79246 Paving the road to better Financial Decision-Making in Tajikistan May 22, 2013 Financial Capability and Consumer Protection Survey Report Financial Inclusion and Consumer Protection Service Line Contents Acknowledgments iii 1 Key Findings and Policy Recommendations 1 1.1 Knowledge of Financial Concepts and Products . . . . . . . . . . . . . . . . . 2 1.2 Financial Attitudes and Behaviors . . . . . . . . . . . . . . . . . . . . . . . . . 4 1.3 Financial Inclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1.4 Consumer Protection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 2 Background 10 3 Financial Capability 12 3.1 Knowledge of �nancial Concepts and Products . . . . . . . . . . . . . . . . . . 12 3.1.1 Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 3.1.2 Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 3.2 Financial Attitudes and Behavior . . . . . . . . . . . . . . . . . . . . . . . . . 21 3.2.1 Dimensions of Financial Capability . . . . . . . . . . . . . . . . . . . . 21 3.2.2 Financial Capability Related to ‘not overspending’ Behavior . . . . . . 22 3.2.3 Financial Capability Related to ‘living within means’ Behavior . . . . . 24 3.2.4 Financial Capability Related to ‘budgeting’ Behavior . . . . . . . . . . 25 3.2.5 Financial Capability Related to ‘saving’ Behavior . . . . . . . . . . . . 26 3.2.6 Financial Capability Related to ‘planning for unexpected and old age expenses’ Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 3.2.7 Financial Capability Related to ‘(non-)impulsiveness’ . . . . . . . . . . 29 3.2.8 Financial Capability Related to ‘achievement orientation’ . . . . . . . . 30 3.2.9 Financial Capability Related to ‘seeking information and learning from others’ Attitudes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 3.2.10 Financial Capability Related to ‘foresightedness’ . . . . . . . . . . . . . 32 3.3 Relationship between Financial Knowledge and Financial Attitudes and Behavior 33 4 Financial Inclusion 34 4.1 Access and Use of Financial Services . . . . . . . . . . . . . . . . . . . . . . . 34 4.1.1 Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 4.1.2 Financial Institutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 4.2 Relationship between Financial Inclusion and Financial Capability . . . . . . . 43 4.2.1 Relationship between Financial Inclusion and Financial Knowledge and Awareness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 4.2.2 Relationship between Financial Inclusion and Financial Attitudes and Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 5 Consumer Protection 45 A Background Tajik Survey 48 B Financial Capability 50 B.1 Knowledge of Financial Concepts and Products . . . . . . . . . . . . . . . . . 50 B.2 Knowledge of �nancial Products . . . . . . . . . . . . . . . . . . . . . . . . . . 52 B.3 Financial Attitudes and Behavior . . . . . . . . . . . . . . . . . . . . . . . . . 55 - ii - C Financial Inclusion: Access and Usage of Financial Services 60 C.1 Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 C.2 Financial Institutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 - iii - Acknowledgments This Financial Capability and Consumer Protection Report was prepared by a team led by Siegfried Zottel (Economist) from the World Bank’s Financial Inclusion and Infrastructure Global Practice, with contributions from Valeria Perotti (Social Protection Specialist) and Tillmann Heidelk (Research Consultant) Peer review comments were received from Aidan Coville (Economist) and Uzma Khalil (Senior Financial Sector Specialist). Douglas Pearce (Acting Service Line Manager, FFIMS) provided valuable guidance and the report also bene�ted from comments received from Andrej Popovic (Financial Sector Specialist) and Juan Carlos Izaguirre (Financial Sector Consultant). In addition, statistical support provided by Nodar Kapanadze (Survey Consultant) is gratefully acknowledged. The team expresses its appreciation to the Tajik authorities, including the National Bank of Tajikistan (NBT) and the Statistical Agency under President of the Republic of Tajikistan (RT) for their cooperation and collaboration during the preparation and implementation of the survey. The team would also like to thank the �rm ‘Zerkalo’ which was selected to undertake this survey work. Finally, the team owes particular appreciation to all Tajik women and men who patiently responded to the survey. The Report was prepared as part of the Bank-Netherlands Partnership Program grant on “Improving Financial Literacy and Consumer Protection in Low Income Countries�, funded by the Government of the Netherlands and the World Bank. - iv - List of Figures 1 Financial Literacy Score, by Region . . . . . . . . . . . . . . . . . . . . . . . . 