington, DC: World Bank.

--. 2016a. "Investing Back Home: The Potential Economic Role of the Caribbean Diaspora." Washington, DC: World Bank.

--. 2016b. "Global Financial Development."

--. 2017. "World Development Indicators."

Yiheyis, Zelealem, and Kasahun Woldemariam. 2016. "The Effect of Remittances on Domestic Capital Formation in Select African Countries: A Comparative Empirical Analysis." Journal of International Development 28: 243-65. doi:10.1002/jid.3194.

(1) Authors' calculations based on data from the United Nations (2015a, 2015b).

(2) Special note on Suriname: The ~5% measure of FDI/GDP is actually negative. It occupies the positive half of the graph for convenience.

(3) Antigua & Barbuda, Grenada, St. Vincent & the Grenadines (1986-2011), Belize (1984-2011), Barbados, St. Kitts & Nevis (1980-2011), Dominica (1977-2011), Jamaica (1976-2011), St. Lucia (1983-2011), Suriname (1977-2010), Trinidad & Tobago (1975-2009).

(4) The indicators that are expressed as a percentage of GDP are obtained in a deflated format, which helps to avoid the stock-flow problem (Beck, Demirgiig-Kunt, and Levine 2000).

(5) Optimal lag length determination and diagnostic tests available upon request.

(6) The focus of this study is on the long-run relationship between remittances and economic development. As such, we do not explore short-run linkages in cases where cointegration is not found.

Remesas y Desarrollo Economico: Evidencia del Caribe

Regan Deonanan y Benjamin Ramkissoon

Este articulo examina empiricamente si las remesas afectan el desarrollo economico en once paises del Caribe durante el periodo 1975-2011 y como lo hacen. Empleamos la prueba de limites y el marco ARDL para evaluar la existencia de relaciones a largo plazo y la causalidad, y adoptamos un enfoque especifico del pais para acomodar las diferencias entre los paises. Encontramos evidencia de relaciones causales heterogeneas a largo plazo en siete paises. Nuestros hallazgos contradicen los estudios recientes sobre el impacto de las remesas sobre el desarrollo en el Caribe, y sugieren que las remesas pueden tener un rol, junto con politicas relevantes, en el avance economico de la region.

Palabras clave: Caribe, desarrollo economico, remesas, desarrollo financiero, ARDL, prueba de limites

Les Envois de Fonds et le Developpement Economique: l'Experience des Caraibes

Cet article examine empiriquement si les envois de fonds affectent le developpement economique dans onze pays des Caraibes au cours de la periode 1975-2011 et comment ils le font. Nous utilisons le test des limites et le cadre ARDL pour evaluer l'existence de relations a long terme et la causalite, et adoptons une approche specifique au pays pour tenir compte des differences entre les pays. Nous trouvons des preuves de relations causales a long terme heterogenes dans sept pays. Nos resultats contredisent les etudes recentes sur l'impact des envois de fonds sur le developpement dans les Caraibes et suggerent que les envois de fonds peuvent jouer un role, avec les politiques pertinentes, dans le developpement economique de la region.

Mots-cles: Caraibes, developpement economique, transferts de fonds, developpement financier, ARDL, test des limites Trinite-et-Tobago
Table 1: Definitions and data sources

Variable                                    Data Source

Real GDP, per capita (constant 2005 US$)    World Bank (2017);
                                            Jamaica--United Nations
Real personal remittances received,         Calculated from World Bank
per capita (constant 2005 US$)              (2017).
Real gross fixed capital formation,         Calculated from World Bank
per capita (constant 2005 US$)              (2017).
Liquid Liabilities (% of GDP)               World Bank (2016b)
Private Credit by Deposit Money Banks       World Bank (2016b)
(% of GDP)
Deposit Money Bank Assets to Deposit Money  World Bank (2016b)
Bank Assets and Central Bank Assets (%)
Bank Deposits (% of GDP)                    World Bank (2016b).

Table 2: Results of the principal components analysis

Country            Components  Variance   Eigenvectors
                                          M        C        A

Antigua & Barbuda  1           94.27%     0.5113   0.4921   0.4844
Belize             2           76.12%     0.5613   0.5583   0.2257
                               22.28%    -0.1521  -0.1101   0.9738
Barbados           1           89.59%     0.5266   0.5256   0.4096
Dominica           1           95.70%     0.5168   0.4967   0.4686
Grenada            1           55.06%     0.5068   0.4877   0.4970
Jamaica            2           26.79%     0.6240   0.4916  -0.0683
                               81.64%     0.2132  -0.2893   0.9254
St. Kitts & Nevis  1           93.27%     0.5226   0.4737   0.4784
St. Lucia          1           82.02%     0.5105   0.5011   0.4761
Suriname           1           81.79%     0.5472   0.4870  -0.4159
Trinidad & Tobago  1           79.87%     0.5432   0.5387  -0.3587
St. Vincent & the  1                      0.5093   0.5426   0.3978


Antigua & Barbuda   0.5117
Belize              0.5677
Barbados            0.5279
Dominica            0.5163
Grenada             0.5082
Jamaica             0.6036
St. Kitts & Nevis   0.5231
St. Lucia           0.5116
Suriname            0.5390
Trinidad & Tobago   0.5348
St. Vincent & the   0.5365

Notes: M represents liquid liabilities to GDP in logarithmic form, C is
private credit to GDP in logarithmic form, A is the deposit money bank
assets ratio in logarithmic form and D is bank deposits to GDP in
logarithmic form.

