Remittances and financial inclusion in development
The effect of remittance inflows on access to and use of financial services in developing countries
Helen S. Toxopeus
February 2006
Supervisor: Prof. Dr. B.W. Lensink Co-assessor: Prof. Dr. H.H. van Ark International Economics and Business Faculty of Economics
University of Groningen, the Netherlands
Research conducted at the United Nations Department for Economic and Social Affairs, New York
PREFACE
I would like to thank Bart van Ark, Henk-Jan Brinkman and Barry Herman for making it possible to spend five months at the United Nations Department for Economic and Social Affairs in New York for this research paper. I thoroughly enjoyed being part of the ‘Blue Book’ team, which exposed me to many knowledgeable people in the area of economics. It brought me to my research topic and provided me with an inspiring surrounding for conducting research. I would like to thank Peter Kooi for his advice regarding thesis-writing, which I have kept in mind throughout. Finally I want to thank Robert Lensink for his enthusiasm, quick feedback and useful comments throughout the supervision process.
ABSTRACT
This paper investigates the effect of remittance inflows on financial inclusion in developing countries.
Building on literature identifying the effect of financial outreach on development, I look at whether remittances have a development impact by increasing access to and use of formal financial services in developing countries.
I first present remittance and financial inclusion trends, focusing on developing countries. I build a framework using asymmetric information and demand, supply, policy and regulation arguments to explain a causal effect of remittance inflows on financial inclusion. This paper is the first to give cross- country evidence of a relationship between remittance inflows and financial inclusion in developing countries, and some evidence of causality from remittances to financial inclusion. There is no evidence that this relationship does not exist for developed countries.
CONTENT
Section 1: Introduction 5
Section 2: Remittance and Financial Inclusion Trends 8
A. Remittance Trends 8
B. Financial Inclusion Trends 13
C. Summary 17
Section 3: The Effect of Remittances on Financial Inclusion 18
A. Demand Factors 18
B. Supply Factors 19
C. The Access Frontier 21
D. Policy and Regulatory Issues 22
E. Preliminary Evidence for a Causal Relationship 23
F. Reverse Causality 26
G. Other Factors affecting Financial Inclusion 26
H. Summary and Hypotheses 27
Section 4: Research Design 28
A. Sampling Design 28
B. Data Collection 29
C. Summary 32
Section 5: Data Analysis 33
A. Descriptive Analysis 33
B. Partial Correlation Coefficient Analysis 34
C. Multiple Regression Analysis 35
D. Summary 41
Section 6: Conclusions, Limitations and Recommendations 42
A. Conclusions 42
B. Limitations 44
C. Recommendations 45
Bibliography 47
Further Reading 52
Appendices 54
SECTION 1: INTRODUCTION
Each year, millions of migrants send money earned abroad back to their country of origin. They participate in globalisation by applying arbitrage in international labour markets, creating family bonds and obligations across countries. The development impact of migration and the ensuing international remittance flow has increasingly become subject of research and policy discussions, as the vast scale of international ‘people and money flows’ has become apparent. It is no longer uncommon for remittance inflows to constitute between 5-10% of total GDP in (small) developing countries (World Bank, 2005b).
Remittance inflows surpass official development flows in middle-income countries, and foreign direct investment in low-income countries. For 2005, the World Bank estimates total flows of US$ 250 billion (including informal flows).
This trend is unlikely to reverse in the medium to long term. Migration is expected to continue and costs of remitting are falling, providing a lower threshold for migration. Part of the surge in remittance numbers is caused by the tightened regulation on money flows, pushing remittances from unrecorded, informal channels to the recorded formal system. Another factor influencing remittances numbers is improved data recording by central banks, led by increased recognition of the importance of these flows. These two factors are causing a temporary fast increase in remittance flows that is unlikely to continue at this pace. However, the falling costs and high levels of migration indicate that in the medium term, remittance flows are expected to keep growing at a 7-8 percent annual rate, similar to growth rates in the 1990s (World Bank 2005c:92-3).
The effects of this large scale movement of capital are multiple, both positive and negative. On the positive side, remittance inflows increase capital availability for consumption in receiving countries, and can create multiplier effects in local economies on GDP, job creation, consumption, income and investment (Stahl and Arnold, 1986, in World Bank 2003b, and De Vasconcelos, 2005). Remittances also supply foreign exchange, complementing national savings levels and providing finance for investment, notably for small-scale projects, hence providing for output growth (Solimano, 2003).
Bugamelli and Paternò (2005) show that a large flow of remittances into a country can help reduce the probability of current account reversals, and thus lower the chance of a financial crisis. Furthermore, remittances are a person-to-person flow of money without government intervention, often delivered directly to the lower income segment of a country. It can therefore stimulate development without increasing debt or administrative burden.
On the downside, the phenomenon of economic migration and remittance sending indicates the failure of migrants’ home countries to provide sufficient job opportunities in order to sustain the livelihood of its people in a globally competitive way. A large inflow of remittances may lead to currency
appreciation, thereby lowering competitiveness of export products (World Bank 2005c:104). One IMF study argues that work efforts of remittance recipients may decrease, thereby dampening growth (Chami, Fullenkamp and Jahjah, 2005). Social costs are considerably large, both for the family that now lives at a distance from the father or mother, and for the migrant who lives abroad without his or her previous social surrounding and possibly at a low standard of living to save money to send home.
However, in economic terms, migration is generally beneficial for migrant families and sending and receiving countries (for a full discussion, see World Bank 2005c:99-105).
