Remittance And Financial Development

Cam Nguyen Karla La Rosa Graduate Seminar Prof Veitch December 1, 2006 Does Remittance have a positive impact on Financial Development? The two main capital inflows in developing countries are foreign direct investments (FDI) and remittances. For the past two decades the supply of remittances has been increasing in developing countries and has become a steady flow of income for the families of the migrant workers (see appendix Graph 1 and Graph 2). Remittance in developing countries has been seen as source of income, sometimes primary or supplemental income, for families, especially in rural areas. Remittances may also be sometimes vital for developing countries since it does help the economy of a developing nation by the additional cash flow received. According to previous papers and studies, remittances can help alleviate poverty in developing countries. As such, research in the topic of remittances expanded due to the rapid increase in remittance income to remittance receiving countries (RRC). Most remittance receiving countries are developing countries. People from these countries migrate to seek for better job opportunities and higher compensation. Many of the papers we found focused on the effects of remittances on household consumption, education and poverty alleviation. The continuous increase of remittances in developing countries and its positive contribution to growth and poverty alleviation led to a number of interesting studies; studies that directly link remittances to poverty alleviation, studies that indirectly link remittances to growth and studies that discusses the relationship because remittances and financial development.

The main question that we pose in this research is -- do

remittances lead to or promote financial development? Many papers have been written about financial development and how it can promote growth and reduce poverty. Some papers have also made mention that a more developed financial system can help make better use of remittances and lead to growth. Our overall conclusion is that financial market development can influence the long-run effects of remittances on the economies chosen for our analysis. Our results suggest that making financial services more generally available than now should lead to even better use of remittances, so as to boost growth in these countries (Mundaca, 2005). This topic about remittances and financial development seems to be a fairly new topic that further needs to be further explored. We wish to explore this topic in a different light by looking at it in a different perspective. Mundaca’s study of remittances looked at it in a perspective wherein a developed financial market can lead to more remittances. Financial institutions such as banks may improve their services and provide new services that can cater to the needs of the recipients.

Our paper will focus on the effect of remittances on

financial development- mainly, does remittance positively impact financial development.

Literature Review There are numerous papers that deal with remittance and financial development separately starting from as early as the 1970’s which tries to explain qualitative results of the impact of remittance on growth and/or financial development. However, papers involving remittance and financial development had only been dealt with empirically only fairly recent- the earliest paper started in this decade. Empirical researches on this

topic is fairly new and have yielded mixed results- some research say remittance has a positive impact some say negative on growth via financial development, current accounts, etc. Reviewing past research in regards to remittances which start as early as in the 70’s until now, one can see that the idea and sentiments of remittance have changed throughout the years. Research in the topic of remittances expanded in the 80’s due to rapid increase in the remittance income to remitting receiving countries (RRC). The initial feelings about remittances in the 70’s and early 80’s seem to be that remittances are not beneficial. But in the later part of the 80’s and to the present, another voice started emerging and viewed remittances in a different light- a positive one. In the recent years, many research in this field focus on empirical evidence rather than theoretical arguments like that of the 70’s and early 80’s. There are positive and negative effects of everything- including remittances. Many who deem remittances as a negative solution to poverty seem to adopt similar reasoning/arguments. They argue that: Remittances increase dependency, contribute to economic and political instability and development distortion, and lead to economic decline that overshadows a temporary advantage for a fortunate few.1

Aside from the arguments stated above, many also believe that remittances offer little contribution to development since the amount of capital being remitted will most likely

1

Keely, Charles B. and Tran, Bao Nga. “Remittance from Labor Migration: Evaluations, Performance, and Implications”. International Migration Review, Vol.23, No.3. Autumn 1989. pp.500-525

be used for consumption and only marginally productive enterprises2. Charles B.Keely and Bao Nga Tran in the late 1980’s looked at the labor markets in Europe and the Middle East to test whether or not remittance has a positive or negative impact on the labor force and also economic conditions. They did not find that remittance increases dependency but also stated that the contribution of remittances to economic performance should not be overstated. The disadvantages that some researchers point out do sound alarming and their arguments do make sense but when we analyze tools, we need to see if the benefits outweigh the cost of implementing the tool. Researchers who are opposite side of the argument that are in support of remittance points out that the argument against remittance tends to “focus attention on the immediate use of remittance income” and is ignoring “the considerable stimulus it provides to indigenous industries, as well as its contribution to the supply of loanable funds, i.e. investment capital”3. Although consumption may be seen as a negative impact by researchers who are not in support of remittance, one must not undermine its multiplier effect. Other researchers point out that consumption behavior has multiplier effects that increase the demand for goods and services, as well as indirect investments. This is true especially when remittances are used for health, education, and shelter, which impact human development4 (Bagasao, 2004).