12 2 Overview Financial Literacy I . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 3 Overview Financial Literacy II . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 4 Financial Literacy Score, by being Head of Household . . . . . . . . . . . . . . 15 5 Financial Literacy Score, by Gender . . . . . . . . . . . . . . . . . . . . . . . . 15 6 Financial Literacy Score, by Urbanization . . . . . . . . . . . . . . . . . . . . 15 7 Financial Literacy Score, by Education . . . . . . . . . . . . . . . . . . . . . . 15 8 Selected Financial Literacy Score, by approximate Household Debt . . . . . . 16 9 Financial Literacy Score, by approximate Household Debt . . . . . . . . . . . 16 10 Financial Products Known in Regions . . . . . . . . . . . . . . . . . . . . . . . 17 11 Overview Awareness Financial Institutions Service Offers I . . . . . . . . . . . 17 12 Overview Awareness Financial Institutions Service Offers II . . . . . . . . . . . 17 13 Knowledge Financial Institutions Service Offers, by Number of Economically Dependent Children . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 14 Knowledge Financial Institutions Service Offers, by Employment Status . . . . 19 15 Knowledge Financial Institutions Service Offers, by Education . . . . . . . . . 20 16 Knowledge Financial Institutions Service Offers, by Financial Literacy . . . . . 20 17 Knowledge Financial Institutions Service Offers, by Media Usage . . . . . . . . 20 18 Overview Financial Capability Scores . . . . . . . . . . . . . . . . . . . . . . . 21 19 Financial Capability Related to ‘not overspending’ Behavior in Regions . . . . 22 20 Financial Capability Related to ‘not overspending’ Behavior, by approximate monthly household income. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 21 Financial Capability Related to ‘living within means’ Behavior in Regions . . . 24 22 Financial Capability Related to ‘living within means’ Behavior, by approximate monthly household income. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 23 Financial Capability Related to ‘budgeting’ Behavior in Regions . . . . . . . . 25 24 Financial Capability Related to ‘budgeting’ Behavior, by Employment Status. 25 25 Financial Capability Related to ‘budgeting’ Behavior, by Cohabiting Status. . 25 26 Financial Capability Related to ‘saving’ Behavior in Regions . . . . . . . . . . 26 27 Financial Capability Related to ‘saving’ Behavior, by approximate monthly household income. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 28 Financial Capability Related to ‘saving’ Behavior, by Saving as a Child. . . . . 26 29 Financial Capability Related to ‘planning for unexpected and old age expenses’ Behavior in Regions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 30 Financial Capability Related to ‘planning for unexpected and old age expenses’ Behavior, by Age. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 31 Financial Capability Related to ‘planning for unexpected and old age expenses’ Behavior, by Saving as a Child. . . . . . . . . . . . . . . . . . . . . . . . . . . 28 32 Financial Capability Related to ‘(non-)impulsiveness’ Behavior in Regions . . . 29 33 Financial Capability Related to ‘(non-)impulsiveness’ Behavior, by Age. . . . . 29 34 Financial Capability Related to ‘achievement orientation’ Behavior in Regions 30 35 Financial Capability Related to ‘achievement orientation’ Behavior, by Saving as a Child. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 36 Financial Capability Related to ‘seeking information and learning from others’ Behavior in Regions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 37 Financial Capability Related to ‘foresightedness’ Behavior, by Cohabiting Status. 31 38 Financial Capability Related to ‘foresightedness’ Behavior, by Education. . . . 31 -v- 39 Financial Capability Related to ‘foresightedness’ Behavior in Regions . . . . . 32 40 Financial Capability Related to ‘foresightedness’ Behavior, by Cohabiting Status. 32 41 Financial Capability Related to ‘foresightedness’ Behavior, by Education. . . . 32 42 Sum of Financial Products, by Region . . . . . . . . . . . . . . . . . . . . . . 34 43 Overview Financial Products I . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 44 Overview Financial Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 45 Overview Financial Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 46 Overview Financial Products II . . . . . . . . . . . . . . . . . . . . . . . . . . 36 47 Usage Financial Products, by Education . . . . . . . . . . . . . . . . . . . . . 37 48 Usage Financial Products, by Age . . . . . . . . . . . . . . . . . . . . . . . . . 37 49 Usage Financial Products, by Urbanization . . . . . . . . . . . . . . . . . . . . 37 50 Using Services offered by Financial Institutions, by Region . . . . . . . . . . . 38 51 Utilization Financial Service, Base Population . . . . . . . . . . . . . . . . . . 38 52 Using Services of Currency Exchange Offices, by Region . . . . . . . . . . . . . 40 53 Using Services of Money Transfer Operators, by Region . . . . . . . . . . . . . 40 54 Using Services of Micro�nance Organizations, by Region . . . . . . . . . . . . 40 55 Utilization Financial Service, by being Head of Household . . . . . . . . . . . 41 56 Utilization Financial Service, by Cohabiting Status . . . . . . . . . . . . . . . 41 57 Utilization Financial Service, by Employment Status . . . . . . . . . . . . . . 41 58 Utilization Financial Service, by approximate Household Income . . . . . . . . 42 59 Utilization Financial Service, by Education . . . . . . . . . . . . . . . . . . . . 42 60 Utilization Financial Service, by Saving as a Child . . . . . . . . . . . . . . . . 42 61 Financial Inclusion vs. Financial Literacy . . . . . . . . . . . . . . . . . . . . . 44 62 Financial Product Usage vs. Financial Literacy . . . . . . . . . . . . . . . . . 44 63 Overview of Satisfaction with Services offered by Financial Institutions . . . . 46 64 Satisfaction with Services Financial by Financial Institutions, by Gender . . . 46 65 Satisfaction with Services Financial by Financial Institutions, by Age . . . . . 46 66 Satisfaction with Services Financial by Financial Institutions, by Employment Status . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 67 Comparison FCCP and LSMS: Regions . . . . . . . . . . . . . . . . . . . . . . 48 68 Comparison FCCP and LSMS: Urbanization . . . . . . . . . . . . . . . . . . . 48 69 Comparison FCCP and LSMS: Household Income . . . . . . . . . . . . . . . . 48 70 Comparison FCCP and LSMS: Gender . . . . . . . . . . . . . . . . . . . . . . 49 71 Comparison FCCP and LSMS: Age . . . . . . . . . . . . . . . . . . . . . . . . 49 List of Tables 1 International comparison of knowledge of basic �nancial concepts (in % of adults). 15 2 Information on sampling design. . . . . . . . . . . . . . . . . . . . . . . . . . . 48 3 Financial Knowledge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 4 Knowledge of Speci�c Financial Products . . . . . . . . . . . . . . . . . . . . . 53 5 Financial Capability Scores – Part 1 . . . . . . . . . . . . . . . . . . . . . . . 56 6 Financial Capability Scores – Part 2 . . . . . . . . . . . . . . . . . . . . . . . 57 7 Number of Financial Products . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 8 Usage of Speci�c Financial Products . . . . . . . . . . . . . . . . . . . . . . . 63 - vi - Glossary DRS – Region of Republican Subordination FCCP – Financial Capability and Consumer Protection FI – Financial Institution GBAO – Gorno-Badakhshan Autonomous Province LSMS – Living Standards Measurement Study MFO – Micro�nance Organization MTO – Money Transfer Operator NBFI – Non-Bank Financial Institution NBT – National Bank of Tajikistan PSU – Primary Sampling Unit RT – Republic of Tajikistan TJS – Tajikistani somoni - vii - 1. Key Findings and Policy Recommendations 1. It is suggested that a comprehensive �nancial education strategy be devel- oped based on the results of the �nancial capability survey.1 The survey identi- �ed numerous �nancial capability issues across various segments of the population and a number of initiatives are suggested in this report, which could be undertaken to improve the knowledge of �nancial concepts and products, as well as attitudes and �nancial be- haviors. In order to ensure that available resources are used in the most efficient way, it is important to prioritize those �nancial capability enhancing programs, which are considered to be most feasible and cost-effective. The process of developing a national �nancial education strategy could serve to identify key priorities. Such priorities could be set based on a number of criteria, including the need, costs, outreach, desired im- pacts, and replicability of programs and could also present opportunities to leverage on existing programs. Both the development of a strategy and the setting of priorities would require a wide consultation process including various stake- holders from public, private, and non-pro�t sectors, which would facilitate a wider consensus building about the importance of �nancial education. 2. Further, the survey results and recommendations outlined in this report could provide guidance on the development and implementation of �nancial education programs. They could not only offer useful inputs for the design and monitoring of priority programs, but they could also align priorities and coordinate them with other initiatives and reforms such as those in the area of �nancial consumer protection. One of the key recommendations in the strategy development process would be to ensure that all proposed measures and programs be tested initially on a small scale and only be rolled out upon evaluation of their positive impacts. For more information about the process of strategy development see OECD/INFE High Level Principles on National Strategies for Financial Education.2 1 All results presented in this report are indicative at best and cannot be interpreted as causal. 2 Online available at: http://www.oecd.org/finance/financial-education/OECD_INFE_High_Level_ Principles_National_Strategies_Financial_Education_APEC.pdf -1- 1.1. Knowledge of Financial Concepts and Products Key Findings 3. Financial knowledge and awareness levels are a signi�cant challenge in Tajikistan, as for many countries at a similar income level. Although the vast majority of the population possesses the necessary skills to perform basic �nancial calculations, they lack speci�c knowledge to make savings and investment decisions. Rural dwellers and people with lower educational attainment are, compared to the rest of the population, less knowledgeable about basic �nancial concepts, such as inflation, interest rates, compound interest, and risk diversi�cation. 4. Money transfer operators (MTOs) and currency exchange offices are widely known as �nancial service providers, but awareness of broader �nancial services provided by banks and Micro�nance Organizations (MFOs) is much more limited. While not surprising for the relatively new institutional model of MFOs, this is a striking �nding with regard to banks and suggests a focus area for �nancial knowledge and awareness efforts. Another striking �nding is that residents in the capital city of Dushanbe are less knowledgeable concerning services provided by banks, MFOs, and MTOs, compared to people living in other regions of Tajikistan. Likewise, having more economically dependent children, being less literate, using fewer types of media regularly, and not having saved as a child are also negatively associated with knowledge of �nancial products. -2- Recommendations 5. In order to raise �nancial knowledge of potential users of savings and invest- ment products, innovative and interactive measures, and “edutainment’ in particular, should be considered to reach adult target audience. Research has shown that innovation on delivery matters because unlike children, adults often have stubborn preferences that are difficult to change. They also have short attention spans and many simply switch off and �nd lecture-based �nancial capability programs boring. Conveying �nancial knowledge messages through innovative ways such as using popu- lar TV soap operas, �lms, videos or radio programs can be quite effective in improving knowledge and also in altering behavior. In order to be successful, it is pivotal that these programs be delivered in an engaging and entertaining manner through appealing sto- ries that stick to memories. They are also presumed to be more effective if messages are repeated and reinforced over time For instance, in Kenya, a popular television drama, Makutano Junction, incorporated �nancial education messages into some of its stories. These messages aim to encourage people to save regularly or to open a bank account, rather than to keep money under a mattress. Viewers are able to send text messages to obtain a leaflet related to these topics. As with other soap operas, people watch Makutano Junction because they identify with the characters and enjoy the stories; but in the course of watching the shows, they bene�t from the �nancial education messages. 6. Publications can be a useful means of conveying �nancial education, since each copy can be read by several people and can be retained for future reference. A diverse range of publications can be used, including leaflets, booklets, fliers and posters. Articles in newspapers and magazines are also important tools, especially if contained within general sections of the newspaper or magazine, rather than in specialist �nancial ones. Comic books have been found to be particularly effective in several countries, such as Kenya, India, and South Africa, where literacy levels are low. In such cases comic books effectively facilitate discussion within the family on topics related to �nancial literacy. 