Table 3: PP and ADF unit root tests

Country            [ED.sub.t]  [R.sub.t]  [F.sub.t]  [I.sub.t]

Antigua & Barbuda  I(1)/I(0)   I(1)/I(0)  I(1)/I(1)  I(2)/I(0)
Belize             I(2)/I(2)   I(1)/I(1)  I(1)/I(1)  I(1)/I(1)
Barbados           I(1)/I(0)   I(1)/I(1)  I(1)/I(1)  I(1)/I(1)
Dominica           I(1)/I(1)   I(0)/I(0)  I(1)/I(1)  I(1)/I(1)
Grenada            I(1)/I(1)   I(1)/I(1)  I(1)/I(1)  I(1)/I(1)
Jamaica            I(1)/I(1)   I(0)/I(0)  I(1)/I(1)  I(1)/I(1)
St. Kitts & Nevis  I(1)/I(1)   I(1)/I(0)  I(1)/I(1)  I(1)/I(1)
St. Lucia          I(1)/I(2)   I(0)/I(0)  I(1)/I(1)  I(1)/I(1)
Suriname           I(1)/I(1)   I(1)/I(0)  I(1)/I(1)  I(1)/I(1)
Trinidad & Tobago  I(2)/I(2)   I(1)/I(1)  I(1)/I(1)  I(1)/I(1)
St. Vincent & the
Grenadines         I(1)/I(1)   I(1)/I(0)  I(1)/I(1)  I(1)/I(1)

Notes: The first and second entries for each variable correspond to
results of the PP and ADF tests respectively. For level variables, a
trend and intercept is included in the test. For first differenced
variables, an intercept is included. For the ADF test, the Schwarz
criterion is used to select the lag length, with a maximum lag of 4
considered. For the PP test, the Bartlett kernel method was used with
the Newey-West automatic bandwidth selection.

Table 4: Results of the bounds test for cointegration

Country                   [ED.sub.t]          [R.sub.t]
Panel A: Bounds test      F-stat       Coint  F-stat       Coint

Antigua & Barbuda         3.211        No     4.160 (*)    Yes
Belize                    1.627        No     3.614        No
Barbados                  1.565        No     1.727        No
Dominica                  3.417        No     8.205 (***)  Yes
Grenada                   1.145        No     3.330        No
Jamaica                   7.127 (***)  Yes    3.331        No
St. Kitts & Nevis         4.493 (*)    Yes    5.021 (**)   Yes
St. Lucia                 2.440        No     7.540 (***)  Yes
Suriname                  2.631        No     2.613        No
Trinidad & Tobago                      Yes    5.671 (**)   Yes
St. Vincent & the         4.307 (*)    No     4.484 (*)    Yes
Grenadines                2.904
Panel B: Critical values  1%                  5%
Lower bounds (n=30)       5.333               3.710
Upper bounds (n=30)       7.063               5.018
Lower bounds (n=35)       5.198               3.615
Upper bounds (n=35)       6.845               4.913

Country                   [F.sub.t]           [I.sub.t]
Panel A: Bounds test      F-stat       Coint  F-stat     Coint

Antigua & Barbuda          2.129       No     4.194 (*)  Yes
Belize                     2.960       No     2.068      No
Barbados                   0.704       No     3.590      No
Dominica                   1.338       No     4.578 (*)  Yes
Grenada                    3.389       No     0.392      No
Jamaica                    5.785 (**)  Yes    6.801 (*)  Yes
St. Kitts & Nevis          1.704       No      (*)       No
St. Lucia                  3.255       No     1.302      No
Suriname                   2.155       No     2.473      No
Trinidad & Tobago          3.855       No     4.069      No
St. Vincent & the          5.314 (**)  Yes    2.973      No
Grenadines                                    3.013
Panel B: Critical values  10%
Lower bounds (n=30)        3.008
Upper bounds (n=30)        4.150
Lower bounds (n=35)        2.958
Upper bounds (n=35)        4.100

Notes: The critical values are the Narayan (2005) critical values for
the unrestricted intercept case where k = 3, with k being the number of
independent regressors and n being the number of observations. (***),
(**) and (*) indicate that the F-statistics are above the 1%, 5% and
10% upper bounds, respectively.