2

Castles and Kosack 1973 Stahl, Charles B. 1999 4 Bagasao, 2004. 3

Stahl agrees with the statement above and also states in his paper that the expenditures from remittance income generate a multiplier effect that will positively impact the economy5. A study conducted by four ADB consultants spearheaded by Ildefonso F. Bagasao focused on remittances by Filipino overseas workers, “Enhancing the Efficiency of Overseas Filipino Workers Remittances”, published by the Asian Development Bank. The objectives of the study are the following: (1) reviewed the flows of remittances in the Philippines (2) identified the constraints in policy, regulatory and institutional frameworks that impact these flows and (3) developed proposals that will increase remittances (if possible), facilitate the shift of remitting money from the informal to the formal sector and encourage the use of remittance proceeds for sustainable poverty reduction. The paper provides an in-depth analysis of the negative and positive effects of remittances in the country. Several foreseen negative impacts of remittances are excessive consumption, the “dependency syndrome”, irregular or undocumented workers, human rights violations and brain drain. Consumption becomes a negative effect when remittances are not put into productive means such as investment in human capital, infrastructure or profit generating activities. The so-called “dependency syndrome” takes effect when beneficiaries (families) of remittances rely on it too much that they simply fail to participate in productive endeavors6. Aside for the need to earn a higher income abroad, the demand for cheap labor by developed countries is another reason why

5 6

Stahl, Charles B. 1999 Bagasao, July 2004.

undocumented workers seek jobs abroad. This study was very useful in helping us understand the underpinnings of remittances. As mentioned earlier, empirical research on the effects of remittances on growth started early in this decade and continue on so. These empirical researches tried to explain how remittance affects growth by way of another channel whether it is current account or financial development. Adams and Page (2003) in a published paper by the World Bank, International Migration, Remittances, and Poverty in Developing Countries, tried to explain the impact of international migration and remittances on poverty using a 74 country cross section dataset. The authors looked at how GDP per capita, survey mean income per capita, Gini coefficient, remittance as a share of country GDP affects poverty. Poverty was measured as a headcount of people living less than 1USD per day, the poverty gap and the poverty gap squared. In each of these regressions, the remittance variable came up to be statistically significant. The authors found that remittances had a statistically significant impact on reducing poverty- “10% increase in share of remittance in a country GDP will lead to a 1.6% decline in the share of people living on less than 1USD per person per day”. We wish to explore this topic in a different light by looking at it in a different perspective. Mundaca’s study of remittances looked at it in a perspective wherein a developed financial market can lead to more remittances. The paper by Aggarwal, Kunt and Peria provides a model that will allow us to measure the effect of remittances in the financial sector. Remittances can contribute to the improvement of financial sectors such as the banking industry through the increase of deposits or the amount of credit provided to private sectors (Aggarwal, Demirguc-Kunt, Peria, 2006). The increase in credit

provided can be attributed to the increase in loanable funds which can be increased through remittance deposits. Currently, remittance income is mainly utilized to buy necessary and discretionary items, pay off debts, purchase of non-productive assets such as land and homes. In the paper by Charles Stahl, he notes that the propensity to save is higher for remittance receiving households than that of non-receiving households – this increase of savings rate can have a long run positive effect on the economy if we follow the Solow Model. As you can see an increase in the savings rate does not only promote financial development but can also promote growth. The paper by King and Levine further discusses how financial development is correlated to growth by way of several different types of financial indicators. Aggarwal, Demirguc-Kunt, and Peria uses two of the 4 different financial indicators suggested by King and Levine to create a financial development model that will help us map the relationship between remittances and financial development and in turn affecting growth as suggested by the paper from King and Levine. Furthermore, the paper by Aggarwal uses balance of payments data on remittances of 99 countries that were categorized as low to middle developing economics from 1975-2003. The primary sources of data for these remittances were obtained from the International Monetary Fund (IMF) Balance of Payments Statistics Yearbook.

Data In our paper we will be working with the same data but decided to start and end on a later period from 1980-2004. Also we will not conduct our study on the 99 countries but on the top ten countries and also thirty one other countries that are similar to the top ten remittance receiving countries that are currently receiving the highest inflow of

remittances. These forty-one countries will be chosen with same criteria as the selection criteria in the Aggarwal, Demirguc-Kunt, and Peria paper- all countries are categorized as low to middle developing economies. However, not all of the countries that are chosen in our paper are highly dependent on remittances. We would like to see if there were any biases during the selection of countries in the Aggarwal, Demirguc-Kunt, and Peria paper since most of the 99 countries in their model are dependent on remittances. If there are biases, then the current model only applies to countries that are dependent on remittances and nothing else. Although data is readily available and easy to attain, we are aware of the fact that data on remittances are not as accurate because of there are migrant workers who remit there money through unofficial sectors. These unofficial sectors do not provide records of remittances remitted by migrant workers and because of that there is no way to measure it. These remittances are left unaccounted for and many believe that the range for unaccounted or unrecorded remittances range from 20 to 200 percent (Aggarwal, Demirguc-Kunt, Peria, 2006). Data used in our model is obtained from the International Monetary Fund (IMF), United Nations Conference on Trade and Development (UNCTAD) and the World Bank (WB) websites and statistical yearbooks. The dataset includes countries from South America, Central America, the Middle East, Africa, Southeast Asia, Central Asia, and Europe.

Methodology

We will use balance of payment data on remittance flows from 31 countries categorized as low to middle developing economies over the period of 1981-2005 to study the impact of remittances on financial development. The countries of interest are the top ten remittance receiving countries are the following: India, Mexico, Egypt, Philippines, Turkey, Morocco, Poland, Jordan, Bangladesh, and Brazil. The 31 other countries have similar economic backgrounds as the top ten remittance receiving countries- all categorized as low to middle developing economies by the World Bank. Increases in the level of deposits to banks will be used to measure financial development in our panel dataset because when big projects are implemented, individuals will need loans which will be finance by banks; hence, if there are increases in the level of deposits to banks, more loans are available for usage. King and Levine suggested that the level of demand deposits is a good indicator for financial development since there is a relationship between this financial indicator and economic growth.

To examine the relationship between financial

development and remittances by running regressions on the following model: FDi,t =β0 + β1 Rem/GDP + β2 Ln(GDP) + β3 GDP/Cap + β4 Inflation + β5Flows/GDP + β6 Exports/GDP + β7 Interest Rate + μ The difference between our model and the Aggarwal, Demirguc-Kunt, and Peria is that we included an interest rate variable because we believe that changes in interest rates will affect financial development. A standard measure of financial development, FD, according to the literature by King and Levine is the ratio of bank deposit to GDP. Data that are used to construct this ratio are obtained on the IMF statistics website. Rem/GDP is the ratio of remittances to GDP. Remittance data is obtained from the IMF World Economic Outlook and the UN Statistics Handbook under workers’ remittances which are

the current transfers made by migrant workers working abroad. If our theory that remittances does have a positive impact on financial development, we will observe that Rem/GDP will be a statistically significant positive value. Log of GDP and GDP per capita are included in this model because they allow us to take into account for the country size and the level of economic development. Data for the Log of GDP and GDP per capita are obtained on the UN statistics division website. We can expect that these two variables will have a positive impact on financial development: an increase in these two variables will cause an increase in financial development. To account for inflation, we will look at the GDP deflator during this time period. GDP deflator data was obtained from the UN Statistics Handbook. According to Smith, Levine, and Boyd, 2001, inflation affects individuals’ decision-making choices that also influences savings rate in real assets. We would expect that this variable will have a negative affect on financial development. Another variable that influences financial development is capital inflows to GDP, Flows/GDP. The ratio of Flows to GDP was constructed by data that was obtained from the UN Statistics Handbook. We expect Flows/GDP to have a positive effect on financial development. The level of export to GDP is also important to helping us understand financial development. The ratio of export to GDP was constructed by data that was obtained on UN statistic division website under trade and development. We expect this variable to have a positive affect on financial development because as exports increase, firms will have an incentive to expand due to the raising demand, thus helping financial development. We would also include interest rate as the final independent variable in our model because interest rate does effect individuals’ decision on saving. Interest rate data was obtained on the IMF statistics website. Since there were missing

data for the lending rate and discount rate, we made an interest rate variable as the average percent change in the discount rate and lending rate. We expect this value to be positive also. A correlation matrix (on appendix, Correlation Matrix) for the variables showed that among the seven independent variables, ln_gdp is positively correlated to per capita gdp by .3082 and ln_gdp is negatively correlated to rem_gdp by -.4572 and to exp_gdp by -.2801. Given this, ln_gdp seemed to be the variable that is highly correlated with rem_gdp, per capita gdp and exp_gdp if compared among the other explanatory variables in the model. But these values are not strong enough evidence to state that multicollinearity can be a problem in this model. Also, we can not simply drop ln_gdp in our model because we believe that this variable is relevant, has explanatory power and must be included in the model. Omitting an important variable can result to specification bias which is a more serious problem than multicollinearity. Our model has data from 41 countries over 20 years so there may be some characteristics in each individual country that persist over time which are unobservable in our model. For that reason, we will need to run a fixed effect or random effect regression to address this underpinning issue, time invariant errors, with our data. We will need to perform a Hausman Test to see which regression, fixed or random, best fits our dataset. When running the Hausman Test, the Null Hypothesis is that the Fixed Effects and Random Effects Results are equal and the alternative hypothesis is that the fixed effects and the random effect results are not similar. We would first run this regression but without the remittance to GDP variable:

FDi,t =β0 + β1 Rem/GDP + β2 Ln(GDP) + β3 GDP/Cap + β4 Inflation + β5Flows/GDP + β6 Exports/GDP + β7 Interest Rate + μ We should expect the signs of this model to be similar to that of the Aggarwal, DemirgucKunt, and Peria model. And then we will add the remittance to GDP variable and observe the significance of remittances to financial development. Aggarwal, DemirgucKunt, and Peria suggested that remittance is statistically significant in their 99 country panel dataset so if their model is good, we should see similar results in our regression. We will then add a remittance lagged variables (one year and two year lag) into our model to see if that would yield more robust results with the lagged terms. Adding lagged variables in our model does make economic sense because it usually takes several quarters to see the affects of remittances on financial development. After that is done, we will run both the random and fixed effects regressions on our model above. In addition to the regressions we have mentioned, we will further more split our dataset into two subsamples: one sample including 21 countries that are highly dependent on remittances and the other sample including 20 countries that are not so dependent on remittances. The 41 countries were ranked according to the amount of remittances it receives and from there the dataset was divided into two subsamples. The top 21 countries were included in the first subsample of highly dependent countries and the bottom 20 countries in the subsample moderately dependent countries. If the Aggarwal, Demirguc-Kunt, and Peria model is correct, this means that the model will work for these two subsamples and should yield similar results. However if it is wrong, then we should observe deviations of their interpretation of remittances on financial development.

Empirical Results The appendix pages in back provide all of the regressions that were used in this paper. After running both the Fixed Effects and the Random Effects Regression, I ran the Hausman test to see which model was more appropriate for the dataset.

The null

hypothesis of the Hausman Test states the following: H0=βFE= βRE which means that the difference in coefficients of the Fixed Effects and the Random Effects is not systematic. The Hausman Test rules out that the Fixed Effects model was more appropriate for the dataset. Table 1 reports the fixed effects estimates of the model with and without the variable ratio of remittances to GDP. Comparing the two models, we can see that remittance have a positive coefficient and is also statistically significant in the 1% level of confidence. Furthermore, the R-squared between the two models are quite differentthe model without remittance has an R-Squared of 4.1% whereas the model with remittance has an R-Squared of 12.5%. This means that the model with remittances helps explain the variation of financial development better than the model without the ratio of remittances to GDP. Also, by looking at the model with the ratio of remittance to GDP, we see that a one percentage point increase in the share of remittances to GDP leads to a 0.132 increase in the ratio of deposits to GDP. An F-test was conducted on these models to further validate the importance of the inclusion of the remittance variable in the model. The restricted model is the model without the remittance to GDP variable and the unrestricted being the model that includes this variable. If we accept the null hypothesis this indicates that the restricted model is as good as the unrestricted in explaining the dependent variable. The F-value turned out to be 24.31 is clearly greater than the F

critical value of 3.86. Hence, we can reject the null and this means that the unrestricted model is a better model to use. As we can observe from results of the fixed effects regression on Table 1, financial development is affected by the country’s size, the level of income in the country, the size of capital inflows, and the ratio of exports to GDP positively and also statistically significant ranging from the 10% level of confidence to the 1% level of confidence. Financial development is negatively affected by inflation as we have expected due to the fact that inflation affects individuals’ decision-making choices which also influences savings rate in real assets. Looking more closely at Table 1 we see that the changes in interest rate, though positive, appear to have no significant effect on financial development. The Random Effects estimates yielded similar results to the Fixed Effects results but since the Hausman test ruled that the Fixed Effects model was more appropriate, we will only consider the results of the Fixed Effects estimate for later regression. The results of Table 1 are consistent with Aggarwal’s study. Fixed Effects regressions of the forty-one countries and also the two subsamples, highly dependent and moderately dependent, are located on Table 2. When comparing the whole sample regression to the highly dependent sample, we can see that the highly dependent countries’ fixed effects results are fairly similar to that of the whole sample. We observe that the ratio of remittances to GDP is positive and statistically significant in the 1% level of confidence. However, the coefficient of the ratio of remittances to GDP increased from 0.132 to 0.155 which suggests that a one percentage point increase in the ratio of remittance to GDP leads to a 0.155 increase in the ratio of deposits to GDP.

As expected, the signs of most coefficients of the independent variables explaining financial development did not change for countries that are highly dependent on remittances. The variable ln(GDP), which helps measure the country’s size, is still positive in both models but has dropped in significance from 10% level of confidence to the 20% level of significance. Per capita GDP is still positive and has increased in significance level to the 20% level of confidence in the highly dependent countries’ fixed effects results. Inflation variable is negative and dropped its level of significance which means that inflation does not have a statistically significant effect on financial development in highly dependent countries. The ratio of flows to GDP remains positive and statistically significant in the 1% level of confidence. The coefficient for this variable increased from 0.289 to 0.444- this suggests that for countries that are highly dependent on remittances, FDI accounts for more of the change in financial development than the ratio of remittances to GDP. The same changes are also true to the ratio of exports to GDP. We observe the same behavior, positive effect on financial development and also statistically significant in the 1% level of confidence. As for interest rate, we noticed that interest rate positively affected the whole sample and now negatively effects countries that are highly dependent on remittances- though the sign of the coefficient has changed, this variable remains insignificant in both cases. The R-squared for the model increased from 0.125 to 0.155 which indicates that the specified model works better for countries that are highly dependent on remittances. When comparing the whole sample results to countries that are moderately dependent on remittances results, we observe that the ratio of remittances to GDP is negative and also insignificant.

This means that for countries that are moderately

dependent on remittances, remittances is not a statistically important variable that helps explain their financial development and also negatively effects their financial development. This may be caused by how remittances are used in these countries that are moderately dependent on remittances. Remittances that were bought into their country through financial intermediaries are often withdrawn out immediately for consumption usage which does not help financial development. A reason why remittances are insignificant may be also due to the fact that with these moderately dependent countries, the amount money that is being remitted is often times sporadic and is not a steady source of income of families. The results of the other variables are similar to that of the whole sample results with ln(GDP), GDP per capita, ratio of flows to GDP, ratio of exports to GDP, and interest rate being positive and the variable inflation being negative. We must note that ln(GDP) and the ratio of exports to GDP dropped in significance level. The only two variables that are significant in explaining financial development for countries that are moderately dependent on remittances are inflation and ratio of flows to GDP. The R-squared for the model increased from 0.125 to 0.168 when we split the data into two subsamples. Table 3 shows the fixed effects results with one time lag.

The results of

comparing countries that are highly dependent on remittances to the whole sample are similar to that of the results with no time lags. We observe that the ratio of remittances to GDP is positive and statistically significant in the 1% level of confidence- also with an increase in the coefficient of the ratio of remittances to GDP increased from 0.39 to 0.49 which suggests that a one percentage point increase in the ratio of remittance to GDP leads to a 0.49 increase in the ratio of deposits to GDP. We observe that one time period