One specific effect of remittances inflows will be the main question of this research paper: What is the effect of remittances on access to and use of financial services (financial inclusion) in developing, remittance-receiving countries? In many developing countries, poverty comes hand in hand with low banking outreach. Developing country financial access rates are often comparable to financial exclusion rates in developed countries (Peachey and Roe, 2004:4). The fields of remittances and financial inclusion have both aroused interest of academics and policymakers. Though speculation and country- level studies about the impact of remittance inflows on access to and use of financial services do exist, this paper will be the first to demonstrate the effect of remittance inflows on financial access and usage at a cross-country, worldwide level. This insight can lead to significant policy recommendations about dealing with remittance inflows and on how to use this to affect access to and usage of financial services. I use recent, new data on financial use and access from a substantial number of developing countries (Beck, Demirguc-Kunt and Martinez Peria, 2005), providing a comparable insight into banking outreach across a large set of countries for the first time.
The relevance of the effect of remittances on financial inclusion fundamentally lies in the impact of financial inclusion for development, poverty alleviation and economic growth. Levine (2003:5) organises the functions of the financial system to society in five categories:
Producing information ex ante about possible investments and allocate capital
Monitoring investments and exerting corporate governance after providing finance
Facilitating the trading, diversification, and management of risk
Mobilizing and pooling savings
Easing the exchange of goods and services.
Not everyone underscores the role of financial systems for economic growth and development.
Schumpeter (1911:74) does emphasize the role of the banking system:
“The banker, therefore, is not so much primarily a middleman in the commodity ‘purchasing power’ as a producer of this commodity. [ ] He stands between those who wish to form new combinations and the possessors of productive means. He is essentially a phenomenon of development [ ]”.
More recently, studies have shown empirically that there is a significant correlation of financial access with economic growth. King and Levine (1993) were the first to show empirically that financial systems have some predictive power for growth. Loayza, Levine and Beck (1999) show that financial intermediary development exerts a causal effect on economic growth. Levine (2003) provides an overview of all theoretical and empirical finance and growth literature.
Looking more specifically at the importance of financial system outreach, Beck, Demirguc-Kunt and Martinez Peria (2005) bring across three main arguments. The first builds on the argument above, noting that a well-developed financial system benefits the poor in particular, who are most affected by market imperfections due to lack of credit histories, collateral and connections. Lack of access for the poor and small entrepreneurs means inefficient resource allocation, and leads to forgoing of high return investments. Secondly, by providing access to external capital, new firms are given the opportunity to enter the market based on their potential for success, not just based on internal capital availability or connections. The final argument is taken from Peachey and Roe (2004:9), who relate financial exclusion in developing countries to the broader topic of addressing poverty issues; they note its equivalence to basic needs such as clean water and education.
Building on the knowledge that improving financial access and usage (financial inclusion) provides a stimulus to growth and development, as argued above, I focus on the impact of remittance inflows on financial inclusion (see Figure A1 in appendix A). This research paper should improve the understanding of a specific impact of remittance flows, and may trigger governments to improve remittance-related policy and regulation, develop better measurement tools for monitoring remittances and financial inclusion, as well as induce further research on this topic. Specifically, I will investigate the role that reliable remittance inflows can play in credit markets by serving as a substitute for collateral and credit history. The increased awareness of the relevance of remittance flows should also trigger caution in preserving financial flows of migrants in the face of increased monitoring of international transfers due to the crackdown on terrorist financing. If a positive development impact of remittances through financial inclusion exists, it should be maximised.
This paper will continue as follows. In the next section I will describe trends in remittance flows and in access to and use of formal financial services. I will also investigate why access to finance is so low in many developing countries. I then present arguments for a causal effect of remittance inflow on financial inclusion, followed by preliminary evidence of this relationship. I mention other factors that affect financial inclusion, and factors affecting the relationship between remittances and financial inclusion. In section 4 I will discuss the sampling design and the research design of my methodology.
This is followed by the data analysis. The paper concludes with outcomes and recommendations.
SECTION 2: REMITTANCE AND FINANCIAL INCLUSION TRENDS
A. Remittance trends
Size of flows. An estimated 3% of the world population (175 million people) is currently a migrant (UN census reports, in World Bank 2005c). In line with an increase in migration over the past decades, (international) remittances have also increased. The large differences in economic returns to labour between developed and developing countries are a clear economic incentive to undertake ‘arbitrage’ by moving to where wages are higher. Remittances increased from US$ 15 billion in 1980 to US$ 80 billion in 2002 (Solimano, 2003:4); the World Bank gives an officially recorded flow of US$ 167 billion for 2005, though estimates including informal channels approximate an additional 50% of flows that go unrecorded, around US$250 billion in total (World Bank 2005c). The exact size of the flows is unknown, in spite of attempts to incorporate informal flows into the numbers (see Freund and Spatafora, 2005, and World Bank 2005c:108). Figure 2.1 below shows that remittance flows to developing countries surpass the inflow of Official Development Assistance (ODA), only second in magnitude to Foreign Direct Investment (FDI) inflows at the aggregate level (World Development Indicators, 2005).
For some individual countries, in particular low-income and small economies, remittances are the single largest source of foreign exchange, contributing up to 31.3%1 of total GDP (World Bank 2005c).
Figure 2.1: FDI, ODA and Remittance inflows into Low and Middle Income countries
0 2E+10 4E+10 6E+10 8E+10 1E+11 1E+11 1E+11 2E+11 2E+11 2E+11
1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003
Year
Current US $
Foreign direct investment, net inflows (BoP, current US$)
Official development as sis tance and official aid (current US$)
Workers' remittances and compens ation of em ployees, received (US$)
Source: World Bank 2005b
1 For Tonga.
Additional graphs in appendix B (figures B1) give a more detailed description of the size and importance of remittance inflows in different regions and levels of development world wide2. To put the size of remittance flows into perspective, I compare them with Foreign Direct Investment (FDI) and Official Development Assistance (ODA). I will mention the important trends. Tables 2.2 and 2.3 below compare flows in the year 2003.
Table 2.2: Showing capital flows to developing countries categorized by income level, 2003
2003 capital flow s (US$) low incom e low er m iddle incom e upper m iddle incom e
Foreign Direct Investm ent 16.128.400.000 99.551.800.000 36.096.300.000
Official Developm ent Assistance 32.128.290.000 21.782.960.000 3.779.100.000
Rem ittance inflow s 36.741.300.224 51.519.000.576 27.270.000.640
Source: World Bank 2005a
Table 2.3: Showing size of capital flows to different regions of the world, 2003
2003 capital flows (US$) East Asia and Pacific Europe and Central Asia Latin America and Caribbean
Foreign Direct Investment 59.611.900.000 35.614.200.000 36.532.900.000
Official Development Assistance 7.131.210.000 10.464.520.000 6.150.720.000
Remittance inflows 19.707.000.832 12.880.000.000 34.389.999.616
Middle East and North Africa South Asia Sub-Saharan Africa
Foreign Direct Investment 4.755.500.000 5.162.700.000 10.099.300.000
Official Development Assistance 7.629.430.000 6.170.570.000 24.145.910.000
Remittance inflows 16.083.000.320 26.787.000.320 6.013.300.224
Source: World Bank 2005b
Comparing international capital flows at a world level and in low and middle income countries
At a world level, FDI flows dwarf both ODA and worker’s remittances. FDI takes off at an exponential pace after 1993, to fall again quite dramatically after 2001, though remaining larger than ODA and remittances. At a world level, remittances are larger than ODA. We see that ODA has stabilised since 1991 while remittances kept growing, nearing US$180 billion in 2003.
In middle and low income countries, remittances play a relatively larger role compared to FDI than at world level. In developing countries as a whole, remittance inflows are double the size of ODA.
Remittance inflows are the largest foreign capital flow when combining all low income countries.
They overtake FDI in 1999 and increase at a slightly faster pace to more than US$36 billion in 2003, in absolute figures a larger amount than remittance inflows to upper middle income countries in that year. The inflow to lower middle income countries is larger yet, over US$51 billion.
Remittances constitute the largest percentage of Gross National Income (GNI) in low income countries, compared to upper middle income and lower middle income countries. In absolute figures per capita this order reverses, with upper middle income countries receiving a much larger absolute amount ($73) in 2003 than low income ($15) and lower middle income countries ($21). This shows that although the absolute value of average remittance receipts per person is largest in upper middle
2 Data is taken from Global Development Finance 2005 and World Development Indicators 2005 (World Bank
income countries, the relevance of remittances compared to total income in the country is largest in low income countries.
Comparing international capital flows at a regional level
The highest absolute amount of remittances goes to Latin America and the Caribbean, followed by South Asia. The lowest amount of inflow is in Sub-Saharan Africa.
As a percentage of GNI, we see the largest (rise in) remittance inflow in South Asia and Latin America and the Caribbean. Remittance inflows / GNI are large in the Middle East and North Africa too, with a peak in 1993 of 3,7% of GNI, but a decline after that. Remittances to Europe and Central Asia and the East Asia and Pacific region are the lowest.
The Middle East and North Africa and Latin America and the Caribbean have the largest remittance inflows per capita, $54 and $65 per capita in 2003 respectively. Sub-Saharan Africa and South Asia receive the least at $9 and $19 per capita, respectively.
In South Asia and the Middle East and North Africa, remittance inflows are the largest capital inflow compared to FDI and ODA. Starting around 1990, measured remittances in the Middle East and North Africa increased and remained high. In South Asia, remittances increase fast, in contrast with FDI and ODA.
Where FDI dropped sharply after 1999 in Latin America, remittances increased consistently reaching almost the same level as FDI in 2003, expecting to overtake FDI in magnitude after 2003.
In East Asia and the Pacific region FDI plays an overwhelmingly large role. The financial crisis that started in 1997 affected not only FDI, but also remittance inflows to a certain degree. Both dropped around the same time, and recovered again.
In Europe and Central Asia remittance inflows are as large as ODA. After 1994, FDI increased at a rapid pace, with remittance inflow to increase only steadily. Remittance inflows in 2003 were measured to be only one third of FDI.
In Sub-Saharan Africa remittance inflows are subordinate to ODA and FDI. It seems to be a stable flow, but this could also be due to roughness of estimates. This is the only region where the largest flow is ODA.
The accuracy of the data presented above may vary between regions. There is considerable variation in how remittances are transferred, through formal or informal channels, which affects whether flows are recorded or not (for an overview of players in the remittance market, see Orozco, 2004). The Global Economic Prospects report 2006 (World Bank, 2005c) provides an overview of different channels used for remittances at country-level. This varies from 96% using formal channels in the Dominican Republic to just 20% in Uganda (World Bank 2005c:91, based on World Bank household surveys). In
2005a/b). GDF looks only at developing (low and middle income) countries.