Table 5: Weak exogeneity test results

Country                       Model             LR stat      p-value

Antigua & Barbuda             R(2,1,0,l)        21.06 (***)  0.000
Antigua & Barbuda             I(2,0,1,2)        17.89 (***)  0.000
Dominica                      R(2,0,0,2)        28.43 (***)  0.000
Dominica                      I(4,4,3,3)        31.90 (***)  0.000
Jamaica                       ED(1,1,0,2)       25.05 (***)  0.000
Jamaica                       F(3,3,2,2)        33.10 (***)  0.000
Jamaica                       I(2,1,1,0)        23.94 (***)  0.000
St. Kitts & Nevis             ED(1,0,0,4)       28.79 (***)  0.000
St. Kitts & Nevis             R(2,0,0,2)        21.58 (***)  0.000
St. Lucia                     R(1,3,2,3)        49.80 (***)  0.000
Trinidad & Tobago             [DELTA]ED(1,0,1,  15.84 (***)  0.000
Trinidad & Tobago             0)                28.76 (***)  0.000
St. Vincent & the Grenadines  R(2,3,2,3)        24.32 (***)  0.000
St. Vincent & the Grenadines  R(2,2,2,0)        29.14 (***)  0.000

Country                       Long-run

Antigua & Barbuda             ECT [right arrow] R
Antigua & Barbuda             ECT [right arrow] I
Dominica                      ECT [right arrow] R
Dominica                      ECT [right arrow] I
Jamaica                       ECT [right arrow] ED
Jamaica                       ECT [right arrow] F
Jamaica                       ECT [right arrow] I
St. Kitts & Nevis             ECT [right arrow] G
St. Kitts & Nevis             ECT [right arrow] R
St. Lucia                     ECT [right arrow] R
Trinidad & Tobago             ECT [right arrow] AED
Trinidad & Tobago             ECT [right arrow] R
St. Vincent & the Grenadines  ECT [right arrow] R
St. Vincent & the Grenadines  ECT [right arrow] F

Notes: The values in parentheses are the optimal lag lengths p, q, r
and s. They are chosen by information criteria. The values in the 'LR
stat' column are the likelihood ratio test statistics for the error
correction term. (***) represents the 1% level of significance.

Table 6: ARDL long-run equations

Country         Long-run equation

Antigua &       [R.sub.t] = -4.85[ED.sub.t] (**) + 3.64[F.sub.t] (**) +
                1.84[I.sub.t] (**)
Barbuda         [I.sub.t] = 2.47[ED.sub.t] (***) + 0.04[R.sub.t]
                - 0.91[F.sub.t] (**)
Dominica        [R.sub.t] = 0.34[ED.sub.t] - 0.40[F.sub.t]
                - 0.44[I.sub.t] (**)
                [I.sub.t] = -0.38[ED.sub.t] + 0.03[R.sub.t] +
                3.38[F.sub.t] (*)
Jamaica         [ED.sub.t] = -0.01[R.sub.t] - 0.05[F.sub.t] +
                0.47[I.sub.t] (***)
                [F.sub.t] = -0.02[ED.sub.t] + 0.18[R.sub.t] (***)
                - 0.59[I.sub.t] (**)
                [I.sub.t] = 2.09[ED.sub.t] (***) + 0.03[R.sub.t]
                - 0.03[F.sub.t]
St. Kitts &     [ED.sub.t] = 0.01[R.sub.t] - 0.25[F.sub.t] +
Nevis           [R.sub.t] = 2.38[ED.sub.t] (***) + 0.42[F.sub.t]
                - 0.83[I.sub.t]
St. Lucia       [R.sub.t] = -12.26[DELTA][ED.sub.t] (***)
                - 1.40[F.sub.t] + 0.03[I.sub.t]
Trinidad &      [DELTA][ED.sub.t] = 0.01[R.sub.t] (**) - 0.10[F.sub.t]
                + 0.03[I.sub.t]
Tobago          [R.sub.t] = 41.75[DELTA][ED.sub.t] (***) +
                7.78[F.sub.t] (**) - 2.66[I.sub.t]
St. Vincent &   [R.sub.t] = 4.28[ED.sub.t] (*) + 2.04[F.sub.t]
                - 3.16[I.sub.t]
the Grenadines  [F.sub.t] = -0.72[ED.sub.t] (**) + 0.16R, (***) +

(*) Notes: (***), (**) and (*) indicate rejection of the null at the
1%, 5% and 10% levels, respectively.
COPYRIGHT 2018 University of the West Indies, Sir Arthur Lewis Institute of Social and Economic Studies
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2018 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Deonanan, Regan; Ramkissoon, Benjamin
Publication:Social and Economic Studies
Article Type:Report
Geographic Code:50CAR
Date:Jun 1, 2018
Previous Article:Energy Consumption and Economic Development in Caribbean SIDS.
Next Article:Do SPS and TBT regulations inhibit Guyana's Food and Agriculture Exports to CARICOM markets?

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters