Report No. 35771-TJ Tajikistan Policy Note Enhancing the Development Impact of Remittances June 2006 Poverty Reduction and Economic Management Unit Europe and Central Asia Document of the World Bank TABLEOF CONTENT ACKNOWLEDGEMENTS ......................................................................................................... V EXECUTIVESUMMARY ....................................................................................................... vi1 I INTRODUCTION ............................................................................................................... 1 I1 BACKGROUND .. .................................................................................................................. 2 MIGRATION TAJIKISTAN................................................................................................... IN 2 WHO ARETHE MIGRANTS? ................................................................................................... 4 OVERVIEW OF REMITTANCESINTAJIKISTAN ....................................................................... 6 I11 ALTERNATIVE ESTIMATESOFREMITTANCES 7 IV.. ..................................................... TRANSFERMETHOD, USES,AND IMPACT OF REMITTANCESIN TAJIKISTAN ..................................................................................................................... 11 TRANSFERMETHODSREMITTANCES ............................................................................. OF 11 USES OF REMITTANCESINTAJIKISTAN ............................................................................... 14 IMPACTS OF REMITTANCESINTAJIKISTAN ......................................................................... 15 V. ENHANCINGTHE DEVELOPMENTALIMPACT OF REMITTANCES- LlSSONS LEARNEDFROMINTERNATIONALEXPERIENCE ........................... 18 MEASURES TO MOBILIZEREMITTANCES INTOTHE BANKINGSYSTEM 18 USING MICRO-FINANCE INSTITUTIONS TO MOBILIZE REMIlTANCES ................................. .............................. 20 MEASURESMOBILIZECOLLECTIVEREMIITANCES ....................................................... TO 21 LESSONSLEARNED .............................................................................................................. 24 V I . ENHANCINGDEVELOPMENTIMPACT- WHAT CANTAJIKISTAN DO? ......25 RECOMMENDATIONS ........................................................................................................... Overarching Policies to Enhance the Development Impactof Remittances ..................25 25 Improvements of BusinessEnvironment ................................................................... 25 Strengthening Financial Intermediaries ................................................................... 25 Improving Information Dissemination and Statistics on Remittances and Migration .................................................................................................................................. 27 Policies Measures to Improve the Welfare of TajikMigrants Abroad ..................... 28 AREAS FURTHER FOR STRENGTHENING AND RESEARCH .................................................... 30 ANNEX I:A SUMMARY OF MIGRANT SURVEY ............................................................. 33 ANNEX 11: ESTIMATIONOF REMITTANCEFLOWSINTOTAJIKISTAN .................35 ANNEX 111: TRANSFEROF REMITTANCESTO TAJIKISTAN ..................................... 39 ANNEX IV: MONEY TRANSFERSYSTEMINTAJIKISTAN .......................................... 45 ANNEX V: MICROFINANCE INSTITUTIONSINTAJIKISTAN ..................................... 51 REFERENCES ............................................................................................................................ 53 iii LIST OF TABLES Tab`le2: 1: TajikistanMigrationFlows (thousandpersons)............................................. 3 Tab#le 4 Tablle 3: Financial Snapshot of a Tajik Migrant............................................................ Tajikistan's Demographic Profile .................................................................... 5 Tablle 4: Tajikistan Balance of Payments ....................................................................... Tabl e 5: ECA's LargestDeveloping Country Recipients of Remittances. 2003 .........77 TabIle 6: Correlation between Methods of Transfer and Transport Mode...............11 TabIle 7: Annual Consumption and Remittances per capita by Quintiles .................16 Tabde A-1.1: Distribution of Survey Samples............................................................... 34 Tabde A-11.1: Estimated Remittances ........................................................................... 36 TabIleA-11.2: Choice of Remittance Channel by Tajik Migrants .............................. 37 Tabde A-111.1: Money Transfer Systems inTajikistan................................................ Tabde A-111.2: SelectedIndicators of Financial System (inthousand Somoni).........44 41 Tabde A-IV.l: Pricing of Money GramRemittance Services..................................... 46 LIST OFFIGURES Figure 1: Migrationof Skilled Workers inEastern Europe and the Figure 2: .............................................................................................................................. Former Soviet Union ......................................................................................................... 5 6 Figure 4: Tajikistan Uses of Remittances (% of Respondents) .................................. Figure3: Tajikistan Inflows and Outflows of Remittances.......................................... 6 14 Figure 5: Tajikistan - Impact of Remittances on LivingStandards .......................... 16 Figure7: Remittance Securitization Structure............................................................ Figure6: Employment Sector for Tajik Migrants....................................................... 17 22 FigureA-111.1: FundTransfers through various Commercial Banks in2004 (% of Total Transfers)..................................................................................................... 39 LIST OFBOXES Box 1: Official Estimate of Remittances......................................................................... 8 10 Box 3: Cross-selling by Tajik Commercial Banks....................................................... Box 2: Estimate of Remittances through InformalChannels..................................... 12 Box 4: Tajikistan's Banking System............................................................................. 13 Box 5: Benefits of Involvement of MFIsinRemittances............................................. 21 43 Box A-IV.l: Sending and ReceivingMoney via Western Union................................ Box A-111.1: Tajikistan's ATMs and PTMs.................................................................. 46 iv ACKNOWLEDGEMENTS This Policy Note is part of the Country Economic Memorandum for Tajikistan. The analysis i s based on the survey of migrants conducted by Tajikistan's Sociological Research Center (Zerkalo) from April through May 2005, the findings of several missions that visited Tajikistan between February through July 2005, and the extensive review of international experience on migration andremittances. The Policy Note i s prepared by a team comprising Jariya Hoffman (task team leader, ECSPE), Bryce Ramsey-Quillin (ECAVP), Carlo Segni (ECSPS), and Paul Jacobs (summer intern). The team wishes to thank the Government of Tajikistan for the constructive discussions and inputs provided during and after the various missions. The team i s particularly grateful for the inputsprovidedby Faizullo Kholboboev (State Adviser to the President on Economic Policy), Mr. M. Alimardonov (NBT Chairman), Mr. Jamshed Yusupov (Head, Monetary Policy and Statistics Department, NBT) Mr. Asadullo Gafurov (Head, BOP Division, NBT), Mr. Anvar Babaev (Head State Migration Service, Ministry of Labor and Social Protection), Mr. Asoev Umed Abdulloevich (Head, Department for Labor Market and Population Employment), Mr.Halimjon Rahimov (Director, Tajik Labor Export Agency-Tojikhofijakor, Ministry of Labor and Social Protection), Mr.Fattoh Saidov (Head, Visa and Registration and PassportAffair Division, Ministry of Interior), Samandar Haidarovich Khaidarov (Head Foreign Economic Activities and Forex Operations, AgroInvestBank), Mr. Umed Davlatzod (First Deputy Chairman, Orionbank), Mrs. Larisa Loginova (First Deputy Chairman, Tojiksodirotbonk), Mr.Muzaffar Zaripov (Director IOM-OSCE Information Resource Center for labor migrants), Aziz Sharipov (Chief, International Relations Department Kafolatbank), Ms. Saodat Olimova (Head, Social Service Scientific Research Center "Shark), Mr.KahramonBakozoda (Sociological Research Center, Zerkalo), Alastair Watson (Chief Operating Officer, the First MicroFiannceBank), Michael Schneider (Manager, Project "Support of Microfinance Services in rural areas," GTZ), Shamsiddin Gulov (Branch Director, International Micro-Loan-Fund "IMON')), Mahmoud Naderi (Chief of Mission, IOM) and Mr. Phillip Wanners (Swiss Forumfor Migration and Population Studies, Switzerland). The author would like to express her gratitude to the Swiss Government for its generosity inproviding financing support through the Consultant Trust Fundfor the survey of migrants inTajikistan. The Note is drawn upon several background papers contributed by Bryce Ramsey-Quillin and Paul Jacob (Estimation of Remittances), Carlo Segni (Transfer Channels and Costs and Micro-finance Institutions), and Paul Jacobs (Review of International Experience) provided in Annex 11-IV. It benefits from discussions with Ali Mansoor and Bryce Ramsey-Quillin and insightfulandconstructive comments fromthe peer reviewers, Dilip Ratha andJohn Litwack. This report was prepared under the general guidance and advice of Cheryl Gray, Carlos Felipe Jaramillo, Samuel K. Otoo, Deborah L.Wetzel, Dennis de Tray, Annette Dixon, Cevdet Denizer, Sudharshan Canagarajah, Lilia Burunciuc, and Ashad Sayed. Bryce Ramsey-Quillin and Ali Mansoor kindly provided overall guidance and advice as well as shared the survey data from the ECA migration study. The report also benefited from valuable suggestions and inputs received at different stages of preparation from SudharshanCanagarajah, Jakob von Weizsacker, Samuel Munzele Maimbo, Raul Hernandez-Coss, Utkir Umarov, and Aziz Khaidarov. Paul Jacobs provided valuable and excellent assistance in analyzing the migrant survey database and conducting the literature review. Processing assistancefrom Dammika Somasundaramand Zakia Nekaien-Nowrouz i s gratefully acknowledged. Helpful logistical assistance was also received from Munira Salieva, Janna Yusonova, Takhmina Jumaeva and Zarina Shukurova. vi EXECUTIVESUMMARY 1. Since its transition to a market economy in 1991, Tajikistan has experienced high levels of out migration. The motivations for this migration, however, have evolved over time. Early emigration was motivated by war and the conflicts following independence, while more recent migration has been provoked by economic factors influenced by demand and supply factors. An increase in the demand for labor driven by a decline in the working age population in Russia as well as strong economic growth in Russia, combined with slow job growth and the low wage level inTajikistan, have encouragedthe emigration of Tajiks. 2. Who are these migrants and how many are they? A migration survey conducted in May 2005, shows that the average migrant i s a male ethnic Tajik, with an average age of 34 years, married and with 5 dependents. The majority of migrants have secondary education or vocational training. Most are unemployed or grapple with a low wage level that is insufficient to meet basic household needs. They travel alone and mainly head for Russia for an average of 14 months and then return to Tajikistan. While in Russia, most Tajik migrants work in the construction, services, and manufacturing sectors. Given the dearth and unreliability of migration statistics, the true number of Tajiks living outside the country i s unknown and a large number of them are undocumented migrants. One official estimate suggests the number was about 421,000 in 2004, while other estimates indicate numbers up to 1.2 million Tajiks living overseas. 3. What i s all the mor: remarkable about Tajik migrants i s that they are the source of a large volume of funds that are being sent back to their home country inamounts so significant that they have now become an important source of economic growth for Tajikistan. In fact, the level of remittances ranks second after aluminum and ahead of cotton if considered as an export of labor services in 2004, and towers over those of official development assistance as well as foreign direct investment. These inflows have a strong seasonal orientation that reflects, inparticular, the temporary nature of migration to Russia. Between 2000 and 2004, official estimated remittances increased from 0.1 percent to 12 percent of GDP. In addition to employment opportunities in Russia (and in other host countries), the reasons that official remittance transfers have increased since 2000 include the removal of the 30 percent tax that had been levied on such transfers, reduction in money transfer costs, improvements in the financial system's infrastructure, and improvedtrust in local banks. 4. Though already large, it i s widely agreed that the true size of remittance flows are unquestionably larger than the official balance of payment estimates. Alternative methods of estimating remittances produce widely different numbers depending on the estimation methodology used. For example, the estimate basedon the Living Standard Survey i s even lower than the official estimate, which clearly undershoots the true mark. An estimate based on a micro-survey of returned migrants suggests that remittances could amount to 28 percent of GDP, although the finding i s contentious because of the unreliability of migrant statistics and the biases of the survey methodology. Alternatively, one regression-based estimate suggests that remittances could be as highas 21percent of GDP. 5. Remittance transfers to Tajikistan are relatively efficient as measuredby the simplicity of operation, cost and speed. The most widely used transfer mode among Tajik migrants is the money transfer system (cash-to-cash transfer) that operates in partnership with Tajik commercial banks (about 70 percent of all operations). These require no documents from senders, and have a fee that averages anywhere from 1.5 to 5 percent of the amount sent (which i s charged to senders in host countries and i s considered relatively low compared to similar systems in countries such as Mexico and Philippines). Competition among banks and money transfer operators have driven rates down, and even a global company like Western Union was obliged to lower costs in Tajikistan. Recipients can withdraw cash in a short time (within a day in Dushanbe), and in several currencies (US$, Pound Sterling, Ruble, Euro, and Somoni). Remittances sent through the banking system (account to account transfer) are insignificant due to high transfer costs (that includes a fee of 3 percent paid by both senders and receivers, and the exchange rate spread of 2- 3 percent), limited access to the banking system, and lack of bank accounts abroad and at home. Despite its extensive network in the rural areas, transfer through the post office system i s inefficient due to high fees (10-12 percent) and very slow speed (2-3 weeks). It i s estimated that about 30 percent of migrants send remittances through informal channels, depending mostly on personal courier, friends, and acquaintances. 6. Remittances to Tajikistan are highly significant as they have become a source of income for many low income families to meet basic needs and thereby help reduce poverty. They are clearly welfare enhancing for households who benefit from these flows. The survey of Tajik migrants suggests that the financial situation of Tajik migrants and their families improve after migration. Their income overseas i s ten times what they earned in Tajikistan before migration (US$274 per month compared with a pre-migration income of US$30 per month), of which about 48 percent i s remitted to Tajikistan. Remittance recipients earmark funds for consumption, home repair, education and health, and savings. Consequently, they have contributed to short-term ' economic growth and poverty alleviation through increased domestic demand that leads to increased employment, income and output. At present, their contribution to longer term growth through investments in productive activities i s limited by the low level of migrants' income and an unfavorable business environment that constrains the development of small and medium enterprises (SMEs). However, their contribution to funding the education and health needs of families i s arguably a contribution to investment in human capital that can yield benefits both in the near- and medium-term. 7. From a macro perspective, the international literature shows that large inflows of remittances can also lead to a real appreciation of the exchange rate, which can fuel imports and worsen the country's export competitiveness. Further, they may increase the vulnerability of recipient countries to host countries' migration policies and business cycles. Such macroeconomic impacts have implications for long-term economic growth and thus deserve a comprehensive analysis. To mitigate the potential negative impacts and to enhance their longer term positive impact, the government of Tajikistan i s seeking appropriate policies on how to further enhance the use of remittances beyond increasing the standards of living and welfare of household recipients. This Note attempts to bring them into focus and to suggest measures for the government to consider enhancing the development impact of remittances. 8. The report contains a review of the latest international policy experiments to enhance the development impact of remittances; however, many of these experiments, unfortunately, have either not been properly evaluated or are more relevant for middle income countries, where financial sectors are more developed. This Note considers some of the most relevant international experience concerned with enhancing the development impact o f remittances. Such experience i s instructive and suggests that there are indeed a variety of policy instruments that can be effective in mobilizing remittances into the banking system and channeling their use for more productive purposes. viii 9. The conclusion of this analysis i s that policies need to be designed for Tajikistan's specific needs. After considering the international experience and Tajikistan's particular conditions, the note assesses the potential for replication of the most successful of these policies that were applied elsewhere to the case of Tajikistan. For example, to meet specific development objectives, collective remittances have been mobilized by the Home Town Associations (HTAs) in several Latin American countries and matched by government funds (known in Mexico as Three-for-One). The experience of matching funds that has been piloted in Tajikistan by UNDP in collaboration with the IOM is considered inconclusive and it is too early to evaluate the program's development impact. 10. The Note considers those policies that would increaseremittance inflows and enhance the development impact of remittances, increase welfare and broaden the safety net for migrants and their families. It recommends a set of policies aimed at optimizing their positive impacts by improving the investment climate, and the role of financial intermediaries. It also offers a second set of policy recommendations that focuses entirely on bilateral negotiations, especially with Russia, concerning the rights of Tajik migrants, and resident and work permits when working in host countries. Finally, the report explains the importance of improving the quality of migrant and remittance statistics to improve the understanding of their impact and as an aid in the designing of appropriate policies. 