Sub-Saharan Africa, for example, use of informal channels is more widespread, and therefore not well recorded (Sanders 2003:3-4). This large variation in channels used is a disadvantage of the current state of data on remittances, and should be kept in mind.
Direction of the flows. By matching migration patterns worldwide, Harrison, Britton and Swanson (2003) estimates sizes of remittance flows per continent and for selected countries (origin and destination) for the year 20003. Table 2.4 below shows the aggregate per continent. Remittances flows from North America to Latin America and the Caribbean are large at US$14,2 billion, but remittance flows between Asian countries are twice that size, displaying a large South-South remittance flow in Asia (within-Asia flows are still $29,3 billion when excluding Japan). Remittance flows within Africa are estimated to be larger that the flows from both Europe and Asia. Other large flows are those from North America to Asia and flows within Europe.
Table 2.4 Table showing size of remittance flows between continents (2000)
(Billion $) Africa Asia Europe LAC
North
America Oceania Total*
Remittances coming from
Africa 3,7 0,5 0,1 0 0 0 4,2
Asia 3,4 31,5 3,4 0,5 0,2 0 39
Europe 2,6 3,2 9,5** 0,4 0,4 0,1 16,2
Latin America / Caribbean 0,1 0,6 1,1 0,1 1,8
North America 0,7 7,9 5,7 14,2 0,9 0,1 29,6
Oceania 0 0,2 0,4 0 0,1 0,8
Total* 10,4 43,4 19,6 16,2 1,6 0,3 91,5
Source: Harrison et al., 2003
* Totals may differ slightly to ow n calculations due to rounding
**$ 24,1 billion for European border w orkers are excluded Remittances going to
When we look more closely within regions, flows are not evenly divided between countries. Table 2.5 gives more detail by listing all remittance flows larger than US $ 300 million in 2000 (Harrison et al, 2003). The flow from the US to Mexico is overwhelmingly the largest, accounting for about half of all flows to LAC countries. It also shows that within-Asia remittance flows largely originate in Saudi Arabia, with Japan in second place. More than half of the Europe - Asia flows occur in the Germany – Turkey corridor. Although the largest flows are into developing countries, we also note that some developed countries receive considerable amounts of remittances (Canada, Germany, UK and Italy).
The sending countries are all developed OECD countries, with the exception of Saudi Arabia (developed, but not OECD). Despite these large flows originating in developed countries, about 30% of
3 The data by Harrison et al. (2003) does not exactly match the year 2000 data by the World Bank (2005b). Though categorized somewhat differently, LAC flows reported here are smaller, European and Asian flows are larger. Though both sets are based on IMF Balance of Payment statistics, they have been augmented with own estimates of missing or badly recorded flows. This difference is a good indication of the need for improvement of remittance data.
all remittance flows are ‘South-South’ flows (including Saudi-Arabia as a developing country would raise South-South remittances to 45%, World Bank 2005c:111).
Table 2.5 Largest remittance corridors in 2000 in decreasing order
From To Amount (million $)
United States Mexico 7612,5
Saudi Arabia India 3609,7
Saudi Arabia Pakistan 1804,9
Saudi Arabia Philippines 1582,7
Saudi Arabia Eqypt 1388,4
United States China 1350,5
Germany Turkey 1195,2
United States Philippines 1186,4
Japan Korea 1012,1
United States India 977,7
Saudi Arabia Indonesia 971,8
United States Vietnam 837,9
Saudi Arabia Bangladesh 694,2
France Portugal 659,2
United States Canada 658,2
United States Germany 634
France Morocco 600,1
United States United Kingdom 595,1
France Algeria 568,5
Japan China 534,6
Switzerland Italy 448,4
United States Italy 437,9
United States Poland 432
United States Colombia 422,3
Japan Brazil 405,3
Germany Italy 370,2
United States Russia 353,4
Source: Harrison et al, 2003
Stability of flows. Remittance flows are often claimed to be more stable than other capital inflows.
Using the WDI 2005 (World Bank, 2005b) figures for developed countries as a whole, I calculate and compare the volatility of remittance flows, ODA and FDI over the time period 1979 – 2003. I calculate volatility in two ways. One is the coefficient of variation, calculated as the standard deviation divided by the mean of a set of data (times 100). Dividing by the mean makes this measure comparable across datasets. The coefficients of variation of the three capital flows indicate that the least volatile flow is ODA, followed by remittances, with FDI as the most volatile flow (see table 2.6).
This variable does not take into account increasing trends in the data, though. Since FDI and remittances show an increasing trend during this period, I also want to measure volatility with the assumption of a linear trend for each of the three capital flows. A steady increase in both remittances and FDI are therefore not included as part of their volatility. I measure the best fit trend line for all three flows, by regressing them individually against the period measured (per year), including a constant in the
regression. I then calculate the standard deviation of the residuals of each regression as a measure of volatility (the square root of the sum of the squared residuals). This measure yields similar results as the coefficient of variation; ODA is the least volatile flow (at 3.7 E+10), followed by remittances (4.7 E+10). FDI is the most volatile compared to its linear regression line (1.4 E+11). Compared to FDI, therefore, remittances are indeed quite stable. Compared with ODA, remittances are somewhat more volatile.