lagged variable of remittances is also significant but negatively impacts financial development. The significance of this variable in the whole sample is at the 1% level of confidence whereas it is in the 5% level of confidence for countries that are highly dependent on remittances. The net effect of remittances, ratio of remittances to GDP and the lagged variable, is positive (0.49-0.34=.15). This means that for countries that are highly dependent on remittances, remittances positively affects financial development. The signs of most coefficients of the independent variables explaining financial development did not change for countries that are highly dependent on remittancesyielded similar results as the model without the lagged variable. The variable ln(GDP), which helps measure the country’s size, is still positive in both models but has dropped in significance from 10% level of confidence to the 20% level of significance. Per capita GDP is still positive and became insignificant. Inflation variable is negative and dropped its level of significance. The ratio of flows to GDP remains positive and statistically significant in the 5% level of confidence. The coefficient for this variable increased from 0.246 to 0.396. For the ratio of exports to GDP, we observe a positive effect on financial development and also statistically significant in the 1% level of confidence. As for interest rate, we noticed that interest rate positively affected the whole sample and now negatively effects countries that are highly dependent on remittances- though the sign of the coefficient has changed, this variable remains insignificant in both cases. The results when comparing countries that are moderately dependent on remittances and the whole sample with one time lag are fairly similar to the results when comparing countries that are highly dependent on remittances to the whole sample with one time lag. We notice that the ratio of remittances to GDP is positive but insignificant.

We observe that one time period lagged variable of remittances is also insignificant and negatively impacts financial development.

The net effect of remittances, ratio of

remittances to GDP and the lagged variable, is negative like that of the model without the time lag (0.081-0.097=-0.016).

This means that for countries that are moderately

dependent on remittances, remittances negatively affects financial development even though this results is insignificant. We also tried running the same model with two time lags on remittances with the idea that remittances takes time to affect a countries’ financial development and yielded similar results as to the model with just one period time lag- Table 4.

The main

difference between the one period time lag and the two period time lag models is that the results for the time lags are insignificant for the whole sample, countries that are highly dependent on remittances, and also countries that are moderately dependent on remittances. Besides that main difference, Table 4 had similar results as Table 3 with the one period time lag. In addition to splitting the data into two subsamples, I ran the Hausman Test to see if the variables in both models were significantly different. The null hypothesis of the Hausman Test is as follows: βHighly Dependent=βModerately Dependent- the differences in coefficients are NOT systematic. After running the Hausman test, we reject the null hypothesis that the differences in coefficients are not systematic in the 1% level of confidence- so the differences in coefficients are systematic. The result of the Hausman Test implies that the data should be split into two subsamples. By running all of these models, we notice the underlying theme which carries on throughout all of these models- remittances are positive and significant for countries that

are highly dependent on remittances. As for countries that are moderately dependent on remittances, we observe that remittances are not significant and negative for financial development. This insignificant should not overshadow the fact that remittances do help families and therefore is not necessarily bad.

Conclusion The steady flow of remittances in many low middle income developing countries have paved way for studies of its effect on growth, poverty alleviation and eventually financial development. The main question we posed earlier was do remittances lead to financial development. After further study, research and test conducted on this model it has generated some interesting results. Based on the results of this study, remittances do have a significant and positive impact on financial development. But another interesting result came about upon splitting the data into two subsamples: one sample including 21 countries that are highly dependent on remittances and the other sample including 20 countries that are moderately dependent on remittances, results show that there are indeed differences in the effect of remittances on financial development. For countries that are highly dependent on remittances the significance and positive impact of remittances on financial development is still evident. However for countries that are moderately dependent on remittances, such an impact does not exist. Remittance is not significant meaning that it does not contribute to financial development. With these results, our model seemed to work not for all countries but only for countries that are highly dependent on remittances.

Furthermore, when adding the lag term to see if past period of remittance behavior effects financial development, we observed that the one period time lag value is negative and significant which may be explained through past studies on the usage of remittance income.

Though remittances do provide a steady source of income for

developing economies, the remittance income is used to purchase household/daily goods for living. Through this research, we can see that remittances do effect financial development positively and significantly.

And since remittances effects financial

development, remittance should effect growth due to the fact that the financial development is similar to that of King and Levine’s model and King and Levine showed the positive relationship between financial development and growth. There are some policy implications based on this study. Since there are evidence that remittances can contribute to financial development for countries that are highly dependent on it, policymakers can propose or implement policies that can improve remittance collection through formal channels. Policies that can help improve remittance collection may further contribute to financial development.

Further Research Based on several studies conducted on the use of remittances, remittances are used for consumption purposes by the families of the migrant workers. As such, we believe that a study should be done in a micro level since remittances are used and is more effective in the micro level rather than the macro level. Also, remittances have different effects for each country. For countries that are highly dependent on remittances

it may be vital to their financial development but such an effect may not be evident for countries that are moderately dependent on it. Hence, a study on remittances on a per country basis on a regional or municipality level may further capture the effect of remittances on development. Data for such study may be unavailable since this field has not been thoroughly explored.