11. In terms of policies that can be implemented unilaterally to enhance the development impact of remittances, the Tajik authorities could focus on the following: (a) Encouraging a pro-business environment, This requires maintaining a stable macro- economy and simplifying the licensingand inspectionregimes. (b) Enhancing the role of financial intermediaries for mobilizing savings and investment resulting fromremittances. Such policies include: Encourage entry of foreign banks into Tajikistan to foster modernization and technological improvements, as well as to further increase public trust and confidence inthe banking system; Cross-selling of financial products by commercial banks to migrants as incentives for keeping their savings in the bankingsystem; Promote healthy competition of commercial banks to further reduce transfer fees, especially wire transfers charged by commercial banks; Improving public information dissemination related to financial services by both the government and commercial banks; and Integrating MFIs into the remittance services to mobilize remittance savings from un-banked migrants in the rural areas and channel them towards productive investment including small business and agricultural projects. (c) Inthe near-term, efforts should be undertaken to improve the quality and coverage of statistics on migration and remittances, including estimating informal transfers of remittances inthe balance of payments and improving household budget surveys. i x 12. In terms of policies designed to increase remittance inflows that are not fully under the control of the Tajik authorities. The Note outlines the potential benefits of a proactive engagement in bilateral diplomatic negotiations with Russia to remove obstacles to migration. The main aim of such negotiations should be to: (a) Protect the right of undocumented Tajik migrants on Russian territory; (b) Simplify the registration process of resident permits for Tajik migrants; and (c) Lower the cost to Russian employers when hiringforeign workers. X I. INTRODUCTION 1. Remittances have played an important role as one of the drivers of Tajikistan's robust economic growth during the past several years. The volume of official remittances has significantly increased since 2001 and now represent close to 12 percent of GDP (2004). Remittances have become the most important source o f external financing for the balance of payments, have increased incomes, and as a result helped reduce poverty. Emigration- particularly seasonal migration to Russia-has become an alternative poverty coping measure chosen by the poor. According to the 2003 Tajikistan's Living Standard Survey, remittances and other transfers made up about 10 percent of average household income in 2003, ranking it the second source of cashincome following wages.' 2. Recognizing their importance, the Tajikistan authorities have sought to sustain remittance flows. In order to increase the quantity of legal migration, temporary worker agreements have been signed bilaterally with key receiving countries (Russia, Kyrgyz Republic). A 30 percent tax on remittance transfers was eliminated and commercial banks were allowed to partner with foreign money transfer operators to provide low cost transfer services to migrants. The Ministry of Labor and Social Protection established an employment agency to provide employment services including information on vacancies, training, and orientation for workers choosing to live abroad, 3. However, anecdotal evidence indicates that migrants are not sufficiently integrated into the Tajikistan financial system, much less that of the destination countries. Most do not have bank accounts at home or abroad, and as a result remittances are mostly transferred through money transfer systems on a cash to cash basis. They are immediately withdrawn in cash once reaching Tajikistan in order to meet basic needs. If there are savings, they prefer to keep them at home because of a lack of trust in the banking system. This leads to inadequate integration of migrants and their families into the financial system, thus limiting access to financial services and products that could smooth consumption andprovide additional investment for income generation activities. 4. The survey of Tajik migrants found that the majority o f remittances are used to fund basic household consumption, while an insignificant amount of remittances are spent on productive investment. Most migrants come from poor, rural areas and the bulk of international remittances appear to fund basic consumption and some investment and saving at the margin. Although this pattern of remittance use i s common around the world, it i s recognized that the development impact of investment can be greater than consumption expenditure. To maximize the potential impact of remittances, the Tajik authorities have expressed keen interest inreceiving guidance on how to further encourage their use for greater productive investment. 5. For remittances to have an impact beyond an income effect, two conditions must be met. First, households must have surplus income after all subsistence needs have been met. Second, investment policies and institutions must create the incentives to encourage households to save any surplus income in the formal, financial sector or invest in businesses. In the case of Tajikistan, it i s not yet clear whether households have large of amounts o f surplus income. However, efforts could be made to improve the investment climate not only to encourage greater 'The World Bank, Poverty Assessment Update (June 2004). savings and investment from remittance income at the margin, but also to take greater advantage of Tajikistan's migrant population. Improved policies to encourage higher levels of remittances will facilitate greater trade and migration complementarities, and provide more opportunities for migrants to use the financial and human capital built abroad back home. 6. This Policy Note has two main objectives: (i)to expand knowledge of the size of remittances, their sending mechanisms (especially through formal channels), and their use; and (ii) proposepolicyrecommendationstoenhancethedevelopmentimpactofremittances. The to Note benefits from a recently completed study (Tajikistan's Trade Diagnostic Study) and from the World Bank's ongoing Study on "Enhancing Gains through International Labor Migration in Europe and Central Asia," and the 2005 Global Economic Prospect featuring "International Migration and Remittances."2 The survey of Tajik migrants conducted by the Zerkalo3 during April-May 2005 for the migration study in the ECA region provides insights to Tajikistan's migration and remittances. It provides demographic characteristics and economic behavior of migrants, thus enabling the computation of economic and financial parameters underpinning the analysis. Finally, reviewing the international experience of other selected labor sending countries around the world provides lessons that can be drawn for the Tajik government in designing relevant policy measures aimingto enhance the development impact of remittances. 7. The Note is organized into six sections, including this Introduction. Section Two provides background on migration and remittances related to Tajikistan. Section Three discusses the estimation of remittance size and data issues. Section Four examines the transfer channels of remittances (their cost, access, and speed), their use, and the impact of remittances. The analysis highlightsimpediments to the efficiency of remittance transfer services through formal channels, examines the use of remittances by recipient households in Tajikistan, and identifies impacts of remittances on the economy. Section Five discusses the international experience of other selected migrant-sending countries in using various instruments to enhance the development impact of remittances. The section draws some of the key lessons learned from the international experience that could be considered in the Tajik context. The final section recommends policies to enhance the development impact of remittances taking into account the Tajik context and identifies topics for further research. 11. BACKGROUND MIGRATIONTAJIKISTAN IN 8. Tajikistan's transition to a market economy has been challenging. The implementation of structural reforms was delayed by the multi-year civil war that broke out after independence in 1991 and did not commence until 1997. The country inherits a difficult geography - it i s land- locked, resource poor, mountainous and faces an unfavorable geo-political backdrop bordered by Afghanistan and Uzbekistan. The civil war contributed to the deterioration of the physical infrastructure and an already weak institutional capacity. Despite strong economic growth, averaging 9.6 percent over 2000-2004, Tajikistan remains the poorest economy in Central Asia with a per capita GDP of US$310 in 2004. Poverty has fallen somewhat, but remains high at The World Bank, 2005. EnhancingGainthroughInternationalLaborMigration inEuropeand CentralAsia. The Zerkalo is a SociologicalResearch Center inTajikistanthat carriedout the survey of 1,200 migrants financed by the World Bank's ECA Migration Study during April-May 2005. The survey i s basedon the same set of questionnaire that was designed for 6 countries that were case studies for the World Bank Study. Based on the survey results, Zerkalopreparedareportentitled"Assessment of Standardof Living of Labor Migrants." 2 64percent and income inequality has widened since 1999.4 Lackluster job growth due to demographic factors combined with delayed structural reform makes it difficult for the population to find employment and earn sufficient income to meet basic needs, especially in the rural areas. A good segment of the population has found its way out of poverty through international migration, primarily to various parts of Russia. 9. International migration in Tajikistan Table 1: Tajikistan Migration Flows i s a relatively new phenomenon relative to (thousand persons) other labor exporting countries around the Arrived Departed Migration world, especially Mexico, Philippines and inflows South Asian countries (India, Pakistan, Sri (+Youtflows (-) Lanka, and Bangladesh). Prior to 1991 74.9 101.3 -26.4 independence in 1991, Tajikistan's migration 1992 51.3 146.0 -94.7 was considered internal within the former 1993 71.4 146.1 -74.7 Soviet Union. The eruption of ethnic conflict, 1994 43.3 88.9 -45.6 followed by a multi-year civil war that began 1995 37.1 74.9 -37.8 in 1991, then liberalization of migration after 1996 26.1 53.7 -27.6 independence and economic transition, all 1997 20.8 37.1 -16.3 resulted in a large flow of international 1998 16.9 32.3 -15.4 migration. Based on the available official 1999 14.7 28.8 -14.1 migration data reported by the State Statistical 2000 14.5 28.2 -13.7 Committee, about 411,623 persons left 2001 16.7 29.1 -12.4 Tajikistan during 1991-2004, of which 2002 17.7 30.2 -12.5 68 percent left during 1991-95 (Table 1). 2003 16.9 27.9 -11.0 They were mostly non-ethnic Tajik (Russian, 2004 15.2 24.7 -9.4 Jews, Crimea Tartars, Ukrainians, Kazakhs, Source: State Statistical Committee. and German) and their main destination countries were Russia, Uzbekistan, Germany, Kazakhstan, Israel, and the Kyrgyz Republic. However, emigration declined sharply in 1997 and leveled off duringthe 1998-2004 period as the political and economic situation gradually normalized. 10. Thepattern of Tajikistan's migration in recent years has evolved. While international migration had been driven by ethnic conflict and civil war, recent emigration has been driven by economic factors, both on the demand and supply sides. The demand factors-demographic crisis in Russia due to a declining and aging population5 and continued economic growth in Russia-generated significant demand for labor, especially in the service sector. The supply factors in Tajikistan included a low wage level, highunemployment6due to skill mismatches and slow job growth relative to labor force g r ~ w t h ,all of which provided great incentives for ~ workers to seek opportunities outside the country. A free visa regime between Tajikistan and the According to the World Bank 2003 Poverty Assessment Update, the inequality of consumption expenditure increased from 0.33 in 1999 to 0.36 in2003. Russia's natural rate of increase (-0.6 percent) ranks the second lowest in the world after Ukraine (-0.8 percent). Population in Russiadeclined to 143.5 million (December2004) from 145.2 million (2002) due to natural causes, a low birth rate and emigration. The World Population Data Sheet projects that Russia's population will decline to 119 million by 2050. According to a demographic forecast for 2050 the share of population in the working age group (between 16 and 55 for females and 60 for males) will be close to 50 percent, which i s considerably lower than 61 percent according to Census 2002 but similar to the Census 1939 with difference in higher proportion of old population, 34 percentin 2050 compared to 9 percentin 1939(Andreev and Vishnevsky, 2004). The World Bank's Poverty Assessment Update estimated the unemployment rate at 30 percent compared to the 'officialrateof 2.2 percent (September 2004). Tajikistan's working population represents 55 percentof total population. 3 CIS countries and Russia, combined with a shared cultural background and language, facilitated easy entry and exit for the Tajik migrants. The main destination countries were, and continue to be, Russia and Kazakhstan. Today's emigration i s characterized as "temporary" as confirmed by responses to the 2005 migrant survey in Tajikistan where 90 percent of migrants reported that they had no intention of settling abroad. According to the State Migration Services, most migrants are from Khatlon, RRS, and Sogd regions. 11. Much of the migrationfrom Tajikistan is undocumented. Despite restrictive migration policy and tight administrative control in Russia, migrants have overcome these barriers. Based on self-determination of legality* by the survey's respondents, about 4.2 percent of the migrants surveyed were categorized as legal; all others (most of whom carried only a passport) were deemed irregularhllegal migrants. The number of illegal Tajik in Russia i s unknown but large as the Ministry of Labor and Social Protection reported that only 10 percent of migrants work legally in Russia.' A free visa regime and weak enforcement enable migrants to cross borders and bribe their way into Russia. Tajik migrants in Russia are considered illegal" because they fail to obtain residence permits upon their registration as foreign citizens within 72 hours upon arrival in the country or they do not have employment permits. Russia still maintains internal administrative controls, including the registration system, a legacy of propiska." Most jobs in Moscow are only open to residents o f the city and the Moscow region even though it i s against Russia's Constitution. Although a lack of legal status makes Tajik migrants vulnerable to bribes and extortion, the survey on income of migrants proves that they are still better off than in their home country. WHOARE THEMIGRANTS? 12. Table 2 provides a summary of the Table 2: Tajikistan's Demographic Profile demographic characteristics of Tajik migrants. Ethnic Tajik 81% They are mainly ethnic Tajik, male, with an Married 72% average age of 34 years, married, and have Male 86% 5 dependents. The travel alone and mainly Average age 34 head for Russia1Y where they find job AverageNo. Of 5 opportunities, a common culture and language, No. Of monthsin migration (median) 9 and existing social network^.'^ The cost of Source: Zerkalo's Survey, M a y 2005 travel i s relatively low at about US$200 per trip by plane or train. Their main motivation to migrate i s higher income to provide food and necessities for their families back in Tajikistan. About 49percent of migrants remain abroad for 9 months. This profile i s similar to that of Mexico, except for the level of education. The flow of circular Mexican migrantlaborers heading to the United States i s mostly made up of males (94.4 percent), between the ages of 25 and 34 T h i s i s based on the respondents' answers whether they camed a series of documents(asylum papers or a work permit and passport)with them when they traveled abroad. Tajikistan Media Monitoring, March 10,2006. lo The Report on "Combating Trafficking in Human, Its Consequences and forms of Forced Labor in Central Asian Countries and RussianFederation," estimatedthat about 15-18 percentof Tajik migrants legally resideinRussia. " Inthe Soviet Union, thepropiskaor residentpermitsystemrequiredpersonsto registerbeforebeingallowed to migrate to a new location. Though this system did not completely capture all migration flows, it was a useful tool in tracking migration flows. Destination cities inRussiaare Moscow, St. Petersburg, Ekaterinburg, Novosibirsk, Chelyabinsk, and Krasnoyarsk. l3 A detailed demographic profile of migrants i s documented in "Migration and Return--The Case Study of Tajikistan," The World Bank, ECA Migration Study, 2005. This profile i s consistent with data provided by the State Migration Service. 4 (37.3 percent), with fewer years of schooling than the national average for Mexico, married (57.9 percent) and are the head of the household (65.1 per~ent).'~ 13. The lack of legal residence and work permits oblige Tajik migrant workers to accept unskilled and low paidjobs. A recent OECD study found that Tajikistan had one of the highest levels of out-migration of skilled workers inthe region (Figure 1). The Zerkalo survey found that one-third of migrants had secondary andor technical education. Despite their education, about 62 percent of Tajik migrants appeared to be employed inthe construction sector. Figure 1: Migration of Skilled Workers inEastern Europeand the Former Soviet Union 7 0 5 10 15 20 25 30 35 40 45 50 Expatnates that are highly skilled (% of total expatriates) 1 Source: OECD. 14. Migration has clearly improved the Table 3: Financial Snapshot of a Tajik Migrant financial situation of Tajik migrants and Average monthly income (excluding Unemployed) their families. Table 3 provides a snapshot Premigration $30 of the financial status of a typical migrant Duringmigration $274 based on self reporting from the survey. Post migration $91 Overall, migration increased the income of Saving/Income 0.015 migrants 10 times compared to what they Remittances/Income 0.476 earned in Tajikistan (US$274 per month Source: Zerkalo compared with pre-migration income of US$30 per month). About 48 percent of this income is remitted to families in Tajikistan, thus resulting in an average monthly income of about US$131 per household. Remittances have enabled families back home to afford food, clothes and utilities to a muchgreater degree. l4The data i s based on the ENEFNEUSurvey (EncuestaNacional de Migracidn a la Frontera Norte y a 10s Estados Unidos)carried out by the Centro Nacional de Informacidn y Estadisticasdel Trabajo to obtain an accurate estimate of Mexican workers in the United States and their socio-economic characteristics. See Germin A. Zirate-Hoyos, 1999. "A New View of Financial Flows from Labor Migration: A Social Accounting Matrix Perspective," E.I.A.L. Vol. 10, No. 2, July-December 1999. 5 OVERVIEW OF REMITTANCESINTAJIKISTAN 15. Remittance flows into Figure 2: Tajikistan Monthly FundsTransfer - Tajikistan have a strong seasonal throughBanking System(mil. US$) orientation that reflects the temporary nature of migration to Russia. The 90 0 private inflows of funds transferred 80 0 70 0 through the banking system and 60 0 captured by the National Bank of 50 0 40 0 Tajikistan (NBT) in the balance of 30 0 payments oscillated during the year 20 0 10 0 (Figure 2). They peak in October after 0 0 the demand for short-term labor in Jan Feb Mar Apr May Iun Iul Aug Sep OctNOV Dec Russia reaches its highest. - c 2 0 0 2+2003 2004 2005 Subsequently, they taper off during November and December as winter slows the demand for labor, especially in the construction and agricultural sectors. sharply during the past few years and are hrge as a portion of GDP (Figure 5000 3). According to the NBT, inflows of 450 0 4000 migrants' remittances tripled from 350 0 3000 US$78 million in 2002 to US$465 2 2500 ~ 2 0 0 0 million in 2005. The rapid growth of official remittance inflows i s due to 50 0 several factors: increased emigration in 0 0 2000 2001 2002 2003 2044 2005 response to the demand for labor, innows O U ~ O W S l5Resolution No. 445, dated 25 September 2001, cancelled the resolution of Council of Ministers of the Republic of Tajikistan No. 587 of 29 November 1993 "On State duty rates" that imposed 30 percent state duty on incoming money transfers from abroad. l6T h i s i s due to a significant enlargement of the network of money transfer offices in all regions of Russia, CIS and FSU. l7Anecdotal evidence suggests that the financial and banking system has been re-monetizing fast, with deposits in banks to assets and to GDP ratios increasing throughout the whole region and Russia. The NBT17estimated that the number of individual accounts in banks has steadily increased during the last three/four years, indicating improved confidence levels and intermediation deepening. 6 17. Remittances play an Table 4: Tajikistan Balance of Payments important role in the balance of 2000 2001 2002 2003 2004 2005 payments. Table 4 shows that if Trade Balance -23 -125 -125 -204 -151 -319 labor services are considered as Exports 788 652 699 799 1088 1107 exports, migrants' remittance of which: Aluminum 424 398 399 430 563 563 Cotton 92 71 128 193 162 144 inflows rank second after Imports 811 777 824 1003 1239 1426 aluminum and ahead of cotton Service Balance 4 -8 -36 -54 -89 -117 fiber. They accounted for nearly Net Income -36 -72 -57 -90 -78 -75 Net Transfer 37 131 184 329 236 432 one-third of merchandise exports of which: Net Migrants' Remittances 0 -1 65 82 133 321 and financed about 70 percent of Inflows 1 4 78 146 252 465 outflows -1 -5 -13 -64 -119 -144 the trade deficit fueled by Current Account Balance -18 -74 -34 -19 -82 -79 imports induced by increased Financing: income associated with Net Public Sector Borrowing 39 6 14 32 -169 50 FDI 24 9 22 32 272 55 remittances since 2000. Changesin Reserves 1 -9 21 35 -54 -41 Remittances also tower above Others 46 -68 23 80 -33 -15 foreign direct investment and Stock of Internatonal Reserves 87 96 96 135 189 224 Memorandum items: other transfers generated by Gross RemittanceslExports 0% 1% 11% 18% 23% 42% official development aid flows. Gross RemittancesmradeDeficit 3% 3% 63% 72% 167% 146% GrossRemittancesFDI 3% 47% 356% 456% 93% 852% In2005, gross remittance inflows GrossRemittmemet Borrowing 2% 70% 560% 456% -149% 932% doubled the stock of Tajikistan's Gross Remittmces/Gross Reserves 1% 4% 82% 108% 133% 207% inter-national reserves. Sources: National Bank of Tajikistan and IMF. 18. International Table 5: ECA's Largest Developing Country Recipients comparisons of remittance of Remittances, 2003 recipients, as a share of GDP, Remittances Migrants GDP Remittances Remittances shows that Tajikistan ranked the Mill.US$ inMill. Mill.US$ per migrant % of GDP fourth among the largest HighEmigration Moldova 465 1 1,566 603 29.70% remittance recipient countries in Bosniaand Herzegovina 1,178 2 5,072 785 23.20% the ECA region in 2003. Table Albania 889 1 4,411 986 20.20% Tajikistan 146 1 1,313 133 11.10% 5 shows that Moldova, Bosnia KyrgyzRepublic 108 1 1,539 163 7.00% and Albania rank the first, Georgia 246 1 3,737 247 6.60% Armenia 168 1 2,700 189 6.20% second, and third, respectively. Macedonia,FYR 171 0 3,565 548 4.80% According to the revised official REGIONALAGGREGATES statistics, remittances to SouthAsia 27,127 14 817,617 1,883 3.30% MiddleEastand NorthAfrica 17,049 13 524,755 1,327 320% Tajikistan accounted for 11 Sub SaharanAfrica 6,026 8 327,117 731 1.80% percent of GDP, compared to an LatinAmerica 34,425 31 1,987,349 1,114 1.70% 12,406 887,472 231 1.30% aggregate of 14 percent of GDP ECA 51 East and SoutheastAsia 24,997 21 8,065,940 1,182 0.30% for all high emigration countries. WORLD 173,103 177 33,527,651 980 0.50% In nominal terms, Tajikistan's Source: The World Bank, "Enhancing Gains through International Labor Migration in Europeand Central Asia" (2005). remittance receipts rank the secondlowest (US$146million) after the Kyrgyz Republic (US$lOS million). 111. ALTERNATIVE ESTIMATESOFREMITTANCES 19. Despite the importance of remittances, there is great uncertainty regarding their actual levels and growth rates. There i s no consensus on the true size of remittances. Remittances are private transfers and are notoriously difficult for any government to monitor and measure as unknown amounts are transmitted through informal channels, including carrying by self and friends or through intermediaries. Ingeneral, official remittance statistics shown in the balance of 7 payments" around the world are grossly under-reported. A recent World Bank survey of central banks found that 75 percent do not attempt to estimate the amounts at all.19 Tajikistan's official estimate, for example, excludes remittances sent through informal channels. It i s well known that a large number of Tajik migrants work illegally abroad and travel back to Tajikistan with their earned salaries in cash or send money through informal channels. They do not trust the banking system or cannot access official money transfer systems because they are abroad illegally and do not have access to foreign bank accounts. Moreover, they may well find it more convenient and cost effective to bring money back home with them directly rather than pay intermediaries as a large number of Tajik migrants work abroad on a short-term seasonalbask2' ____ Box 1: Official Estimate of Remittances The Global Economic Prospects 2006 defines remittances as the total sum of workers' remittances, compensation of employees and migrant transfers. The Tajikistan's balance of payments, however, records only workers' remittances and omits the remainder due to a lack of information. The estimate of workers' remittances i s based on fund transferred through the banking system (wire transfer, travelers check, money orders, and money transfer system) that i s reported periodically by commercial banks. The NBT conducted two surveys o f migrants twice a year. This has enabled the NBT to differentiate migrants' transfers from export receipts. Previously, NBT classified all single transfers o f more than US$l,OOO plus 50 percent of transfers between US$l,OOO and US$3,000 as migrants' transfers; while all single transfers of more than US$3,000 and 50 percent o f transfers between US$l,OOO and US$3,000 are export receipts. This method yields the ratio of workers' remittances to export receipts of 60:40. The inflows o f workers' remittances in the balance of payments were estimated at US$252 and US$380 million, respectively in 2004 and 2005, compared to total fund transferred through the banking system of US$430 and US$664 million during the same period. The official estimate of Tajikistan's workers' remittance inflows now accounts for 12 percent in2004 and 17percent of GDP in2005. The NBT is currently improving the above estimation methodology by calculating the ratio of migrants' transfer to exports directly from the surveys in stead of using all single transfers (over and above US$3,000). The surveys in 2005 indicated that the share of migrants' transfers in total bank transfers has increased to 70-75 percent. Source: World Bank Staff. 18The Technical Sub-Groupon the Movements of Persons (TSG), chairedby UN Statistics Division with membership from centralbanks, internationaland national statistics organizations recommended that workers' remittances in the balanceof paymentsbe replacedby a new component "personal transfer." This includes all current transfer in cash or in kinds received by resident households to or from other non-resident households. A new aggregate, "personal remittances" should be reported in the balance of payments as a memorandum item including current and capital transfer bothin cash and inkind made or receivedby resident households to or from non-residenthouseholds, and net compensationof employees from persons working abroadfor ashort periodof time (less than one year). l9 Martinez, JosC de Luna (2005). Workers' Remittances to Developing Countries: A Survey with Central Bank on Selected Public Policy Issues, Washington, DC: World Bank, Mimeo. *'Three main types of Tajik labor migrants are distinguished-permanent, seasonal, and shuttle. Permanent migrants, hired under formal working contacts or informal working relationships,remain in host countries for more than one year. Seasonal labor migrants, usually unskilled workers travel individually or in groups to Russiathroughout March to November to work mostly in construction, mining or agriculture, and are mostly employed under informal contractual agreements. Shuttle workers travel individually or in groups and more periodicallythan seasonal workers for short period of times under established and de facto labor networks. These types of workers are relatively more skilled than seasonal workers. Also these workers tend to travel with cash rather then sending money through formal channels. 8 20. Understanding the true size of remittances is essential for policy makers. Reliable remittance data can contribute to prudent monetary policy managernent.'l It also provides inputs for the formulation of appropriate policies and products to improve income and living conditions of migrants and their families. Further, reliable data on the size of remittances can encourage private investment to improve infrastructure for transfer services that will reduce cost, improve access, and develop innovative financial products. This will better integrate migrants into the banking system, increase the deposit base and thus deepen the financial sector and increase productive investment. Finally, better remittance data will enable monitoring of cross border flows of money and the timely detection of money laundering or other illicit activities that undermine the integrity of the financial system. 21. Absent reliable data, efforts have been made to estimate the true size of remittances; however, there is no consensus as to their size. These estimates are based on different assumptions regarding migrants' remitting behavior and number of migrants. A variation of techniques to improve remittance statistics in labor sending countries has been developed during the past several years. The standard techniques commonly employed in the literature include both a direct and indirect approach. The direct approach relies on scaling up the estimate of remittances derived from the household budget survey2* and survey of labor migrants, while the indirect approach i s based on an application of estimated parameters of a cross country regression analysis. 22. Alternatively, remittunces can be estimatedfrom income remitted to migrants' families drawn from a labor migrant survey. Such a survey would provide two estimated values of the annual average remittances per capita that are congruous. The estimated remittances for the whole country inany year are generatedby the product of the estimated annual average per capita remittance and the stock of migrants. The estimated remittances using both values o f annual average remittance per capita are slightly different and amounted to 28 and 29 percent of GDP in 2004. However, this method i s also subject to the accuracy of the stock of migrants and the bias of the survey. An indirect estimate o f remittances was performed by using a percentage of migrants remitting funds through an informal channel to scale up the official remittances reported in the balance of payments. This method gives a remittance total of US$360 million or equivalent to 17 percent of GDP in 2004. Annex 2 discusses the estimated values of remittances using the data drawn out from the survey of migrants and the available estimated stock of migrants by various researchers and government sources. 21Ifunrecordedremittances are extremely large, then setting sound macroeconomic policy may be complicated by an appreciation of the national currency and inflationary pressure in light of limited monetary policy instruments and a shallow foreign exchange market. 22 The household budget survey, which ideally should capture a complete picture of household remittance income transmitted through formal or informal channels, produces a disappointing estimateof remittances. Assuming that the survey sample is nationally representative,it is possible to `scale' up the responses into a national picture of aggregate remittance amounts. However, the estimated remittances obtained from scaling up the Tajikistan Living Standard Survey in 2003 is US$23 million or only about 1percent of GDP, lower than the official remittancesin the balanceof payments (US$146 million), establishing a lower bound estimate of remittance levels. This is clearly due to under- reporting by Tajik householdsto questions about their expenditures and revenues. The householdsurvey approachmay not represent the best method to produce reliable point estimates. However, the approach is useful in constructing broaderinterval estimateof remittance levels when combined with other estimation methodssubsequentlydiscussed. 9 23. The recent estimation method of remittances is based on regression analysis.23 A recent World Bank and IMF study conducted a regression panel that sought to predict remittance levels with respect to a number of explanatory variables-including workers' income, predicted costs of sending remittances through informal channels, effect of dual exchange regime, net errors and omissions (capital flight). Applying the estimated parameters to Tajikistan, it results in remittance inflows equivalent to 22 to 30 percent of GDP in 2004, where the lower estimate corresponds to reducing the fee of remittances transferred through informal channels to 5 percent, and the higher estimate corresponds to reducing the cost to 2 percent. 24. The foregoing discussions on the estimation of remittances single out the need to improve the quality of statistics on migration and remittances in Tajikistan. In January 2004, the State Migration Committee introduced the Migration Registration Card24 to monitor international migration. However, the statistics are likely underestimated becauseof misreporting caused by weak capacity in data processing and compilation, and weak administrative control by the State Border Control Committee responsible for collecting cards from travelers. The National Bank of Tajikistan could improve the coverage of the remittance by providing detailed remittances sent through various channels including informal channels. The statistics from the latter sources can be captured through a survey of migrants. According to the survey of the central banks in 40 countries, 10 of them (Costa Rica, Ecuador, El Salvador, Honduras, Indonesia, Mexico, Moldova, Nicaragua, Peru, and Russia) have developed tools to measure the volume of remittance inflows that occur through informal channels.25 In these countries, information i s normally collected through surveys of migrants visiting their home country and/or household surveys. Box 2: Estimate of Remittances through Informal Channels The Moldovan National Bank's procedures for recording migrant's remittances in the Balance of Payments does record some transfer through informal channels. In addition to the standard practice of obtaining remittances transfer data from banks, the central bank monitors foreign exchange transactions made at bureaus de change and in the purchase of goods in which foreign currencies are often used (e.g., automobiles). Inthis way, the Bank i s able to pick up some of the funds that move into the country from abroad but are not sent through formal channels. Ecuador and Honduras conduct surveys of migrants at the country's main points of entry (usually airports) every year, especially during holiday seasons when the inflow o f migrant visitors reaches its peak (e.g., Christmas and New Year, Easter, Chinese New Year).26 I El Salvador conducts household surveys every year to estimate the total volume that they receive in remittances either through formal or informal channels. Sources: The World Bank, Moldova: Opportunities for Accelerated Growth- A Country Economic Memorandum, 2005. 1JosC de Luna Martinez (2005). 23 Freund, Caroline and Nikola Spatafora (2005)," Remittances: Costs, Determinants, and Informality," Washington, DC: World Bank and IMF. 24 The cardincludes informationon purpose, terms of departure, and estimateddurationof trip. 25 Jost de Luna Martinez, Workers' Remittances to DevelopingCountries: A Survey with CentralBanks on Selected Public PolicyIssues, the World Bank (2005). 26 The survey did not assess the scope, methodology and adequacy of the surveys conductedby this group of countries. T h i s i s an issue that deserves much more attentioninthe future. 10 IV. TRANSFER METHOD, USES,AND IMPACT OF REMITTANCES INTAJIKISTAN TRANSFER METHODS OF REMITTANCES 25. The Policy Note now turns to examine the main methods of transferring remittances in Tajikistan. Remittances "flow" to Tajikistan via both formalz7and informalz8channels. Table 6 shows that about 70 percent of migrants interviewed remitted money through a formal channel, the remaining remittedthrough informal channels. Due to the low level of the average remittance and the amount of money transfer per migrant, it i s quite common that remittances are pooled and transferred to Tajikistan ina single transaction by individuals residing inthe sending co~ntry.'~ informalThechannels 26. choice between formal and Table 6: Correlation between Methods of Transfer is driven by and Transport Mode transaction cost, level of trust in Modeof Transport electronic uavments and banks, Methodsof Transfer Automobile Plane Train preference fo; cash, level of education 2% 10% 13% and skill. A low transaction cost Operators 3% 20% 20% postofices 0% 1% 1% (transfer fee), increased confidence in the Friends/Acquaintan=s 3% 10% 12% 0% 0% 1% banking system, weak preference for cash, and a high level of education and ~ ~ ~ n T ~ ' 0% ~ ~ & 1% ~ ~1% 10% 42% 48% skilled migrants likely promote the Use Of Source: Zerkalo, Migration Survey Report, Tajikistan 2005. formal transfer channels through the bankingsystem. There are also preferences related to proximity and accessibility of migrants an2 relatives to banks and other intermediaries. Although the cost of sending cash through personal courier and friends is zero, the risk of carrying cash i s high for Tajik migrants due to informal payments to Russian, Uzbek and Tajik customs and immigration officers, particularly for those who travel by train and car that cross multiple borders. The majority of migrants traveling by plane (74 percent) and by train (71 percent) send money through official channels, while only a small number of migrants usedinformal channels to remit funds (Table 6). 27. The most popular method of formal transfer among migrants is through money transfer systems because they are relatively cheap, fast and they require no bank accounts. Money transfer systems have become a preferred mode to remit funds to Tajikistan among migrants. About 60 percent of respondents who remitted through formal channels chose money transfer operators, while 36 percent chose bank transfers. There are a number of money transfer ''Formal channels include banks and specialized money transfer systems, and the postal system. Banks and money transfer systems form the financial system infrastructure in the recipient country. The postal system in Tajikistan supports the financial system infrastructure, although it does not provide savings or other payments services. Non-bank financial institutions do not play any role in this segment, given the nature o f their business and their chronic underdevelopment. Microfinance institutions, although relatively developed and active in Tajikistan, do not yet intermediate in the remittances segment, due to their limitation from providing basic financial services and their inability to access the payments system. z8 Informal channels include personal carrier of cash through the borders by self or friends, or pool their earnings and transfer them through trustworthy travelers, who in turn take care of the delivery to workers' and migrants' relatives. T h i s i s quite common in Tajikistan, mainly due to the low density o f the average remittance and money transfer per migrant. Pooling in fact helps abate transaction costs, as some remittance service providers, and especially banks, charge flat fees irrespective of the amounts sent. 29 Pooling in fact helps abate transaction costs, as some remittance service providers, and especially banks, charge flat fees irrespective of the amounts. The same individuals organize the distribution o f remittances to migrant relatives in the respective recipient country. 11 systems operating in Tajikistan and they all integrate into the banking system through partnerships and contract agreements. They operate from the premises of commercial banks and share fees almost equally. Most transfers through money transfer systems are made on a cash-to- cash basis. Competition among the money transfer systems helps drive down the cost of transfers. (Details on transfer fees and speed of these systems operating in Tajikistan are described in Annex 3.) B y contrast, inter-bank money transfers (account-to-account) are infrequently used among migrants because of difficulties in opening bank accounts and foreign exchange regulations limiting the amount that can be transferred. The Zerkalo survey (2005) indicates that only 18 percent of migrants have bank accounts because their illegal status limits access to the banking system in host countries. Transfers through post offices are insignificant due to high cost (10-12 percent) and slow speed (2-3 weeks) despite their extended outreach into rural areas. (Annex 2 provides details on transfer channels and costs for remitting funds by Tajik migrants.) 28. Informal channels for transferring remittances include self-carrier, friends and acquaintancesdue to the zero cost of transfer or their illegal status. For example, illegal Tajik migrants residing in Russia do not leave their residence to go to commercial banks to transfer funds for fear of harassment by Russian police and deportation. Instead, they send money to friends or acquaintances that in turn will carry or transfer funds and distribute them to their families back home. According to the survey, about 26 percent of migrants sent money to their families inTajikistan through friends. 29. In addition, anecdotal evidence shows that some banks have formalized business relationships with a number of Tajik individuals permanently residing in Russia involved in providing labor opportunities for seasonal and shuttle migrants. These individuals are generally trustworthy and well known by Russian employers and Tajik workers. Together with providing job opportunities, they also provide remittance services by sending- receiving- distributing remittances through their bank accounts in both Russia and Tajikistan. They work against a commission, the value of which remains unknown. The benefits of these business relationships are significant: migrants find temporary jobs and a secure and accessible way to send money to their relatives all the while addressing cultural and language barriers when it comes to dealing with Russian money transfer systems and banks. Tajik banks build these relationships in order to expand the client base, profit from remittance commissions, and increase the average value of remittances and thus decrease the transaction cost for the bank. This i s because banks and money transfer commissions and fees are in most cases based on a percentage of the transferred amount, irrespective of the value of the transaction. As a result, the higher the value of the remittance, the less paper work and fixed costs for the bank and the money transfer system. Box3: Cross-sellingby Tajik CommercialBanks Some commercial banks have slowly begun to sell financial products to migrants. In general, Tajik banks are neither particularly focused on, nor capable of, servicing remittances. A few banks with a more extensive branch network in the country (AIB, Orion Bank, Tajiksodirot and Eskhata Bank) have just begun targeting migrants and remittances as an important element of their business. Two Tajik banks have so far established specialized units to provide a range of banking services other than simple savings to migrant clients and their relatives (including small loans). 12 30. Tajikistan'sformal transfer systemfor remittances is efl~ient,~'especially through the money transfer operators. The country has been successful in reducing cost by removing a transfer tax and creating an enabling environment that promotes competition among banks in partnership with the money transfer systems. Further, the exchange rate spread i s insignificant due to no restrictions on cash withdrawal in foreign currencies. The analysis of financial infrastructure, costs, speed, and outreach discussed in Annex I1 reveals that money transfer operators offer highspeed services at a relatively low cost.` Senders are not required to have bank accounts and only need to show an ID card if they use cash-to-cash or cash-to-account transfers. The cost of transfer that i s paid entirely by remitters in sending countries to money transfer operators i s relatively low (between 1.5 percent and 5 percent of the remitted amount), compared to other countries.31 31. Disbursement of remittances mostly takes place in Dushanbe at locations where commercial banks operate. The low density of the rural economy, weak communication infrastructure, and low profit for retail banking in the regions and rural areas limits the outreach of banks. Two commercial banks, A B and Amonat, have the largest branch networks in the country, with about 80 branches each. Amonat bank, in particular, plays a critical role in delivering fiscal and quasi-fiscal payments in the rural areas (almost on a monopolistic basis); however, its technology i s cumbersome. Only 10 percent of the branches are connected to headquarters, which makes it impossible to consolidate bank accounts on a daily basis. Most transfers are made on a cash-to-cash basis becauseof, inpart, a low level of thrust in the banking system due to prior unstable macro-economic environment, the effects of banks failures and closures and the three mandatory monetary conversions during the last decade, which sharply eroded savings (Box 4). Box 4: Tajikistan's Banking System The level of development o f the Tajik banking sector i s relatively low compared to other countries in the FSUand the CIS. I t is currently undergoing a profound stabilization and consolidation process that has resulted in the exit of insolvent banks and the divestiture of KreditInvest from AIB bank. Consequently, asset quality, the overall capitalization o f banks, and the capital adequacy ratio have improved, and non-performing loans have declined. Although the banking sector i s growing rapidly in terms of credit to the private sector, it still faces a lack of long-term financing, an inadequate deposit base, and low capacity to intermediatetheir capital base. The banking sector includes 14 banks, with only one bank fully owned by the state, Amonat Bank (former savings bank). There i s only one foreign commercial bank operating in the country, and one microfinance bank fully owned by the Aga Khan Foundation(established in 2003). The sector remains concentrated, with four large banks (Orion Bank, AIB, Tajiksodirot and Amonat Bank) controlling 70 percent of assets, 80 percent of household deposits and 70 percent o f credits to the economy. The economy i s highly dollarized and the inter-bank market i s not yet developed to acceptable standards, mainly due to weakness in monetary policy management, lack of a government securities market, high degree of dollarization and lack of trust among bankers. 30 An efficient transfer service is characterizedby low transfer costs, increasedspeed, and better access. 31 Bank charges in the Philippines, exclusive of the foreign exchange spread and recipient charge, range from about US$5 for a transfer from a foreign account to a branch account in a single bank, to about US$16 which may include courier services (Kevin Mellyn, "Worker Remittances as a Development Tool Opportunity for the Philippines, p.12 (June, 2003). The average cost of transfer money from the US. to Mexico was 7.3 percentinFebruary2004. 13 USESOF REMITTANCES TAJIKISTANIN 32. According to the survey of Figure4: TajikistanUsesof Remittances migrants by Zerkalo, most recipients used (% of Respondents) funds for purposes of consumption, education, health, home repair and savings. Nearly one-half of respondents (47 percent) indicated that they used remittances for cons~mption,~~22 percent used remittances for education and health care, 15 percent for home repair, and 10percent for saving. These findings show that about 37 percent of respondents , . . uscd remittances for personal investment 0% 5% 10% 15% E3% 25% 30% 35% 4% 45% 50% (human capital and home repair), while only 3 percent of respondents used remittances for income generation activities and asset accumulation. Unfortunately, further insight i s limited as information on the share of these expenditures in total remittance use i s not available from the survey. 33. The level of income dictates the type of spending. Given a low level of income of about US$131 per month (based on estimates of remittances a migrant transfers to his family in Tajikistan), a Tajik household with 5 dependents will likely spend the majority of their income- including remittances-on priority expenditures (consumption and other basic needs). For this reason, it i s unlikely that any significant amount of funds will be left over for productive investment. However, an increase in income after several years of working abroad tends to lower the marginal propensity to consume, thus increasing the scope for recipients to use remittances for productive investment, and more so given a favorable business environment. This includes income generation activities in rural areas, small and medium businesses, and bank deposits that will be invested indirectly. 34. However, migrants' ability to save and invest is limited by severalfactors. Bank saving - deposits are discouraged due to a lack of trust in the banking system33and a low level of interest rate. As remittances reach Tajikistan, they are mostly withdrawn in cash in various currencies (U.S. Dollar, Russian Ruble, Euro, and Somoni) and typically kept at home "under the mattress" or used to accumulate assets (car, house, and land). The NBT survey indicated that 82 percent of respondents were not ready to deposit remitted funds in the banking system and 62 percent said that they needed money for current household consumption and business transaction^.^^ Only small amounts are converted into savings inthe banking system; however, deposits in commercial banks has increased in recent years (deposits by individuals doubled from 1percent in 2002 to 2 percent of GDP in 2005 and deposits by legal entity and individuals increased from 4 percent in 2002 to 7 percent of GDP in 2005). 32Consumption is defined as the sum of expenditures on food and clothing, purchases of goods, social activities, charity, and paying off debts. 33This i s due to the prior unstable macro-economicenvironment, the effects of bank failures and closures and the three mandatory monetary conversions during the last decade. All this resultedin people's savings being sharply devalued and eroded. 34IMF,"Labor Migrationand Remittances" inTajikistan--Selected Issues and Statistical Appendix, March, 2005. 14 35. Similarly, an unfavorable business environment in Tajikistan hinders the use of remittancesfor investment, specifically, for income generation activities in rural areas andfor small businesses in urban areas. According to an IFC survey of the business environment in Tajikistan (2003), the majority of SMEs found that administrative procedures are overly complex (such as access to external financing, foreign trade operations, taxation, inspection, and licensing procedure^).^^ The regulatory system i s non-transparent, application of the tax code i s uneven, licensing, inspection, standardization and certification are excessive, and harassment and corruption are frequent. Although the Business Environment Enterprise Performance Survey (BEEPS) in 2005 shows that the overall environment for doing business in Tajikistan has improved (compared to 2002); there remain serious problems regarding access to land, title or leasing of land, business licensing and permits, labor regulations, skill and education of workers, judiciary functions, and street crime, thefts and disorders. In relation to CIS countries, doing business inTajikistan i s worse off in several areas includingjudiciary function, business licensing and permits, tax administration, transport and electricity. The access to landproblem discourages remittance recipients from using funds for income generation activities in the agriculture sector. Similarly, the businesslicensing, permits and inspection problems limit new business entries, thus limitingthe use of remittances to a few choices includingaccumulating personal wealth and home repairs. The Government has recently amended the law on licensing that significantly reduced the number of activities subject to licensing, and adopted a law on micro-finance to support access of SMEs to finance; however, much remains to be done to further improve the business environment. IMPACTS OF REMITTANCES INTAJIKISTAN 36. The extent to which remittances will have a maximum impact on the economy depends largely on how they are used. It i s widely recognized that remittances provide a source of foreign exchange for financing the current account deficit, which i s associated with increased income of recipient households fueling domestic consumption and imports. The long-term developmental impact of remittances, however, depends on the type of remittance spending, Le., consumption versus investment. This i s important because remittances can help reduce poverty and fuel high rates of household savings and investment. If remittance income i s spent on consumption, the impact will be short-term through a spill-over effect transmitted through a multiplier. Consumption spending raises the effective demand for local goods and services and fuels imports. Future consumption will therefore depend on future remittances. On the other hand, if remittances are invested directly or indirectly (through bank deposits) for the development of local infrastructure, income generation activities and small and medium businesses, they will increase capital and generate a future income stream to finance future consumption. This will contribute to long-term economic growth3' and reduce the vulnerability of recipients to a sharp falloff inremittances and to potential remittance decay. 35 These recommendations call for: alternative financing mechanisms for SMEs, reform o f procedures and processes related to foreign trade operations, simplification of the tax regime for SMEs, reduction of inspections and the number of activities that require licensing, streamlined standardization and certification, and unification o f registration procedures for all businesses,p.12-14, (IFC, 2004). 36 Empirical evidence from 13 Caribbean countries shows that a 1 percentage point increase in remittance inflows increased private investment by 0.6 percent points as a share of GDP (Mishra, 2005). 37 The investment multiplier i s estimated to increase the amount initially spent by three times. See Norbert Bielefeld and Antonique Koning, "International Remittances: Delivering Fair Value," in Remittances Develomnent Impact and Future Prospects, edited by Samuel Munzele Mambo and Dilip Ratha, The World Bank, 2005, p.195. 15 37. Remittances have increased national income and accelerated consumption in Tajikistan. As a share of GDP, remittances increased from 0.1 percent in 2000 to 12.2 percent in 2004, and over the same period contributed to the country's 58 percent increase in real GDP by accelerating consumption. As a result, the share of the service sector (construction, trade, informal sector activities) in GDP increased from 49 percent in 2000 to 59 percent of GDP in 2004. Tajikistan's gross national disposable income per capita was 15 percent higher than per capita GDP in2004 and is growing at a faster rate.38 38. Figure 5 illustrates that remittances Figure 5: Tajikistan Impact of Remittances - have enabled households to meet their on Living Standards nutrition needs. Clearly, remittances have enhanced welfare and raised living standards. b The survey indicates that the number of Nomvam respondents reporting that they find it hard to P provide foodstuffs significantly declined after DVBblE6rnrnllWee m- Fm4utilitiesc h t k bd migrating. A large number of respondents Wlduables Food bdmlulili~%6 reporting that they could afford food and cbw Hardlopm\ndeloodrlul~ utilities (but not durables) significantly increased after migration to 38 percent, Pmml 0 10 20 30 40 50 compared to 16percent before migration. 1BPdlgratrn" .h.rmgatonI 34. Remittances have increased purchasing power of thepoor `undprovided a safety netfor their families. Remittances accounted for a larger share of income for the poorest households. According to the World Bank Poverty Assessment Update, remittances constituted 25 percent of income for the poorest quintile, compared to 9 percent of income for the richest quintile in 2003.39 The majority of remittances sent to Tajikistan go to the rural areas (almost 70 percent) where the poverty rate i s high, while in Armenia and Georgia the situation i s reversed-almost 70percent o f remittances go to metropolitan and other cities.40 Table 7 shows that poorer households' purchasing power i s heavily tied to migration-generated income. Remittances per capita of the poorest quintile were only 42 percent of that of the richest quintile, while the share of remittances in consumption expenditure for the poorest quintile doubled that of the richest quintile. In the medium term, Tajikistan's ability to sustain and accelerate its macroeconomic growth and acceIerate poverty reduction will be strongly linked to the sustainability of remittances. Table 7: Annual Consumption and Remittances per capita by Quintiles Quintile 1 2 3 4 5 Consumption per capita (all households) $67.20 $103.88 $139.03 $188.13 $344.35 Shareof receiving households 8.01% 9.82% 9.33% 8.96% 7.66% Remittances per capita (receiving 23.56 28.12 34.25 41.85 55.68 households) Remittances/Consumption(receiving 35.07% 27.07% 24.63% 22.25% 16.17% households) Source: The World Bank, "Enhancing Gains through International Labor Migration in Europe and Central Asia Region," (2005). 38 International Monetary Fund (2005), Republic of Tajikistan: Selected Issues and Statistical Appendix, IMF Washington, DC, IMFCountry Report: 05/131, April. 39 The World Bank, PovertyAssessment Update (2005). 40The World Bank, "Enhancing Gains throughInternationalLabor Migration inEasternEuropeand the Former Soviet Union," 2005. 16 growth, if not properly managed. Remittance inflows can produce its own version of "Dutch 40. An increased flow of remittances can undermine competitiveness and long-term disease" by contributing to an appreciation of the local currency41 and thus international Further, while reducing pressure on the domestic labor market in the short- term, continued outflows of educated and skilled labor forces will adversely affect the domestic labor market over the long-term. On one hand, migration reduces the supply of skilled labor and human capital and thus increases wages. This increases the cost of production and erodes competitiveness of the country in world markets. On the other hand, increased remittance inflows to recipient households could reduce output through lowered work effort. The Zerkalo survey of migrants indicated that the number of respondentsreporting that they were unemployed increased to 44 percent after migration compared to 22 percent before migration; however, it is not clear if this higher rate of unemployment i s due to reduced work effort or lack of suitable employment. The impact of migration and remittances on Tajikistan's labor productivity, .work effort, and its international competitiveness should be considered as an areafor further analysis.43 growth. These risks include unfavorable changes inimmigrationpolicy and changes inthe business cycle in host countries. Nationalism and Others animosity towards migrants in Russia could put Aglcullwe political pressure on the Russian girvernment to Transporlalmn Prafesslonal implement restrictive measures towards illegal M~nlnflmf~rliltles Wholesale/Relad Tajik migrants. This would discourage temporary Conwrwllon migration, and thereby lower the number of 0 0 6 1006 200% 300% 400% 500% 600% 700% migrants to Russia. Secondly, although ~ 41 Amuedo-Dorantes and Pozo (2004) showed that doubling of workers' remittancesin 13 LAC countries resulted in a 22 percentreal exchange rate appreciation. 42 According to empirical evidence conducted by Rajan and Subramanian (2005), there i s no evidence that remittance flows slow down growth by affecting competitiveness. 43The analysis of unit labor cost in 2002 indicated that Tajikistan remained competitive among the CIS despite low labor productivity. This was mainly due to an extremely low level of wage. Low productivity can to an important extent be ascribed to the high cost of doing business in Tajikistan. See the World Bank (2005), "Trade Diagnostic Study," Washington, D.C. 44The counter cyclical are person to person flows targeted to meet the needs of recipients. The World Bank (2005) "Global Economic Prospects: International Remittances and Migration," Washington, DC. 17 estimate45based on the data from the 2005 survey of migrants suggested that every one percent increase in income increasesremittance levels by 0.35 percent, and vice versa. 42. Tajikistan needs to mitigate these risks by enhancing the use of surplus remittance income, after meeting basic subsistence needs, for development purposes as well as encouraging private sector led development to create jobs at home. Policies to promote these activities will make a positive contribution to increased productivity and improved competitiveness, and will help generate domestic employment and output over the long-run. Over time, economic growth in labor sending countries will deter migration (as confirmed by empirical evidence from South and South East Asia to the Middle East). One study found that a one percent increase in GDP per capita at home i s estimated to lower migration per capita by nearly two Government policies that support private sector led development will likely generate employment at home and therefore reduce workers' incentives to emigrate temporarily. Inother words, remittances partly reflect the failure of development. 43. The foregoing discussion of remittance use in Tajikistan i s not a unique phenomenon; most labor sending countries around the world have encountered similar experiences. Countries with a long history of international migration over several decades have sought to develop and implement various policies to influence the choice of spending between consumption and investment. The following section reviews the international literature to suggest potential ways to enhance the development impact of remittances and assesses the potential of replication in the case of Tajikistan. The review draws some lessons learned that can be used in the design of specific policy measuresfor Tajikistan. V. ENHANCINGTHE DEVELOPMENTALIMPACT OFREMITTANCES- LESSONSLEARNEDFROMINTERNATIONALEXPERIENCE 44. Over the past few decades, several labor exporting countries have attempted to maximize the benefit of remittances by implementing various policy instruments to improve incentives to use remittances for productive investment. These policy instruments, tailored to specific circumstances and characteristics of migrants in each country, have been used to mobilize remittances from individuals and collective groups. The ultimate goal of these instruments i s to maintain remittances in the banking system so that they can be channeled more effectively to fund productive projects and to encourage migrants to use remittances for investment. This section describes the key features of these experiences and reflects on the applicability to Tajikistanof policy instruments that have been adoptedelsewhere. MEASURES TO MOBILIZE REMITTANCES INTO THE BANKINGSYSTEM 45. Various types of policy instruments have been implemented by labor sending countries in an effort to bring greater amounts of remittances into the formal banking system in recipient countries. As part of an incentive approach, some countries sought to introduce repatriate foreign currency accounts and foreign currency bonds (South and South East Asia region). The mandatory requirement approach was implemented by South East Asian countries. The Note considers thesein more detail below. 45 R= 26.90 + 0.35 Y, where R = represents monthly remittances, Y = monthly income. T-statistics for both parameters are statistically significant at 1percent confidence interval, andAdjust R-squarei s 0.44. 46Robert Lucas, International Migration Regime and Economic Development, p.41, 18 46. Repatriate Foreign Currency Accounts and Foreign Currency Bonds.47 Due to restrictions on foreign capital movement, foreign currency accounts were allowed by several labor exporting countries48to target skilled migrants with retained savings in overseas personal saving accounts. Foreign currency accounts49provided various incentives including exemptions from foreign exchange regulations when repatriated, a premium over the international interest rate (Eurodollar deposit with similar terms), and tax-free earned interest income to attract informal remittances from individual migrants into domestic banks. Foreign currency denominated bonds - providing incentives similar to that of repatriating foreign currency accounts - have beenintroduced inseveral countries50to attract remittances into the formal banking system for financing of public investment programs. They targeted members of the Diaspora, and these were considered more effective than foreign currency accounts in attracting unrecorded remittances becauseof the anonymity provided. 47. The foreign currency accounts and bonds can not be replicated in Tajikistan. Generally, the incentive schemes that aim to attract remittances into the formal banking system have had mixed experience as they do not address factors causing migrants to transfer remittances through informal channels to home countries. These factors include foreign exchange restriction, overvaluation of currency, and a low level of financial intermediati~n.~~Additionally, a potential downside to certain tax incentives are the opportunities they create for tax evasion. Inthe case of Tajikistan, the incentives are not appealing because the capital account i s relatively liberalized. The central bank has recently relaxed threshold for exporting foreign currency without permission to facilitate international trade. The economy i s somewhat dollarized as fund transfers and deposit accounts can be disbursed and denominated in any foreign currency. Further, the majority of migrants do not have bank accounts and lack knowledge of financial assets. Finally, a weak banking system including imprudent management of commercial banks can counteract the gains from mobilizing remittances into the banking system. This could further erode migrants' confidence intransferring remittances through the formal channel. 48. Mandatory Req~irernent.~~the 1980s, the Korean government imposed a mandatory In requirement such that Korean workers must remit 80 percent of earnings through the banking system. This experience was considered successful because of its unique labor export process by which Korean contractors abroad employed mostly Korean workers and their collaboration with state companies. Additionally, the Philippines and China also required workers to remit a significant percentage of their earnings. It should be noted that the mandatory remittance approach i s effective only if labor flows are organized through bilateral official agreements and a government has access and control over the migrant's earnings. Further, forced remittances can be detrimental as migrants likely under-report their income and channel remittances through informal channels. 47 Shivani Puri and Tineke Ritzema, "Migrant Worker Remittances, Micro-Finance and the Informal Economy: Prospects and Issues," Working Paper No.21, Social Finance Unit, International Labor Office, Geneva. 48 India, Pakistan, Sn Lanka, Korea, the Philippines, and Thailand 49 By India and Pakistan 50 by Pakistan,BangladeshandIndia. 5 1Shivani Pun and Tineke Ritzema, "Worker Migrant, Remittances, Micro-finance and the Informal Economy - Prospects and Issues, Working Paper #21, Enterprisesand Cooperation Development Department, Social Finance Unit, p.22. s2 As cited by Puri and Ritzeman (2001) and Carling (2005), other countries that have implemented similar policies (Egypt, Pakistan, Bangladesh,Philippines andThailand) have had a lesser degree of success. 19 49. Direct Government Involvements. Governments can address obstacles to remitting funds through the banking system in both origin and host countries through direct interventions. With a large number of unskilled migrants who are un-banked, the Mexican government pursued diplomatic dialog that led to close collaboration with the U S government and commercial banks. Consequently, a Mexican migrant can open a bank account with a commercial bank in the U.S. A migrant only needs to display a "Mexican Identification Card" to open a deposit account without disclosing their immigration status to a U.S. commercial bank. Consequently, transfer of remittances through the banking system to Mexico has increased significantly. The Philippines developed a successful "pre-departure orientation service" that disseminated information about savings and investment options to encourage migrants to open deposit accounts and use financial products inthe Philippines. 50. Only selected aspects of this experience can be replicated in Tajikistan. The "Mexican Identification Card" approach would require close collaboration between commercial banks and the governments in Tajikistan and Russia and would take time to achieve results. Enforcement of the migration law inside Russia i s more hostile towards illegal migrants despite its free visa regime with the CIS countries due to the anti-immigrant attitudes that prevail in some segments of the population. However, a similar program to the pre-departure orientation service in the Philippines has been initiated in Tajikistan by the International Labor Organization (IOM) through the project entitled "Awareness Campaign and Assistance for Migrants". The project established a network of "Information Resource Centers" for labor migrants in Dushanbe. The center ,lisseminates information on migrants' rights, migration laws, and remittance transfers. It also provides consultation to migrants (mobile and on-line). USINGMICRO-FINANCE INSTITUTIONSTO MOBILIZE REMITTANCES 51. The insufficient banking network in Tajikistan provides scope for penetration of micro-finance institutions (MFZs) to deliver remittances to recipients and mobilize savingsfor development purposes. Due to the higher administrative costs associated with maintaining the typical small account held by migrants, commercial banks do not have an extensive retail banking network. This provides an opportunity for MFIs to play role in delivering remittances, capturing informal remittances that are kept outside the banking system for productive investment, and providing lending backed by future remittance flows to migrants' families.53 Empirical evidence in the U.S. and Latin America indicates that improving remittance intermediation can have a dramatic impact on the living conditions of transnational migrant families, and result in sustainable and equitable development of communities in both labor sending and receiving countries.54 In El Salvador, FEDECACES has been working with its affiliates to strengthen capacity to receive international remittances. As a result, the number of transfers received increased significantly from 3,000 in 2001 to more than 100,000 transactions in 2004. Financiera Calpia processes thousands of remittances monthly after it launched a concerted market campaign. It also offers other services including savings accounts, mortgages, and insurance products. In the Philippines, the government supports micro-finance banks and non- governmental organizations that have been active in mobilizing savings from migrants' families and offering financial services such as lending products to migrants in rural areas. In Mexico, National Financiera (NAFIN), established in 1994, provides migrants technical and financial 53 Pedro de Vasconcelos, "The Multilateral Investment Fund of the Inter-American Development Bank and Remittances," FourthCoordination Meeting on International Migration, UnitedNations, October2005. 54Raul Hinojosa Ojeda, "Transnational migration, remittances, and development in North America: Globalization Lessons from the Oxaca California Transnational Village/Community Modeling Project," (NAID Center, UCLA), October 2003. 20 support to channel remittances for productive uses, reduces intermediaries to the minimumlevel, and offers support to women, youth, and elderly. I Box 5: Benefits of Involvement of MFIsin Remittances Having the MFIs involved in the remittance market i s desirable for several reasons. First, MFIs can make remittance transfers more attractive to recipients due to their outreach and accessibility that surpass that of commercial banks. Secondly, remittance transfer as a fee-based product could provide revenue to MFIs and thus help them achieve financial sustainability without compromising the objectives of poverty orientation and impact maximization. Thirdly, MFIs provision of services to non-members receiving remittances could expand the client base and reduce the number o f un-banked households as they later become MFImembers. Finally, auxiliary training and marketing services offered by MFIs can enhance the capacity of return migrants or recipients to use remittances for productive investment in origin countries. Already, there are MFIs that operate in remittance markets in several countries. In the ECA region, microfinance banks in Kosovo, Serbia, and Ukraine registered as commercial banks are already offering remittance services. Source: Jorgen Carling. 2005. "Migrant Remittances and Development Cooperation," PRIO Report Y2005, International Peace ResearchInstitute, Oslo (PRIO). 52. However, it is likely that the role of MFIs in delivering remittances in Tajikistan will remain limited over the medium-term. The legislative and regulatory frameworks for MFIs have only recently been developed and international experience suggests that the maturity of this industry will take time. The industry i s still going through consolidation, capitalization, and capacity improvements. The majority of MFIs emerged from donor-financed organizations, have limited capacity infinancial and portfolio management, and have no capital to meet the minimum requirement as required by the micro-finance law. These factors likely limit the role of micro- finance institutions in delivering remittances in the near future due to concerns about potential security risks to a banking system and to delivery of remittance to recipients. Alternatively, the role of MFIs in the provisioning of financial services in Tajikistan, such as credits to close income-consumption gaps and to support income generating activities in rural and urban areas, can be foreseenin the near future. MEASURES TO MOBILIZE COLLECTIVEREMITTANCES 53. The enormous size of remittances has drawn the attention of governments, commercial banks, and non-governmental organizations intent on capturing collective funds and channeling them to productive investments. Some commercial banks have raised funds from the international financial market by securitizing remittances flows, and many non-governmental organizations have mobilized remittances to meet the development objectives of local communities. 54. Securitization of Remittances. Emerging markets with a large volume of remittances and deep financial markets have designed financial instruments that can attract a large level of remittances for use by the private sector. Commercial banks in emerging markets (Turkey, Mexico, and Brazil) with developed financial systems have raised hard currency for their long- term capital (5-10 years) through securitization of future remittance receivables. Securitization of remittances involves the borrowing entity (such as a bank) pledging its future remittance receivables to an offshore Special PurposeVehicle (SPV). The SPV issues the debt. Designated correspondent banks are directed to channel all remittance flows of the borrowing bank to an offshore collection account managed by a trustee. The trustee makes principal and interest payments to the investors and sends the excess collection to the borrowing bank. Remittances do 21 not enter the issuer's home countries; therefore, the rating agency believes that it mitigates the usual sovereign and convertibility risks (Figure 7).55 I I I I I I I I I Remittancepayments Issuingbankcredits I I I I (foreign currency) beneficiary'saccount I I in domesticcurrency. I I I I I I Correspondent Message I I banks Issuingba I I I I I I I I I I I I Remittance flows (foreign currency) I I I I 1 I I Excess cash I I I I I I b . - - L , - - - - - - - , - - - - - - - ~ I I I I Trustee collateral Issuingbanks' accoi nt I I I I (foreign currency) U L Lqnmkdnmk I I I I I I I I I I I I I I I III Offshord Domestic I I I I I I I I 55. The largest remittance notes (as of June 30, 2005) were placed by Turkey's Akbank Remittances Trust Securitization, Ltd (A.R.T.S.) in the amount of US$2.525 billion in sixteen tranches. The latest issuance of US$l billion (tranches 10through 16) represents the largest-ever single issuance of this asset class in emerging markets, The underlying collateral supporting the notes is composed of remittances generated by Akbank56 T.A.S. through its specified correspondent banks. Mexico i s second in the world's remittance securitization market- averaging 24 percent of the worldwide amount raised through such bonds (1994-2004).57 Mexican banks were some of the first in the world to offer bonds secured by future remittance flows in the mid-1990s. Banco de Brasil issued 5-year bonds worth US$300 million usingfuture yen remittances from Brazilian workers inJapan as collateral in 2001. 56. There is little scopefor securitization of future remittance receivables in Tajikistandue to weakness in thefinancial sector. The legal and regulatory structure of banking institutions in Tajikistan i s still developing. The financial market i s shallow and the sector remains underdeveloped. Further, commercial banks are small and have limited experience in issuing advanced debt instruments. Frequent violations of prudential regulations and a relatively high ratio of non-performing loans indicate structural management weaknesses. Finally, local banks do not have alternative sources of liquidity to ensure timely cash deliveries to remittance 55 The World Bank, "Global Economic Prospects, 2006." 56 In November 2004, Isbank, another large financial institution, completed a US$600 million securitization transaction based on diversified payment rights. Standard Chartered Bank managed the transaction. A special purpose vehicle in three tranches issuedthe notes with maturity dates ranging from 7 to 10years. This deal provided the longest maturity offered by a Turkishbank to date. 57 Ratha, Dilip, "Securitization of Future Remittance Flows: A Global Overview," presentation at the International Forumon Remittances,Inter-American Development Bank, June 28,2005. 22 beneficiaries. Their activities, besides transfer services, are limited to financing of short-term trade, often to bank insiders, and to speculation in the foreign exchange markets. The limited sources of liquidity could delay payments of remittances, and thus negatively affect public confidence inthe banking system. 57. Home Town Associations (HTAs) and Matching Funds. HTAs are established in host countries to raise funds from diasporas and have them transferred to their home towns for development purposes. Due to strong ties to their communities in a host country-most commonly the United States and Canada-migrant diasporas (often first-generation) from a particular town or region in Latin America have joined forces to establish HTAs. These associations raise funds in host countries through various activities and remit them to communities in their home towns. The funds are used for construction, basic infrastructure, and other development projects. 58. To attract funds for specific development projects, the governments (central and local) of various Latin American countries, notably Mexico and El Salvador, have worked closely with HTAs by providing "matching funds" to complement their donation. A widely known matching fund, the Three-for-One programs introduced by several states in Mexico (started in 1997), have been successful in attracting funds for specific investment needs. For every dollar remitted, three . dollars are donated to HTA projects by federal, state and municipal governments (Le., the three levels of government each contribute a dollar for each dollar raised by HTAs). B y 2002, the programs had established projects totaling US$43.5 million. These projects include road and school construction, healthcare, utility and agriculture improvements and local investments. Through close collaboration between HTAs and the government, the matching funds have been considered effective inusingresources for specialized local needs. 59. The matchingfund was piloted in Tajikistan by UNDP in collaboration with IOM. In 2005, a project "Enhancing the Development Impact of Migrant Remittances in Rural Communities" was piloted in two jamoats to enhance the use of remittances. The I O M has worked closely with the Jamoat Development Committee (JDC) and Jamoat Resource Center (JRC), a community based organization developed by UNDP. The project has two components. The first component provides micro credits to return labor migrants and migrant households to capitalize and promote the investment of migrant remittances for the development of viable livelihoods for migrant families affected by emigration. Under the first component, the JRC selected participants from migrants' households and assessed their capacity based on the availability of their own investment resources (savings and non-cash). Selected participants would be trained on business start-up including development of business plans, finance, and business management. Subsequently, they would be eligible for micro credits to match savings to finance agriculture, livestock, and small trade projects. The amount of credit can be as high as 50 percent of total investment. The terms of the credit are: interest of 1 percent per month, 10 month maturity, and repayment at the end of maturity. The second component provided participative financing facilities (matching funds) for community initiatives and infrastructure by matching grant to migrants' contributions5* for implementation of local infrastructure projects that were selected and prioritized by the JRC. These projects included rehabilitation of schools, bridges, a water reservoir, an electricity transformer, etc. While developments to date seem encouraging, it i s too early to evaluate results of this initiative. ______~ 58 Total value of these projects constituted US$8,798, of which US$3,913 was contributed by labor migrantsand the remainingwas fundedby UNDPAOM project. 23 LESSONS LEARNED 60. The international experience shows that there is no single answer regarding the right policy that a country could pursue to enhance the development impact of remittances. Each of the policy measures discussed in the foregoing section has been designed and implemented to address issues specific to migrants' characteristics and the macroeconomic environment in each country. This highlights the importance of having reliable information on migrants overseas and their behavior related to the sending and using of remittances. Further, there has been no clear empirical evidence on the eventual development outcomes and impacts of these policies. 61. Nonetheless, instruments that aim to improve access of migrants to the banking system in both origin and host countries have been effective in steadily increasing transfers of remittances through the banking system. Interventions by the governments of Mexico and the Philippines have increased transfers of remittances through the banking systems inboth countries. With the active involvement of labor sending governments in conducting intensive negotiations with their counterparts, migrants are now able to open bank accounts in certain host countries (Mexico and the US). An information campaign, supported by the Philippines government, provides information on savings and investment options to migrants. 