Table 2.6: Measures of volatility for FDI, ODA and Remittances
FDI ODA Remittances
Coefficient of Variation 95 32 65
Standard Deviation of residuals of a
linear regression line 1,4E+11 3,7E+10 4,7E+10
Source: World Bank 2005b
Finally, though not measured here, remittance flows are said to be counter cyclical, thus smoothing consumption patterns. Migrants’ transfers are not driven by investment-like incentives, but are based on a desire to increase the welfare of relatives in their country of origin (Bugamelli and Paternò, 2005:1).
Money flows home thus tend to increase in economically bad times as a form of income support (Solimano, 2003:5). There is country-level evidence from various sources that a decrease in GDP, related to natural disasters or other economic shocks, leads to an increase in remittance inflows (World Bank, 2005c:99-100). At the same time, remittances represent a stable source of foreign exchange in case of macroeconomic shocks (World Bank, 2005c).
B. Financial Inclusion Trends
The financial landscape in developing countries. Before looking at numbers of people with access to financial services, it is useful to briefly indicate what sorts of financial service providers there are.
Although many people in developing countries do not have any relations with banks, this does not mean that they do not have any financial access. There is an array of informal and semi-formal financial services that lies outside the official, regulated sector. There is large variation in the institutions and people offering financial services, the most relevant of which I will briefly mention here. Although categorised, note that different organisational forms overlap and are sometimes combined.
In the formal financial sector, commercial banks and development / government banks play the largest role. Commercial banks in developing countries tend to concentrate on the more affluent segments in an economy. Government or development banks, including postal banks, tend to carry a mandate to serve all people, but have often failed to do so in developing countries.
Credit unions/co-operatives and microfinance institutions operate in the semi-formal sector. Though they are not usually under the same regulatory system as banks, they are often officially recognized organizations, offering financial services with a social objective. Credit co-operatives are owned by its members that each have a vote in its organization. Microfinance organizations provide small-scale financial services to those excluded from formal financial institutions. Though microfinance institutions have developed services in many different organisational forms, their commonality lies in the use of informal market lending methods to lower informational costs and to solve the problem of lack of collateral, combined with access to capital of the formal financial system. There are also commercial banks that have set up microfinance offices. For a detailed overview of informal sector services and typologies see Montiel, Agénor and Ul Haque (1993).
The emphasis in this paper lies on access to and use of formal financial services. I offer two reasons.
One is fundamental: I believe that ideally, a formal financial sector should be inclusive, offering access to all members of a population and open to regulation. However, I would have liked to include access and usage of semi-formal financial services here, such as credit unions and microfinance institutions, if it were not for the more practical issue of data availability. The comparable data available on access and use of financial services comes from the regulated, formal sector. For quantitative analysis this is therefore the most sensible choice.
Levels of access to and use of financial services. The number of people with access to financial services world-wide is difficult to quantify; just as the kind of services offered varies, the access and use of financial services vary considerably per country, also. Because of these factors, and due to the unregulated nature of large parts of the financial system, the exact size of these markets is not systematically reported. I will present the data that is available.
Access ranges from a few percent of the population in financially repressed economies, for example Nicaragua (4.7%) and Armenia (8.9%), to 99% in Denmark and 97.7% in Singapore (Beck et al., 2005:36). This data reflects share of households with a bank account; although this is a particular category of service (savings and/or transactions), it is a useful measure of access. It plays a key role in offering other financial services such as credit and insurance, and there is considerable data availability on usage of bank accounts (Porteous, 2004:24). The data for selected countries can be found below (table 2.7).
The Consultative Group to Assist the Poor (CGAP) estimates that on a global scale there are about 3 billion people of working age without access to financial services (CGAP Press, Microfinance by the Numbers, undated). They state that hundreds of millions are served by state-owned agricultural, development and postal banks; member-owned collectives; savings banks and low-capital local banks.
It is estimated that in 2003, 70 million individuals were served by microfinance institutions (Daley- Harris, 2003:3).
Table 2.7: Table showing household share with bank account, in order of increasing financial access Country Household share with bank account Country Household share with bank account
Bulgaria 0.0024 Colombia 0.412
Nicagarua 0.047 Brazil 0.427
Armenia 0.089 Italy 0.704
Kenya 0.100 Greece 0.789
Pakistan 0.122 Austria 0.814
Guyana 0.137 Spain 0.916
Ecuador 0.161 Belgium 0.927
Guatamala 0.178 France 0.963
Mexico 0.250 Denmark 0.991
Namibia 0.284
Source: Claessens (2005) and Gasparini et at (2005) in Beck et al. (2005:36)
Peachey and Roe (2004) analyse the balance between cash in circulation and deposits mobilized together with deposit-to-GDP ratios to draw conclusions about financial access at a regional level.
Lower cash-to-deposit ratios and higher deposit-to-GDP ratios indicate better financial access5. Sub- Saharan Africa lags behind other regions, which they link to the extreme poverty in this part of the world. Secondly, Asian countries are well underway in expanding financial access, especially in the industrializing countries. Central European advancement has been mixed; the advanced transition economies have shown very good progress, others have not. Finally, progress in building access in Central and South America has been slow in spite of a relatively robust economic base.
The picture for developed, urbanised social market economies in Europe (Germany, Scandinavian countries, France, The Netherlands and Spain) is quite different, with access figures above 95%. Anglo- Saxon market economies such as the UK and the US have lower access, around 91% (average for industrialized countries). Below average (70-80%) are the Southern European countries and Ireland, possibly due to the larger proportions of ‘rural dwellers’ and larger regional income inequality (Peachey and Roe, 2004:5, 13).