62. Mobilization of collective remittances has been limited due to country specific conditions. For example, remittance securitization requires a large inflow of remittances through a well developed banking system (account to account basis) and an adequate legal structure of institutions in the issuing countries. The amount of wire transfers that are used mostly for securitization is insignificant in Tajikistan. The HTAs and matching funds require strong diaspora networks that can play a role in mobilizing remittances in host countries and strong local community networks in recipient countries that can manage and monitor the use of remittances and matching funds. Further, the use of matching funds requires good institutional capacity of local governments inprioritizing development projects to ensure that scarce public funds are used effectively, 63. Integration of Micro-finance institutions into remittance-related services in remote and rural areas should be promoted. Integration of MFIs into remittance services will help increase access of rural households and individuals to financial services. Access to bank services i s limited (or not available) because retail banking remains unprofitable and too costly due to the low density of population in these areas and the weak communication infrastructure (see detailed discussions on micro finance organizations in Tajikistan in Annex V). Due to the MFIs' reputation among the rural population and better penetration (as a result of being associated with a large number of active organizations and outreach initiatives), they can engage in remittance activities by physically delivering remittances to migrant families and mobilizing savings sent through both formal and informal channels in support of income generation activities, consumption smoothing, and a social safety-net. New financial products need to be developed to enable MFIs to provide low-cost financial services for consumption smoothing (such as advances), income generation activities (in the form of micro credits), for personal investment (home improvement loans and or education loans), To ensure efficient intermediation of remittances through linkages to MFIs, their institutionalcapacity needs to be strengthened. While MFIs have a comparative advantage in a segment of the remittance market that is naturally excluded from the outreach of commercial banks, the capacity of MFIs in financial and portfolio management i s weak and a large number of MFIs in developing countries are not registered as commercial banks. 24 VI. ENHANCINGDEVELOPMENTIMPACT- WHAT CANTAJIKISTAN DO? 64. This section proposes some policy measures to enhance the development impact of remittances in Tajikistan. The Note proposes a series of recommendations for Tajikistan based on lessons learned from international experience and highlights areas for additional research. The proposed measures differentiate policies under the government control and those that are not. The policies under government control that aim at stimulating migrant savings and investment focus on improving the business environment and the capacities of financial intermediaries. The policies outside the control of government focus on actions of host countries, for which diplomatic negotiations would be needed. RECOMMENDATIONS 65. The note strongly encourages that the overarching priority of a strategy to enhance the development impact of remittances should be the improvement of the investment climate and strengthening operations of banking intermediaries that can mobilize savings and channel them to productive investment. Efforts to improve the treatment of Tajik migrants in host countries are also encouraged, through diplomatic negotiations. By contrast, policies pursued in very different settings such as mandatory requirements and remittance bonds are discouraged because their impact could be marginalized or adversely affected due to an unfavorable business environment and weak financial sector. By contrast, policies beyond government control should receive lower priority. OverarchingPoliciesto Enhancethe DevelopmentImpactof Remittances Improvements of BusinessEnvironment 66. Improvement of the business environment in Tajikistan will promote the use of remittances for investments. A prudent macroeconomic framework will help provide an environment conducive for investment by contributing to a stable economy with declining inflation and a stable exchange rate. Improving the overall business environment remains a necessary condition for encouraging migrants to increase levels of investment, especially in SMEs, the sector in which individual migrant households would most likely invest their remittance income. These problems hamper new entries of enterprises, increase the cost and uncertainties of doing businessinTajikistan. 67. The World Bank i s currently supporting the implementation of policies to improve the business environment in Tajikistan through a development policy credit. The International Finance Corporation Private Enterprise Partnership (IFC-PEP), Foreign Investment Advisory Service (FIAS) and the Multilateral Investment Guarantee Agency (MIGA) are providing technical assistance for implementing the new legal and regulatory framework for improving investment climate inTajikistan. StrengtheningFinancial Intermediaries 68. The NBT should accelerate ihe penetration of reputable foreign banks in Tajikistan and enforce compliance of commercial banks with prudential regulation. Penetration by reputable foreign banks can help modernize and bring new technologies to this under-developed sector as well as improve confidence in the Tajik banking system. This should increase bank deposit accounts, and as a result reduce cash disbursements. Further, foreign banks can facilitate 25 account to account remittance transfers if these banks have offices in host countries, especially Russia. Migrants can deposit their earnings into the accounts of a foreign bank in Russia, while their families can withdraw from the accounts of that foreign bank inTajikistan. This would help strengthen public confidence in the overall Tajik banking system and promote modernization and competition in the financial sector. The government has already taken important steps as evidenced by recent Parliamentary approval o f amendments to the Banking Law liberalizing the entry of foreign banks into Tajikistan and a decline in the number of banks violating prudential norms. 69. Partnerships between commercial banks and MFIs would enable remittance operations by licensed MFIs without accessing the payment system. Clients o f microfinance organizations might be allowed to send remittances through the correspondent accounts of partner banks and cashpayments through the windows of MFIs intheir villages. MFIs would receive remittances in their own accounts with partner banks, and deliver cash and/or credit to clients against their debt servicing payments to their relatives while working abroad through the same system. A partnershipbetween MFIs and banks has several advantages. First, it would broaden the product and client base and consequently the outreach as MFIs are generally more engaged in rural and remote areas of the country. Second, the level of information throughout the financial system would increase, thus enabling banks to scale down and the MFIs to receive long-term funding and other banking services from banks. Third, MFIs' credit risk policies and their cash management would be enhanced. Finally, partnerships would open opportunities for sharing financial and credit information on individual labor migrants and their relatives and families. This would help banks downscale portfolios, further develop retail banking and improve the outreach to creditworthy households in rural areas. However, successful partnerships will depend on MFIs' ability to mobilize a critical mass of money transfer transactions and a sufficiently large demand for money transfer services by micro-borrowers, especially in the rural areas. 70. Tofacilitate the association of MFIs with remittance operations, a number of actions need to be taken. First, an evaluation of the untapped demand for financing in the rural areas during the seasonal migration cycle (not just for consumption) should be conducted to inform MFIsof their potential market. Further, the legal andregulatory framework for MFIs needsto be strengthened. A number of legal and regulatory requirements should be revised to enable MFI operations in remittance related services. The minimumcapital requirements for a deposit-taking MFI should be increased from US$lOO,OOO to US$300,000 to further consolidate the MFIs to facilitate market entry. The US$20,000 maximum single loan requirement should be eliminated to enable well-capitalized and efficient MFIs to expand their portfolios. The functions of the NBT related to microfinance supervision should be developed and enhanced, and MFIs should be included in the development of the credit information bureau system. Finally, the capacity of the microfinance sector should be strengthened via seminars and workshops as well as training programs on micro-finance offered by the MFIAssociation and the banking community. 71. Commercial banks should promote cross-selling other financial products to migrants and their families to increase bank deposits and create opportunities for lending to potential investors. Commercial banks should adopt the position that labor migrants have the potential to be bankable clients and thus cater financial products to this segment. They face limited profit opportunities given the high liquidity in both domestic and foreign currencies and the low returns related to money transfers and electronic payments. But if migrants are allowed to build creditworthiness overtime due to their stable cash flows and family relations, commercial banks could tap into new profit sources by cross-selling financial products to this group. In fact, it i s more profitable to cross-sell financial products due to the significantly higher interest on small 26 and retail credits59compared to fees on money transfers and remittances. Further, cross-selling could motivate migrants to change remittances into bank deposits as they can be used as collateral for future consumer credits that can close the income-consumption gap. Banks can also use savings to finance investment loans of that would be of interest to potential investors. 72. Cross selling of financial products to migrants can lead to the development of health and life insurance markets which would enhance welfare. Cross-selling will increase penetration of financial services into migrant households and integrate them into the financial sector. This will deepen the financial market and thus facilitate the development of new financial services and products which can enhance social safety-net and welfare of the Tajik population. Such products include remittance-back mortgages and health and life insurance products that could be supported by the expected future stream of remittance inflows. Particularly, health insurance products can provide a safety net to migrants overseas and their families left behind. Finally, remittance-backed mortgages could increase opportunities of home ownership. 73. Promote greater competition could further reduce the cost of transfers charged by MTOs and by commercial banks, thus encourage migrants to remit funds through formal channels. The average cost of transfers through money transfer operators and wire systems could be further reduced. Anecdotal evidence indicates that a bank would still be profitable even if the fee for transfer via MTOs i s lowered to 1percent. Fees for wire transfers could also be reduced to encourage the frequent use of account to account or cash to account transfers. These reductions could be implemented in conjunction with the promotion of cross selling of financial products. Commercial banks would likely benefit from these measures due to the additional income gains from increased remittance transfers and from income related to cross selling of financial products. Discuss with Felipe to consider dropping it. Improving Information Dissemination and Statistics on Remittances and Migration 74. The Government should develop an effective public information campaign to increase awareness among migrants and citizens of the existing laws, regulations, and practices concerning international migration and transfer of remittances. The government should develop andor improve its public information campaign to target migrants in specific regions. The information should cover migration procedures, related laws and legislation inhost countries, their rights, the procedures related to sending money home and other resources available to migrants at home and abroad. Commercial banks should be encouraged to improve their public information campaign to further increase awareness of options for transferring funds, savings, investment, and financial services. Because of competition for funds transferred through money transfer systems, large Tajik banks have already advertised their transfer services using their partner operators, describing their fees, speed, and documents required at various locations where migrants depart Tajikistan. However, the information omits available options for saving, investment and financial services eligible to a migrant in the future once one builds creditworthiness with a commercial bank. The International Labor Organization currently operates an Information Resource Center for Labor Migrant to promote the awareness level of labor migrants who choose to go abroad both in Dushanbe and a few locations inrural areas. The information provided includes labor regulation overseas, employment abroad, the migration process, livingoverseas (especially in Russia), and training. 59Retail banking was only introduced in 2004, but since then individual deposits and small loans to individualshave increased. 27 75. The State Migration Service needs to improve its migration statistics by strengthening administrative control of migration, including passport control and collection of migration cards. This could be achieved by better coordination with the State Border control and the Tajikistan State Airline company in the distribution and collection of the migration card. The definition of migrants (especially for those with dual citizenship), as well as the length of stay abroad, should be clarified in order to distinguish between temporary and permanent migration. The International Organization for Migration is currently providing technical assistance to the State Migration Service under two programs: "Building Capacity for Migration Management Program" and "Enhancing Border Management and the Security of Travel Documents." The assistance focus on reviewing the State Migration Service mandate and associated legislation, improving border management through training and provision of equipment to checkpoints to make procedures more human and effective, enhancing travel document management, and improving data collection and data sharing. 76. The NBT can improve remittance statistics in the balance of payments through better classification of remittances and broadening the coverage of statistics. It i s recommended that the NBT uses the definition of remittances that include compensation of employees, workers' remittances, and migrants' transfers in the balance of payments. The areas for improvements include: (a) identify the length of stay that will enable the NBT to distinguish between compensation of employees and workers' remittances. Currently, the balance of payments only records migrants' transfers because of a lack of information on how long migrants stay outside Tajikistan; (b) the: threshold to distinguishremittances and export receipts should be based on the results of migrant surveys. The NBT i s currently working with the IMF to improve the methodology for estimation of remittances; and (c) the coverage of remittance statistics should be broadened to the extent possible to include remittances by various official transfer methods and by source countries and to cover ;emittances transferred through informal channels and capital transfers. 77. Finally, the State Statistical Committee should improve the household budget survey. Both the household budget survey that includes a remittance module as well as the Labor Force Survey with an internal and international migration module should be regularly conducted. They will contribute to improving statistics and help point to trends and the dynamics of remittances and migration that are required to design adequate policies. This can be supported by the World Bank STATCAP project that provides technical assistance for improving the sampling frame for remittances and for conducting Tajikistan's Living Standard Survey using the new sampling frame. PoliciesMeasures to Improve the Welfareof Tajik MigrantsAbroad `78. Tajikistan should continue bilateral diplomatic negotiations with Russia (and other host countries) to improve access of Tajik migrants to labor markets over the medium- and long- term. The international experience has shown that successful bilateral negotiations can serve as a major vehicle for low-skilled, seasonal workers in the agriculture, tourism, and construction sectors6'. Successful negotiations related to such agreements can increase certainty of access for migrants by removing or minimizing restrictions and thereby improving their conditions in host countries. A bilateral agreement between the Tajik and Russian governments, signed in October 2004, sought to reduce the number of illegal labor migrants and enhance the legal framework for guest-workers with proper documentation. Subsequently, the Russian government approved legislation aimed at increasing penalties against those who violate the rights of migrants. 6oThe WorldBank.2005. "Global Economic Prospects", Washington, D.C. 28 Nevertheless, the existing legal framework has not yet been successful in curbing either abuses or illegal migration. The Tajik Government should continue negotiations with Russia, the main market for migrants on (a) preventing the violation of Tajik migrants' rights on Russian territory, (b) simplification of registration procedures for a resident permit, and (c) easing the conditions to obtain a work permit. 79. Anecdotal evidence indicates that the rights of Tajik migrants are violated in Russia. Due to lack of information of their rights on Russian territory and their illegal status, Tajik migrants are subject to abusive treatments by Russian law enforcement officers (customs, immigration, and police) once they are in Russia. Statistics show that Tajik migrants have suffered physically inRussia (arrests, beatings) and a number of them have died. This constitutes one of the reasons that 90 percent of respondents to the survey of migrants indicated that they would return to Tajikistan. Recently, Russia and Tajikistan are discussing the establishment of a joint employment center to legalize migrants and protect their rightsin Russia. 80. The Tajik authorities should press their Russian counterparts to simplifv the registrationprocedure for obtaining a resident permit. Possession of a residence permit will minimize the risk of harassmentand deportation by Russianpolice and enable a migrant to open a bank account in Russia that requires a passport as well as a permit. Bilateral negotiations between Tajikistan and Russia should aim at simplifying the residency registration process for Tajik workers upon arrival. All foreign visitors arriving inRussia are required to register with the Department of Passport and Visa Services within 3 days after arrival. The length of stay will depend on the purpose of visit as noted on the migration card. For example, businessesas well as private visitors can stay up to 90 days. Currently, to obtain a resident permit, an emigrant needs to produce several documents: a national passport, a migration card with stamp from the customs control, the national passport of and an application written by the local host (owner of the residence where the migrant stays), a certificate of payment for utility services (by the local host), prepaidregistration fee, and medical insurance. Anecdotal evidence indicates this procedure is excessively complicated, and attempts to circumvent some of the requirements by offering informal payments to Russian officers for a residence permit are prone to risk and typically unaffordable for migrants. 81. Further reductionsin cost to employerswho contract migrant labor in Russia would be , beneficial toforeign workers and local employers. Given that the demand for labor migrants i s likely to continue increasing over time due to a decline in the working age population in Russia and the unwillingness of Russian workers to take unskilledjobs, lowering the cost of employment borne by Russian employers that hire foreign workers would increase benefits on both sides. Such costs include a fee for a work permit of about 580 rubles and employee's tax. Lowering such costs would encourage employers to increase hiringlegal migrants' workers and thus reduce the price of products paid by Russian consumers. Similarly, Tajik migrants would be able to increasetheir wages and thereby increase the value of remittances sent back home. 82. Addressing these issues would go a long way to eliminating the obstacles to remitting funds via the banking system and would thus increase the volume of funds transferred on an account-to-account basis (provided that the transfer fees charged by commercial banks are competitive compared to the fees charged by money transfer operators). The clear advantage here i s further integration of remittances into the formal financial sector, which could ultimately contribute to greater productive investment. 29 AREASFOR FURTHER STRENGTHENINGAND RESEARCH 83. This section offers recommendations to improve the future survey of migrants and the estimated statistics basedon the findings of the Policy Note. It also identifies areas where further researchi s required to support public policy related to migration and remittances. To improve the survey, the Note recommends the following. First, a sampling method that better represents the migrant population in various regions in Tajikistan and various countries of destination should be employed. (Due to a budget constraint, the past survey method was biased as it used a snowball method that resulted in a sample of labor migrants headed to Russia rather than expatriate stocks or flows to other CIS and non-CIS countries.) Secondly, the questionnaire should yield more information that can offer a better understanding of migrants and their problems; for example, obstacles encountered by migrants in the process of transferring remittances to their household, available formal information networks for labor migrants in both the sending and receiving countries, and social implications of households that receive remittances and those that do not. Finally, the quality of the interview could be improvedthrough professional training. 