There are four different basic types of financial services usually distinguished: credit, saving, insurance and transactions (with international remittance sending as a subcategory). The above data related to bank accounts; the table below (table 2.8): shows data on all four retail financial services for African countries. This gives a more holistic view of the services that are accessible in a country, and also shows that there is considerable variation between services.
4This figure seems remarkably low; I expect it to be higher.
5See Peachey and Roe (2004: 35-37) for a full explanation of their methodology.
Table 2.8: Table showing use of four different financial services in five African countries Country Savings account (% of
adult population)
Access to credit / including
estimate of micro loans Transactions Insurance
Botswana 41.3 16.4 / 24.7 15.9 15.4
Lesotho 28.5 1.8 1.1 1.8
Swaziland 28.0 12.7 7.4 4.6
Namibia 44.8 11.8 /23.3 9.8 12.8
South Africa 49 57 48 42
Source: Genesis Analytics (2003a, b, c and d) and Porteous (2004)
Reasons for lack of access to and use of formal financial services. The data above leaves us with the question: Why is there so little access to formal financial services in developing countries? The answer to this question serves as an introduction to the next section that focuses on the effect of remittances on financial inclusion.
In basic economic reasoning, the missing market is surprising. Diminishing marginal returns to capital should lead to higher returns on investment from entrepreneurs with little capital than from enterprises with a lot of capital, as noted by Murdoch and Armendariz de Aghion (2005). Capital should flow from large enterprises in developed economies to small-scale entrepreneurs in countries or regions where capital is relatively scarce. Although investing in these regions may be relatively more risky, interest rates should adjust upwards accordingly. Though restrictions on interest rates are one explanation why this often does not work, there are more fundamental factors that play a role.
Financial markets should be distinguished from auction markets, where sellers choose the highest bidder. Financial markets involve delivery in the future, which is always uncertain (Fry, 1995:62). This
‘payment promise’ carries a risk that differs per potential ‘buyer’. In the case of credit products, those willing to pay the highest interest rates are not necessarily the customers that will maximise returns. The expected likelihood of buyers paying back money plays a large role in profitably allocating funds. The allocating mechanism does not just depend on prices but also on screening of buyers (Stiglitz and Greenwald, 1991). This is relevant both for credit markets and the market for savings (where the cash- holding institution is the one to be ‘screened’). The cost level of monitoring individual borrowers will affect whether or not markets will come about, and whether markets are restricted to lower the risks of lending (Fry, 1995:62). The issue of screening and monitoring leads me to the fundamental problem that seems to cause the missing financial markets: informational asymmetries (unequal access to information between suppliers and buyers), notably adverse selection and moral hazard.
Adverse selection occurs because lenders do not know how risky their customers are. They therefore cannot charge risky customers a higher interest rate, and raising interest rates for all drives low-risk customers out of the market. Unlike in classical markets, the level of interest rates does not function as a signalling mechanism. Moral hazard arises since banks cannot monitor their customers fully to assure that they are putting in a maximum effort to make their projects successful, or to make sure that they do
not run off with the borrowed funds. A weak regulatory environment, often the case in developing countries, enhances this risk (Murdoch and Armendariz de Aghion, 2005:6-8).
There are three straightforward solutions to these problems (though their practical application may not be so straightforward). Firstly, if banks could gather information about their clients more cheaply, they would have knowledge about the risk level of their clients, for example by knowing their financial history (thus solving the adverse selection problem). If banks could enforce contracts better, clients would face higher costs in breaking them, which solves part of the moral hazard problem. Thirdly, if borrowers had marketable collateral, assets would cover loans, thereby lowering risk for banks and motivating clients to pay back loans. This would therefore solve both problems of adverse selection and moral hazard. These solutions, generally, do not work for the poor: they generally do not yet have these assets, nor do they have a formal financial history, and they need loans to break out of the poverty trap.
In order for this to happen, informational problems and transaction costs need to be reduced (Murdoch and Armendariz de Aghion, 2005:6-8).
Other factors play a role, too. Corruption in many developing countries increases transaction costs.
Costs of deposits and lending rise if officials have to be bribed. Also, inadequate public services make operation of financial institutions more tedious. In areas with unreliable electricity supplies (or none at all), systems will work only sporadically or will be paper-based (‘A Survey on Microfinance’, The Economist 05/11/05). The combination of high (perceived) costs and small scale fails to attract profit seeking players to the market. Other arguments for lack of access are geographic isolation (often rural areas are less well served by banks) and social prejudice toward poor communities (Prahalad, 2004:293).
C. Summary
The data presented shows the vastness of remittance flows to developing countries, though they vary considerably between continents and countries. A few remittance corridors sometimes transfer the bulk of remittance flows in a region. Financial inclusion in developing countries is low compared to developed countries. Variation in financial usage figures is also large within the developing country set.
Informational asymmetries cause adverse selection and moral hazard problems, arguably fundamental causes of failing financial markets in developing countries. The overview in this section of remittance and financial inclusion trends leads us to the theoretical framework, which combines the two phenomenons and looks at their interaction on the basis of existing literature.
SECTION 3: THE EFFECT OF REMITTANCES ON FINANCIAL INCLUSION
The continuously growing and widespread demand for remittance services has interested suppliers in a segment of the market that was previously largely unserved. Not only suppliers, but also policy makers are looking at using these flows for improving financial access for the previously ‘unbanked’. As stated by Orozco and Fedewa (2005): “The notion that financial intermediation can lead to development is rooted in the concept that money transferred through financial institutions paves the way for senders and recipients to gain access to other financial products and services, which they might not have otherwise.”