84. To improve the reliability of the estimated statistics related to migrants and remittances, it i s suggestedthat two additional surveys of migrants in source countries be conducted, as well as a survey of Tajik diaspora in Russia. The surveys would shed some light on remitting and saving behavior of migrants in host countries, and contribute to the design of policy measures geared to attract remittances for productive investment from both temporary and permanent migrants. The surveys could alsc! help improve the estimated number of both legal and undocumented labor migrants, and thus workers' remittances versus compensation of employees (Le., permanent versus circular migration). 85. The note recommends additional research areas where additional analyses are required for future policy formulation. First, the future research could examine the impact of migration and remittances on labor market, human capital and growth as well as the social impact. The subjects include: 0 Migration and the Labor Market. This should address the effect of emigration on the supply of labor and wages, the effect of remittances on work effort, and the long-term impact on productivity, international competitiveness, and outputs. Such analysis will inform policy makers of actions necessary to mitigate the adverse impact of remittances. Remittances, Human Capital and Growth. The use of remittances for investment in human capital, their impact on the demand for social services, especially education and health, and the contribution to improved human capital and long-term economic growth inTajikistan. The social impact of migration and remittances can provide insightsto Tajik society. These issues include the impact on gender and the viability of family units. 86. Secondly, additional research should assess the needs related to strengthening the financial infrastructure. This will contribute to an increase in the transfer of remittances and accelerating the development of micro-finance institutions so that productive investment i s enhanced.61 T h i s subject will be investigated further by the World Bank forthcoming regional report on "Financial Sector and Remittancesin Central Asia and South Caucasus". 30 87. Finally, the future research could focus on a comprehensive policy and framework for migration and cooperative arrangements with key destination countries. These could include appropriate policies and structures for Tajikistan's migration management, identification of a more flexible worker programs in Russia to respond to the seasonal and cyclical nature of migration, and the role of Russia's private sector inpersuading their government to put inplace a framework for employing legal labor migrants. 31 L ANNEX I:A SUMMARY OF MIGRANTSURVEY 1. The survey was conducted in Tajikistan by the Center for Sociological Researches (Zerkalo) during April-May 2005, jointly with the Swiss Foundation for Migration and Population Studies (SFM) Switzerland. The main purpose of the research was to study the standard of living of labor migrants before departure, changes in their lives after their work abroad, and their plans for the future. It also attempted to develop a wide representation of the legal and legislative frameworks in place concerning migration, the characteristics of migration flows and the economic and social impact of labor migration in the republic. To achieve these aims, a survey of labor migrants with 1,200 respondentswas conducted. Survey Questionnaire 2. The questionnaire for this survey has been presented by the World Bank, ECA region. A number of pilot interviews were carried out in Dushanbe to help revise the questionnaire to conditions of Tajikistan. The final version of the questionnaire has been prepared in two languages: Tajik and Russian. The questionnaire consists of several sections. The first section of the questionnaire concerned the living standards and employment experience of labor migrants before they migrated. The second block of the questionnaire concerned the ways and means by which respondents used to travel to a destination and what documents/papers they had during travel. In a separate section of the questionnaire, migrants were asked to describe the specific nature of work carried out abroad (sphere of employment, status/post, and income). The main focus of the questionnaire was a block of questions Concerning the financial and social aspects of living as a labor migrant abroad. Also, a number of questions were designed to assess the labor migrants' standardof living after returning home. SamplingMethod 3. The survey used a multistage, proportional, stratified, and random sampling of the adult population of the republic (18 years of age and older). When selecting sample regions for the purpose of carrying out the survey among labor migrants, Dushanbe city, RRS and GBAO were considered to be one region. Such a decision can be rationalized by the fact that departure of labor migrants from GBAO and RRS i s complicated due to lack of accessibility of transportation in these areas, including airports, railway stations, and bus stations. Therefore, labor migrants start their travel abroad mainly through the capital-Dushanbe. Thus, for study purposes, we selectedDushanbe, Sughd oblast and Khatlon oblast as sample regions. 4. As each of the selected regions represents a separate stratum, the next step of the sampling process was the distribution of 1,200 respondents over the three strata, taking into account the population census of the Republic of Tajikistan. The number of potential respondents 18 years of age and older was calculated for every strata, and the number of necessary interviews inproportionto the number of adult populationof every stratawas determined: Table A-1.1: Distributionof Survey Samples Number of Populationtotal Percent interviews Dushanbe(including CRRS and GBO) 1,083,796 35.0 436 Hatlon oblast 1,001,429 32.4 373 Sogdii oblast 1,008,174 32.6 391 Total 3,093,399 100.0 1,200 5. This proportional method of sampling distribution allows every adult citizen of the republic to have an equal opportunity to be sampled, which corresponds to a principle of random selection of respondents. Place of Survey 6. Places with the highest congestion of labor migrants were selected as survey points. These included airports and railway / bus stations, state agencies, private companies and NGOs dealing with labor migrant issues. For distinguishing the more convenient households, arrangements were made with the main representative of makhalli (amin, jamoat), who indicated to the interviewers the households where labor migrants lived. Selection of Respondents 7. The selection of respondents was based on the following criteria: the individual b.aving work experience; the individual having earned an income on two or more trips abroad, and the individual being eighteen years of age or older. The selection of respondents was made by the "snowball method," which allows researchers to be directed to additional labor migrants in the community through word of mouth. Methodof Interview 8. The survey was carried out by "face to face" interviews, which were administered through a standardized questionnaire. The field research took place from April 21 until May 9, 2005. Interviews in regions were carried out through regional representative offices and partner organizations of CSR (Zerkalo). `The field study involved 52 interviewers working in Dushanbe and other cities. Before launching each field study, one-day training sessions for interviewers were arranged in order to familiarize the staff with the purpose of the survey, the questionnaire, the methodology for selecting survey respondents and training in other skills necessary for the successful administration of interviews. Regional supervisors monitored the work of interviewers. Data Compilation 9. The data collected were processedusingstatistical pack SPSS 11.5 for Windows. 34 ANNEX 11: ESTIMATIONOF REMITTANCEFLOWS INTO TAJIKISTAN 1. This Annex uses the stock approach to estimate remittance inflows through the use of a database obtained from a labor migrant survey conducted by Zerkalo in 2005. The stock of migrants i s basedon various estimates in order to derive a range of estimates of remittances. Two statistics are derived from this dataset-average annual per capita remittances, and percentage of migrants remitting informally. The average annual per capita remittance is calculated two ways-first by using a self-reported average propensity to remit and average monthly income, and secondly by a self-reported figure describing monthly remittance amount. These two averages are then multiplied by estimates of the stock of Tajik migrants, which vary considerably. The second statistic-the percentage of informal remittances by migrants-is used in conjunction with the IMF's 2004 remittance figure from the balance of payments figure to produce this third estimate of flows into Tajikistan. 2. The estimate of the first statistic, average annual per capita remittances, i s based on assumptions related to migrants' remitting behavior that was not elucidated by the survey. First, it i s assumedthat those individuals who did not "send money back" on their last trip remitted all of their savings when they migrated back to Tajikistan. Unfortunately, the survey design bifurcates the 1,200 respondents into remitters and non-remitters (savers). This results in a lack of information about the choice of remittance transfer method and the banking decision of non- remitters, as well as confusion about the possible savings stock for remitting migrants. Indeed, the questionnaire was designed to elicit total saving stocks for only those migrants who did not remit. The reported savings figures, since they only apply to non-remitting migrants, are most probably returned either in a lump sum remittance before return migration, or carried over in cash or goods in kind. However, the means by which savings are transported i s left unclear by the survey design.62 3. Secondly, several hundred respondents who answered that they remitted "more than 50 percent of what [they] earned" needed to be quantified to obtain an accurate estimation. This was achieved by finding the average marginalpropensity to remit of all individuals inthis greater than 50 percent category by looking at their reported income and reported remittance data, and then usingthese numbers to produce a marginalpropensity to remit statistic. This average was 64 percent of income earned abroad and was used as their answer to the marginalpropensity to remit question. Several obvious outliers were excluded in this calculation, which included anyone claiming to have remittedmore than 100percent of what they earned. 4. Thirdly, any individual who reported that they returned from their last migration trip before 2004 were excluded from our calculation. To estimate a remittance inflow for 2004, migrants who returned in 2004 or 2005 were included. Weights (i.e., the length in months o f the most recent migration) for migrants departing their home country in 2004 and returningin 2005 were calculated by subtracting the months they were on a migration trip in 2005 from their total length of stay measured in months. These weights were then multipliedby the average reported 62 A quick calculation of the savings behavior of non-remitters reveals they saved on average US$137 per month. This figure i s nearly identical to the average monthly remittance for non-savers (Le., remitters), which was US$129. monthly income and then by the marginal propensity to remit (Method 1) or by the average monthly remittance amount (Method 2). 5. The two estimates for average annual per capita remittances seem robust as they generate the estimated values of remittances that are nearly congruous. The first estimate of the average annual per capita remittances i s generated by multiplying migrants' reported monthly income during migration by their reported length of stay in 2004, and then multiplied by the self- estimated propensity to remit. This method renders an average annual per capita remittance of US$1,373.3 in 2004. The second method, nearly identical to the first, renders an estimated average annual per capita remittance of 1,406.7. The average annual per capita remittance was calculated as the product of the reported monthly remittance (instead of self-reported propensity to remit multiplied by income), and the same length of stay weights were used to calculate this figure. The difference (US$33) is considered statistically insignificantly. Further, these two values are also comparable.to the annual per capita remittance of US$1,300 as estimated by the World Bank`s Trade Diagnostic Study. 6. The estimate of the true Table A-11.1: Estimated Remittances size of remittances, however, i s EstimatedNumbei EstimatedRemittances obscured by the unreliability of of Migrants Method#1 Method#2 the stock of Tajik migrants AveragePer CapitalAnnual RemittanceEstimate $1373.33 $1,406.74 abroad. TO derive the value O f Parsons'BilateralMatrix (Worldwide) Stock 1,094.834 $1,504 $1,540 remittance in-flows to Parsons'BilateralMatrix (CIS Couneies) Stock 696.873 $957 $980 World Bank(TDS) Stock 1.200,m $1,648 d.688 Tajikistan in2004, the two Stock 120,m $165 $169 InternationalOrganization of Migration RussianCensusEstimate estimates of the average annual Flow 500.m $687 $703 Minimum F. Hill (Brwkings) Migrants in Russia Flow 300.000 $412 $422 per capita remittances are then MaximumF.Hill (Brwkings) Migrants in Russia Flow 600,000 $824 $844 State Migration Statistics(2004) Flow 420,622 $578 $592 multiplied by the stock of migrants. The lack of reliable data on the stock of migrants leads to divergence between the estimated remittances that are generated by various values of migrant stock. The official estimated number of migrants (420,622) by the State Migration Committee produces the estimated remittance of US$578 (method 1) and US$591.7 million (method 2) in 2004. Similarly, when using the stock of migrants obtained from a bilateral matrix of migrant flows between host countries around the world and Tajikistan (1,094,83463million persons), we find both estimated remittances are very close to US$1.5 billion. 7. The estimate method that is basedon the stock of migrant approach should be interpreted with caution. Some of the estimates below are most likely irrelevant, but were included nonethelessto elucidate the range of stock and flow figures and the varying remittance estimates available from these calculations. It may also be worthwhile to consider that the sample bias inherent in the Zerkalo survey-i.e., including only labor migrants to Russia rather than expatriate stocks or flows to other CIS and non-CIS countries-may make estimates of remittance flows a more accurate application of these data than that of stocks. The self-reported values of remittances are likely underreported by respondents due to safety concerns (Le., the police or community discovering their financial position) and/or concerns relating to customs and/or tax authorities. The quality of responses i s likely poor due to ambiguities arising from poor or insufficient questioning by interviewers during the surveys and limitations relating to respondent's ability to accurately and comprehensively recall transfers over the relevant period. Further, the estimates rely on migrant stock estimates, which may often be inaccurate. Finally, the method implicitly assumes that migrants from ECA sent on average the same amount of money as those from other regions. This may not always be true, especially for migrants from 63This estimateof Tajikistan's migrantstock was reportedinChrisParsons. 36 OECD countries that on average most likely send more remittances than their peers from the developing world. 8. The second statistic, the percentage of informal remittances by migrants, used in conjunction with the remittances reported in the balance of payments generates an alternative value of remittances. With the aim of estimating the total inflow of remittances to Tajikistan by calculating an informal multiplier, we can use the 755 respondents who reported that they remitted out of the total of 1,200 - 455 respondents are "non-remitters." Isolating this subgroup of remitting migrants raises immediate questions about the biases in their characteristics and behavior. Specifically, they are slightly less likely to be irregular migrants-95 percent instead of 97 percent for non-remitters. And remitters earn on average about 29 percent more than non- remitting migrants-US$306.82 versus US$237.94. 9. However, if the informal percentage can be estimated from a group that i s statistically .more likely to remit formally, then this estimate should be a low-end estimate of the entire remittance market size in Tajikistan. Using questions designed to elicit information about the most common channels used to remit wages, one can generate a figure for the percentage of formal remittances (70 percent) versus those that are sent through informal channels. The inverse of the percentage of formal remittances to Tajikistan (0.7) gives 1.43, and when multiplied by remittances reported in the BOP (US$252 million) it gives estimated value of remittances at US$360 million, respectively. Table A-11.2: Choice of Remittance Channel by Ta3k Migrants Method of Transfer Formal/ Informal Survey Percentage Response Money Transfer Organizations Formal 435 57.6% BankTransfers Formal 78 10.3% Post Offices Formal 15 2.0% Friendand acquaintances Informal 194 25.7% Anonymous Individuals Informal 19 2.5% InformalTransfer Offices Informal 14 1.9% Didnot remit/ Brought back by hand? ? 445 37 ANNEX 111: TRANSFER OFREMITTANCESTO TAJIKISTAN 1. This annex begins with a description of the mechanisms used by Tajik migrants to remit funds through formal channels (banks, money transfer systems, and postal service), their respective transfer costs, access to the transfer services by migrants and recipients, and their speed. Subsequently, the annex examines the disbursement of remittances in Tajikistan and highlightsimpediments to the use of formal transfer channelsby Tajik migrants. RemittingthroughCommercialBanks 2. Tajikistan's local commercial banks are active in money transfer services directly (bank- to-bank) and indirectly through money transfer systems (such as Western Union, Anelik or Contact) on a partnership basis. Banks provide direct money transfer and remittance services (i) internationally through correspondent accounts in foreign banks,64 and (ii)domestically through a national payments system operated by the NBT.65 The main remittances transfer services include Agroinvestment, Tajprombank, and Orion Bank. Their market shares in the transfer services to Tajikistan in 2004 were 45 percent, 21percent, and 15 percent, respectively. Figure A-111.1: FundTransfers through various Commercial Banksin2004 (96 of Total Transfers) 'a AIB rn Orionbank 0 Tajpronbank 0 Bonk EskhatarnAnomt Tajiksodirotbank rn .~ _~ 3. Technically, whenever a migrant remits money to Tajikistan from a bank in Russd, the correspondent account of the Tajik bank will be credited for the sent amount. Once the correspondent account i s credited, a notification to the headquarters of the Tajik bank i s sent via the SWIFT coding system (telex transmissions) that includes information on the account to be ~ ~ ~ 64 The most active Tajik banks have several correspondent accounts with the major Russian banks, including SberBank. SberBank of Russiais the largest bank in the country, with the most extensivebranch network and deposit base. The bank also operates the second largest bulk clearing system in the country for small intra- and inter-bank payments, serving about 150 banks in Russia and abroad. 65 Most banks confirmed that the NationalPaymentsSystem i s well functioning and consistentwith the existing needs of the financial system of the country. The payments system runs on a single clearing module for all payments, large and small. The settlement of inter-bank positions is done twice per day, and integratedwith the liquidity facility of the NBT. credited or the recipient of the transfer. Upon receiving such notification, the bank account of the sender or recipient is accredited, or issues a money order payment in the name of the recipient. 4. Funds can be transferred on a cash-to-cash, cash-to-account, or account-to-account basis. In theory, a labor migrant with an account in Russia could withdraw funds directly in Dushanbe from his Moscow bank account. The cost of money transfer from account-to-account and from cash-to-account charged to a sender i s about 2 percent for non-clients, and 1 to 2 percent for clients of the transferred amount, on average, if in the same currency. The recipient of the funds must also pay a flat fee of about US$30. Fees for foreign currency conversion would apply if withdrawals are made in different currencies. These fees vary considerably: banks generally apply a flat fee on currency conversion, but they also apply a spread on the foreign exchange rate, which i s particularly wide for small retail operations (spreads could vary between 2 and 5 percent). 