In this section I will further explore in what channels exactly changes are taking place that may cause remittances to increase financial inclusion. I will look at demand, supply and policy factors. I will also apply the access frontier theory to financial markets in developing countries.
A. Demand factors
Remittance senders, by definition, are in demand of at least one financial service: one that provides for international payments. This demand can be an incentive to turn towards the banking sector or other financial institutions as a supplier. The demand for receiving remittances, at the other end of the transaction, may induce people to look for financial services beyond their neighbourhood for the first time. The Global Economic Prospects report 2006 (World Bank, 2005c) notes that “in contrast to cash transactions, remittances channelled through bank accounts may encourage savings and enable a better match for savings and investment in the economy”. Migration and subsequent remittance sending can thus be the first personal interaction with the global economy for many.
The migrant who sends remittances induces the receiving person to contact the institution through which the money is being sent. If this institution is a bank offering supplementary financial products (versus a money transfer organisation or informal channel that only offer remittances) this interaction with financial institutions can create demand for products such as saving, credit, mortgages and insurance. In this manner, increased financial literacy of the migrant abroad can be a driving force for increased literacy at the receiving end. Estimates show that around 10 percent of remittance receipts are being saved, invested, and used for entrepreneurial activity (Orozco and Fedewa, 2005:4). The fact that some cash inflow is invested indicates that demand among remittance receivers for complementary financial products does exist.
An opportunity on the demand side is the high transfer costs in the market. The cost of sending $200 through a money transfer organisation (i.e. Western Union and MoneyGram) is around 10%. There is demand for cheap, reliable remittance services; factors which banks can use to compete. Orozco (2004:25) finds that transfer costs are somewhat lower when sent through regulated financial institutions, such as banks and credit unions. IRnet, a credit union system initiative, has linked credit
union co-operatives with Citibank branches, both in the US and in receiving countries, at a lower cost than informal networks (Ratha, 2003:35). The challenge lies on the supply side, where banks need to target this demand.
Demand for channelling remittances. Some remittances are sent in-kind, in order to stipulate the use of the remitted ‘capital’. This implies that demand exist from the sender’s side to influence the use of their money (i.e. by sending an airline ticket or vouchers). Linking other financial products, such as different payout options or mortgages, to the remitted amount is a service sometimes already asked for by customers. Increasing the possibilities of formal money transfer services in this manner could be a response to existing demand (Sander and Maimbo 2005:68).
B. Supply factors
To adhere to the vast demand for remittance sending services, an array of remittance-sending institutions exists. Besides many informal methods of sending remittances (‘Hawala’-type institutions6) and the money transfer organisations that capture a large share of the market, other more diversified and formal financial institutions also offer remittance services, though many have been slow to catch up.
Commercial banks that become aware of the vast size of remittance flows (an annual average of $700- 1000 per sender is estimated by Harris, 2002 in Orozco, 2003), however small the individual amounts, have increasingly become interested in targeting this new market segment. Besides capturing the money flows, the remittance channel can be used to sell a package of financial services geared towards low- income individuals. Hernandez-Coss (2004) states that “by developing formal remittance channels that are competitive with informal ones, the formal financial sector has an incentive to develop and benefit from the overall opportunity to grow and expand through the remittance market.” Credit unions world- wide have also jumped on remittances and have collectively created a remittance service (IRnet) that allows their members and non-members to send money electronically. In the process, they offer users other financial services such as savings accounts (see Grace, 2005).
Building financial history and marketable collateral through remittance receipts. The perceived benefits of serving the low-income market have increased due to poorer people’s demand for remittance services and the ensuing constant money inflow. In the previous section, the reasoning for missing credit markets (and related, savings markets) was explained. Where do remittances fit into this missing financial market? It seems that they may be able to solve part of the problem. Remittance inflows can lie at the basis of two of the three solutions mentioned earlier: reducing informational problems and having marketable collateral7.
6 ‘Hawala’ is an often used term referring to various sorts of informal methods of transferring money.
7 With regard to the third solution, regulatory strength, no direct relationship can be found with remittance inflows.
Regular remittance inflows can serve to reduce informational problems by building financial history.
The constant inflow of money attracted from abroad allows the lower-income segment of the population to build a sound financial history with a financial institution. The income received, unlike prior to migration, now needs some form of intermediation in order to get the money at its destination. Banks can apply cross-selling to obtain new clients and allow them to build financial history: offer an international transfer service together with complementary services, such as a savings or checking account. Due to the remittance inflow, the bank gets some insight into the income of the client and into expected future funds. Remittance receivers can hereby inform banks of their creditworthiness, since a constant (future) inflow can repay loans. Additionally, since in adverse circumstances at home, remittances from abroad often increase8, this can potentially lowers the risk profile of the client. Banks thus obtain information about prospective clients for loans and reduce the problem of adverse selection.
The argument for remittance inflows as a marketable collateral is similar. Remittance inflows do not only have an informational function, they also carry a direct value to the bank. If remittances enter through a bank, clients can use both current and future inflows as ‘collateral’. If accepted by banks as such, loans could be (partly) covered by remittance inflows, thereby lowering risk for banks and motivating pay-back and optimal project management. This therefore could provide a solution for both adverse selection and moral hazard. Nevertheless, remittance receivers can still order the sender to stop sending via the bank, and run off with a loan; or the remittance sender may not be able to send regularly him/herself due to irregular employment abroad.