5. The speed of transfer i s quite rapid. The transaction can take a few minutes, depending on the back-office and cashier speed. The electronic transfer i s real-time. Fixed costs for banks are relatively high when it comes to money transfers, and this discourages banks from engaging directly in remittances. Membership in the SWIFT system i s relatively inexpensive (about US$15,000 per year), but the system works on telexes. Managing a correspondent account instead i s costly in terms of profitability: Russian banks for example pay less than the overnight rate to Tajik banks. 6. Despite the number o f benefits o f opening an account in Russia,66inter-bank money transfers (account-to-account) are infrequently used among migrants because of difficulties in opening an account and foreign exchange regulations limiting the maximum amount that can be transferred. The survey of migrants indicates that only 18 percent of migrants have bank accounts because a Tajik migrant needs to show a passport and a "migration card." The latter can only be obtained with a valid working permit and i s required to open an account with a Russian bank. As most Tajik migrants are temporary seasonal workers and have no work permit, account- to-account transfers are less frequent. In contrast, cash-to-account transfers do not require particular documents, only a valid passport and the codes of the recipient bank and account. Furthermore, foreign exchange regulations in Russia and Tajikistan set limits on transfer amounts. InRussia, resident aliens cannot send more than US$5,000 per day, while non-residents do not have these limitations. The source of all cash deposited for money transfers above US$5,000 must be indicated in the memo of the transfer order. InTajikistan, the origin of funds mustbe declared for transfers exceeding US$2,000 incash. 7. Lastly, the educational levels of Tajik labor migrants also limit them from opening bank accounts. Anecdotal evidence shows that Russian banks prefer not to serve Tajik migrants for remittances, mainly because of lack of language and basic banking skills. Russian banks indicated that it takes sometimes 30 minutes of a teller's time to process a remittance of a few hundred U.S. dollars, which makes the transaction unprofitable. RemittingthroughMoneyTransfer Systems 8. There are a number of money transfer systems operating in Tajikistan. They all integrate into the banking system through partnerships and contract agreements based on which money transfer systems operate from the premises of the commercial banks; fees are shared almost 66These include access to basic banking services such as savings, ATMs and debit cards, and security. 40 equally. Many of these companies also have local offices in department stores, foreign exchange bureaus, post offices, airports, etc. Table A-111.1: Money Transfer Systemsin Tajikistan Western Money Inter- STB VIP Money Union Gram Anelik Migrom Express Express Contact Transfer FastMail Unistream HeadOffce USA USA Armenia Russia Russia Russia Russia Russia Russia Russia SettlementCenter Western American Anelik CB Inter- STB-Card JSCB VIPBank InpexBank Unistream Union Express Eurotrust Express Russian Bank Bank Paymenttime 15 minutes 10-15 3-24 hours 5-20 10-15 10-15 4-24 hours 5-10 minutes 10-15 10-15 minutes minutes minutes minutes minutes minutes Feesand Charges Minimum U S 6 1.5% 3% 2% 1.5% 2,5070 2% 1.50% 270 1% Maximum u s 5 3 6% 4% 4% 3% 3% 3% 2.50% 2% 9. In Tajikistan, money transfer systems are extremely active and technologically well equipped. The types of technologies used are relatively simple. Most of the money transfer systems use the internet for their Management Information System. Practically, remittances travel through the correspondent accounts of money transfer systems with the banks, while the communication on the remittance transfer i s transmitted to the local offices of the money transfer system by email. As a result, when a migrant sends a remittance, the money transfer system sends a message by internet to its local (recipient) office, with a code and details of the sender and recipient person. The deposited funds are credited in the account of the money transfer system with the commercial bank in the sending country. The money transfer system then settles all positions on its bank accounts, with different frequency, and credits and debits correspondent accounts in different recipient countries, depending on the flow of remittances and the daily expected payments. Some money transfer systems instead run real clearing houses, and thus integrate with the clearing systems of large bank networks. These systems, though, are more expensive as they require continuous clearing of each transaction and do not allow for less frequent settlement of correspondent accounts. 10. Money transfer systems differ slightly in terms of pricing and quality of service, thus indicating that remittance costs could be close to break-even, or that there i s a saturation effect that increases competition and lowers transaction costs. Most money transfer systems (except Western Union) establish a fee structure as a flat rate as a percentage of remittances sent. A minimum transfer fee charged to senders in host countries i s as low as 1 percent (Unistream), while a maximum fee i s 6 percent (Moneygram). All transfer transactions take only 15 minutes to complete and as a result speedi s close to real time. PostalTransfer 11. Tajikistan's postal system i s subordinated to the Ministry of Communications. The system i s complex and has a large infrastructure network with more than 600 offices spread all over the country with 1,500 employees. The technology at the Tajik post is very weak and outdated. The government tried to revamp the system in 2003 by issuing a new law on the National Postal System, which enables the postal system to provide basic financial services, including small/micro credits and savings services. However, the chronic lack of funding i s de facto impeding the further development o f the system and the opening to financial services operations. 41 12. Currently, the postal system receives assistancefrom USAID and UPU to automate some of the treasury operations, which now run on a paper based system. UPU i s planning to install new software in the automation center in Dushanbe, which will enable it to automate treasury operations throughout the whole 4 regional automation centers of the Tajik postal system. Further developments are hampered by the low level of connectivity of the branch network, which only has telephone lines (but not in all offices), and the overall required investment that runsinthe order of several millionUS dollars. 13. The postal system is active in the remittances segment through its partnershipswith the postal systems of Russia, FSU and CIS, but volumes are extremely low: US$683,000 of incoming and US$165,000 of outgoing remittances in 2004. The main problem with the postal system i s the low speed of the money transfer (a process that takes about two-to-three weeks), and the relatively high transfer cost (which runs in the order of 10 to 12 percent). Commissions and fees are paid by the recipient. The commissions between postal systems are regulated by international agreements, which are standard throughout the world: Tajik Post pays, for example, Russia Post SDR0.90 per transaction, irrespective of the amount and destination. The law limits the postal operations and ability to deal with remittance services as it is not authorize to operate in foreign currencies and it cannot disburse funds in foreign exchange. As a result, fees applied by the Postal System could be discounted by 2 or 5 percent, which i s the average foreign exchange spread applied by banks to consumers. 14. The process of sending reaittances through the postal system i s relatively simple. Tajik, Post has an account with a local commercial bank, Orion Bank, which manages the correspondent banking operations of the Post. As a migrant sends money through a postal system in a sending country, the funds are transferred through a foreign bank to the correspondent account of Tajik Post with Orion Bank. Subsequently, Orion Bank provides for the currency conversion and delivers the cash to the nearest postal office. The postal office then initiates the physical delivery of the funds to the local office of the recipients. Recipients and senders do not require specialized documents,just a passport or valid ID. 15. Clearly, the postal system capacity of providing remittance services i s hampered by the low level of technology and the speed of the transaction as well as transfer fees, which makes it largely uncompetitive. The main advantage of usingthe postal system might be the network and the outreach. However, the costs for modernizing the postal system seem too high, thus exceeding the expected benefits, especially considering the extremely low value of remittances in density areas where Tajik Post would have a competitive advantage. Furthermore, considering that 80 percent of remittances are currently cashed in Dushanbe, there i s little scope or benefit in expanding the outreach at such highcost. Disbursementof Remittances 16. The disbursement of remittances can be made in both local and foreign currencies (US Dollar, Russian Ruble, and Euro), thus increasing the degree of dollarization of the Tajik economy. Because of foreign currency preference, Tajik commercial banks need to purchase large amounts of foreign currency in cash during a peak season. Anecdotal evidence shows that commercial banks purchase foreign currencies from the NBT and from Russian, Kazakh or Kyrgyz banks. The NBT sells foreign currencies at a higher price, between 1.5 and 2 percent, due to its transaction costs structure,67while the prices charged by Russian or Kyrgyz banks are 67 The cost involves cash managementand purchases of foreign currencies in cash from other central and commercial banksin the region. 42 relatively lower than that of the NBT. Questions remain, though, on where Russia and Kyrgyz banks purchase foreign exchange in cash, which could be explained by large inflows from informal trade and businesses. 17. The NBT estimates that 80 percent of the remittances transferred through formal channels are disbursed in cash in Dushanbe, which i s the main port of entry for labor migrants returning from Russia. In the case of account-to-account transfers, disbursement can be made through the ATM or PTM systems. This choice is less popular because of the illegal status of migrants in sending countries, and the high demand for cash among beneficiaries in Tajikistan due to limited trust in the Tajik banking system. However, an increase in individual deposits in the banking system duringthe past few years reflects an increase inthe level of trust among the population. Box A-111.1: Tajikistan's ATMs and PTMs Tajik banks are now developing a network of Automation Teller Machines (ATM) and Post Terminal Machines (PTMs). Currently, the ATM services are being provided by the four banks in Tajikistan via the gradually expanding network. AIB i s the first bank that operates ATM network in Tajikistan. The network comprises 10 ATMs, 6 of which are located in Dushanbe. The bank i s planning to install an additional 20 machines throughout the country. The card clearing system of the AIB card is in Russia. The system is based on the EMV platform, which enables it to read both magnetic and chip cards. The AIB card is a debit card based on the Maestro System, the most common system in Europe. The fees Aiarged to non-clients of AIB are 2 percent for withdrawals in Somoni and 1.5 percent for withdrawal in US$. For clients, fees are 1percent for withdrawal in Somoni and US$2 for U S dollar and other currency withdrawals. PTMs are different from ATMs, with fewer modules, and much cheaper. They are located in stores and shops or department stores, and operated by retailers. Basically, a PTM would read the debit card, inquiry the bank of the client on the existing balance, and withdraw funds from the client account to be transferred to the retailer bank account. The retailer would then provide cash to the client net of applicable fees. Fees for these operations vary considerably, depending on the retailer, but generally 2 percent for Somoni and 4 percent for foreign currency withdrawals. The bank would only charge for the money transfer between accounts (from the client to the retailer) and between banks if the two accounts are in different banks. PTMs are often favored over ATMs due to the low cost of the machines and for certain specific functions, like payment o f salaries. Depending on the model, volumes and functions, an ATM can cost between Euro 25,000 and Euro 75,000, while a PTM could cost US$l,OOO. Also, connectivity i s a decision factor: ATMs need online and fast connections for real-time operations, while PTMs could work off-line, if necessary. The retailer could clear all operations once a day, before closirig of business, if the client i s trustworthy. 18. The analysis of financial infrastructure, costs, speed, and outreach reveals that Tajikistan's transfer system for remittances i s efficient, especially the money transfer systems that offer high speed transfer services at a relative low cost compared to other countries (Philippines, Mexico, etc). Philippines' bank charges, exclusive of the foreign exchange spread and recipient charge, range from about US$5 for a transfer from a foreign account to a branch account in a single bank, to about US$16, which may include courier services.68 19. However, it i s clear that there i s a need to increase intermediation of remittances among migrants and to extend the outreach of banks to rural areas. Cash disbursement of remittances Ibid., p.8. 43 implies a low degree of intermediation among recipients because of recipients' low level of income and lack of trust in the banking system. The concentration of disbursement location in Dushanbe indicates that outreach of banks in rural areas i s limited because of the low density of the rural economy, weak communication infrastructure, and low profit for retail banking in the regions and rural areas. Two banks, AIB and Amonat, have the largest branch networks in the country, with about 80 branches each. Amonat bank, in particular, plays a critical role in delivering fiscal and quasi-fiscal payments in the rural areas (almost on a monopolistic basis); however, its technology i s cumbersome. Only 10 percent of the branches are connected to headquarters, which makes it impossible to consolidate the bank's accounts on a daily basis. Table A-111.2: SelectedIndicators of Financial System (inthousand Somoni) 2001 2002 2003 2004 2005 RealGDP Growth 10.2 9.5 10.2 10.6 6.7 Per capita GDP (US$) 236.0 310.0 354.0 * GDP (thousandSomoni) 2528.8 3365.4 4757.8 6157.5 7201.1 GDP (thousand US$i) 991.7 1121.8 1615.6 2027.5 2404.0 ExchangeRate 2.6 3.0 2.9 3.0 3.1 Inflation 12.5 14.5 13.7 5.6 5.0 Interest rates onDeposits 11.2 10.9 8.6 To Individuals 6.6 8.0 7.6 To LegalEntities 14.6 17.2 16.2 On Credit inLocalCurrency 14.2 16.7 20.3 23.3 OnCredit to Individuals 30.2 29.1 28.5 On Credit to LegalEntities 13.7 15.1 20.1 Deposits Total Deposits/GDP 4% 4% 5% 5% 7% of which inLocal Currency (% of Total) 0.5 0.5 0.6 0.3 LegalEntities 0.0 0.0 0.0 0.0 0.0 of which inLocal Currency (% of Total) 0.0 0.5 0.6 0.6 0.4 Individuals 0.0 0.0 0.0 0.0 0.0 of which inLocal Currency (% of Total) 0.0 0.0 0.1 0.1 0.1 Total CreditdGDP 15% 13% 11% 18% 18% M2/GDP 5.6 6.0 5.5 4.2 3.6 Source: TheNational Bank of Tajikistan 44 ANNEX IV: MONEY TRANSFER SYSTEMINTAJIKISTAN 1. There are a number of money transfer systems operating in the Republic of Tajikistan. They all work with the banking system under partnerships and contract agreements, according to which banks and money transfer systems operate from the premises of the commercial banks. Fees are shared almost equally. Many of these companies also have local offices in department stores, foreign exchange bureaus, post offices, and airports. 2. Money transfer systems have a clear competitive advantage compared to banks. In particular, fees and commissions are relatively low, standardized and easily accessible and readab!e. Considering the fixed costs for banks in running small value transactions, money transfer systems have a clear market niche. In addition, money transfer systems do not require bank accounts and particular documents, only a passport or a valid IDi s necessary. 3. Money transfer systems also have an impressive network of offices and operations through their partnerships with banks. This i s an important aspect because it enables money transfer systems flexibility in expanding and rationalizing the network according to the demand. Another competitive advantage i s the possibility to use different technologies, including the intermt. Many money transfer systems are starting to include additional services to remittances, such as free text messages, the possibility to transfer remittances directly to a bank account, and free assistanceincompiling papers and other documents. 4. Money transfer systems are growing fast in terms of technology and outreach. Technology innovations and improvements in connectivity, including use of the Internet, enabled other money transfer systems to enter the market and drastically lower transaction costs. Fast growth has been accompanied by market liberalization and sharp decreases in service costs. The following sections provide a summary of the main services and costs charged by the money transfer systems operating inthe CIS. Western Union 5. Western Union i s one of the oldest money transfer systems in the world. The U.S.-based Western Union has been engaged in money remittances since 1871 and has since evolved by using modern communication and processing means. The most updated electronic technologies allow transferring money to any destination through the company's network in just a few minutes. Each remittance i s protected by its world-class security system, which guarantees that the money will be paid only to the indicated person. The international system of money transfers established by the company works in 180 countries and includes more than 212,000 bureaus. Just in the CIS, FSUand Baltic countries, Western Union has about 23,000 bureaus. InTajikistan, its services are provided by Orionbank, Amonatbank, Tojiksodirotbank, Tajprombank, Sohibkorbank and Eskhata Bank. 6. Western Union operates money transfers for both individuals and businesses, and it has a well developed internet website with information including a location search engine. Interestingly enough, the website does not provide information on fees and commissions. The cost of services depends on the size of the transaction and the destination. Although Western Union has been lowering fees during the last two years, it remains the most expensive among the money transfer systems. Box A-IV.l: Sending and ReceivingMoney via Western Union A sender must fill-out the form "For Transfer of Money", which requires informationon: 0 amount to be transferred in figures and in words; 0 full name of the beneficiary; full name of the sender; 0 phone number of the sender; 0 sender's address. After signing the form, the cashier would issue a receipt with a ten-digit code of the transfer, which i s communicated to the beneficiary. The following services are offered for additional fees: 0 telephone or cable notification to the beneficiary; brief notice or text message; delivery of a check for the amount of transfer. To receive the funds, a beneficiary must fill-out the form "For Receipt of Money", which requires the following information: 0 full name; 0 the ten-digit code; address; 0 full name of the sender; 0 sender's phone number (optional); expected amount; city and country from which the money was sent. Money Gram 7. Money Gram i s the second largest global network of money remittances in the world, which i s also the main characteristic and competitive advantage of this system. Money Gram i s also an American company and was established by the American Express Group as an affiliate to deal with money transfers. The company performed the first money transfers in 1940 without opening bank accounts. Currently the owners of Money Gram are Viad Corporation and Travelers Express Company. Unlike its closest competitor (Western Union), the Money Gram system i s designed only for individuals. Money Gram has 3 to 6 percent of the market share in this segment. InTajikistan, only KafolatBank provides the services of Money Gram. 8. To transfer funds, the documents required by Money Gram of senders are a passport and/or and ID. Fees are paid by the sender, not the receiver. It takes no more than 10 to 15 minutes for a money transfer to be executed through the system, although there are variations depending on the regions and the recipient countries. Money Gram provides a text message service, according to which senders can send a 10 word message to receivers for free. Fees depend on the size of the transaction and the recipient country as shown below as of July 2005. Table A-IV.l: Pricing of Money Gram Remittance Services