Relationship lending techniques. Relationship lending techniques are another way to solve asymmetric information problems. Relationship lending by banks involves gathering information about a firm through establishing contact with the firm, the owner and the firm’s local surrounding, and using this to make credit availability decisions and loan requirements for this particular firm (Berger and Udell, 2002). Remittances are, just like saving, a relatively risk-free way to establish contact and build a relationship with new clientele. It can use this to base its client analysis more on ‘soft’ data, such as reliability and character of the firm’s owner. The technique of relationship lending (rather than using hard data such as credit history, as noted above) is most suited to small banks, since the information gathered is restricted to the loan officer, and difficult to transmit across a multilayered organisation (Berger and Udell, 2002).
8 For evidence on counter cyclicality of remittance flows see Global Economic Prospects Report. 2006: 99-100 (World Bank 2005c)
Direct income effect. Finally, the direct income effect of remittances may affect supply. When a family member decides to migrate, he or she would, rationally speaking, only do so if the expected benefit from working abroad minus the extra costs of sustaining oneself abroad were larger than the family income before migration. The receiving families will therefore, in general, move into a higher income client group that is more attractive for banks and may therefore boost supply of financial services for this group.
C. The Access Frontier
The access frontier can be applied to approach the relationship between remittance inflows and financial inclusion from a different angle. This method combines demand and supply arguments and is used by Porteous (2004) to look at how a financial market can ‘work for the poor’. The access frontier can be defined as is the maximum usage possible under existing structural conditions of technology, infrastructure and regulation (Porteous, 2004:8). He argues that the access frontier can be moved outward until market development moves into a saturation and consolidation phase, where the market reaches a natural limit. Usage is at its maximum, and non-usage becomes a genuine choice, unhindered by income or supply constraints. Porteous investigates why the access frontier in developing countries is not yet at its maximum (natural limit).
I take the example of South Africa, cited by Porteous, to demonstrate how remittance inflows can move the access frontier outwards. In South Africa, 48% of all adults have a bank account. The two most frequently cited reasons in South Africa for not having a bank account are not having a regular income (35.6%) and not having a job (59.8%). Third largest reason is that earnings are too little to make it worthwhile (11.4%). Less people cite not having an identity document, not qualifying for an account or not wanting to keep a minimum balance or paying service fees as reasons for non-usage (between 1- 6%).
The regular inflow of remittances may move the access frontier outwards by getting rid of the reasons for non-usage. Remittance inflows can function as a substitution for a job and a regular income. The family member that migrated got a job abroad to send home earnings. Since remittances are sent home to sustain a family in most cases, the inflow is often regular – thus comparable to ‘regular income’.
Also, family earnings will tend to increase, thereby lessening the problem of the income effect for having a bank account. All three effects make the receivers of remittances more interesting clients for banks. By taking away the reasons mentioned for not being banked, whether demand or supply-led, the receipt of international remittances may trigger an outward move of the access frontier. Receivers become more interesting clients for banks, and they themselves can also find more reason to make use of banking services.
The effects mentioned above will depend on ability and willingness of banks to adapt. Assuming they are interested in remittance inflows from abroad, their product offerings should expand accordingly.
Examples are banks that offer low cost or free international transactions on the condition that clients have a bank account with them. However, as Prahalad (2004:8) notes, private-sector businesses have a dominant logic that may restrict their ability to see a dynamic and viable market opportunity at the
‘bottom of the pyramid’. On the demand side, the problem of inadequate regulation and distrust of banks can hinder development of financial access considerably. Adequate government regulation and policy in this area, the next topic, will also play a role.
D. Policy and Regulatory Issues
By using policy geared at integrating remittance senders into the formal economy, governments can create a more inclusive financial sector and a more efficient and formal economy. They can increase financial depth of their economy and improve monitoring of financial flows. The incentives to get involved with remittances are numerous, and not solely based on an inclusive financial sector (see arguments in section one). One way to for governments to affect access to formal financial services in the country is by stimulating remittance sending through formal channels. This puts migrants and remittance receivers in touch with diversified financial institutions, and can lead to increased demand and supply of other financial products, in line with the arguments above.
Governments can stimulate remitting through formal channels by removing taxes on incoming remittances, relaxing exchange and capital controls, allowing domestic banks to operate overseas, providing ID cards for migrants (such as the matrícula consular in Mexico), supporting hometown associations and providing matching grants, offering loan/pension schemes and bonds targeted at the diasporas, and by actively supporting the diaspora to help ensure the welfare of their citizens abroad (World Bank 2005c:95). Also, educating its population about the benefits and processes of financial institutions can increase demand for formal financial services (World Bank, 2003b). Each of these measures makes it more attractive for diversified financial institutions to enter the remittance market, and for clientele to send their money through formal channels. Two regulatory issues in moving towards formal channels are highlighted below: migrant identification requirements and regulation on money laundering and terrorist financing.
Migrant Identification. In using formal channels to remit, a valid immigration status is often a problem.
Migrants without legal status abroad lack identification to open a bank account or to use the banking system for transferring funds. Surveys of migrants in Los Angeles and New York show that they are discouraged from opening bank accounts by minimum balance requirements and strict identification requirements (Ratha, 2003:35). In this case, migrants tend to resort to money transfer organisations (MTOs) or to informal networks. By accepting other forms of identification for opening bank accounts,