7. Financial education should be delivered through trusted organizations and individuals with whom the target audience deals with on a regular (day-to- day) basis. Managing one’s personal �nances is an important aspect of everyday living. Therefore, many organizations have an interest in helping people to become �nancially knowledgeable and capable. Especially to reach remote communities in rural areas, and groups that are hard to engage, it could be considered to collaborate with community organizations and trusted intermediaries such as local community leaders, social workers, and to support them with resources, training or funding, if necessary. 8. Particular areas of focus for �nancial education programs include: (i) the meaning of interest rates, (ii) the ability to understand the cost of loans, (iii) how interest earned on interest and inflation affect their savings, (iv) the main purpose of insurance products, and (v) the key bene�ts, features, and risks of basic retail products such as savings accounts, consumer loans, remittances, and mortgage loans. 9. The delivery of �nancial education programs should take advantage of “teach- able moments�. One of the key lessons learned from impact assessments of �nancial education programs is that �nancial education works best when delivered to adults dur- ing teachable moments. Teachable moments are times in people’s lives when they are more likely to be receptive to new information as they can relate it directly to their own life events. In terms of �nancial education, key teachable moments when one may -3- re-examine his personal �nances include marriage, new employment, and the launch of a new business. Given the vital role remittances play in the wellbeing of the Tajik pop- ulation, �nancial education programs should be extended to migrants so that they learn about issues related to remittance transfers. Imparting �nancial education to recipients upon the collection of their remittances may not be as effective as targeting migrant workers just prior to their departure (e.g. when they apply for visas at the Russian embassy). Migration is an important teachable moment and many Tajiks are likely to be more receptive to key �nancial knowledge at that time as they will not yet have the �nancial familiarity or habits needed to effectively manage their higher earnings. 1.2. Financial Attitudes and Behaviors Key Findings 10. Tajiks demonstrate high levels of �nancial capability along most indicators measured in this survey, although they score higher with regard to their attitudes and motivations than their �nancial behaviors. The survey results suggest that on average Tajiks show strengths in not over- spending, as they tend to be not very impulsive. This survey also measures for far-sightedness, inclination to seek information and advice, and achievement orientation, all indicators on which Tajiks score highly. Compared to these �nancial capability aspects, Tajiks score slightly worse with regard to �nancial behaviors such as living within their means and budgeting. By far the lowest average scores are found for �nancial behaviors that relate to planning for one’s own future, such as saving and planning for unexpected expenses and old age. These low scores are worrying given their implications for people’s ability to smooth consumption, to cope with economic shocks, to generate lump sums for productive investments, and eventually for their long-term wellbeing. As compared to the highest income group, people living on lower incomes are signi�cantly more limited in their ability to cope with expected and unexpected �nancial shocks, and to make ends meet. Consequently, solving immediate problems seems to draw their attention away from their long-term needs. Another vulnerable group is Dushanbe’s population which, compared to residents from other regions, is less capable of putting money aside, and planning for unexpected expenses and old age. Likewise, Tajiks who reported to have saved as a child perform signi�cantly better in saving and planning for unexpected expenses and old age than their counterparts who did not learn this �nancial behavior in their early childhood. Recommendations 11. Provision of �nancial education from early age should be encouraged. Start- ing early can have huge value. Opportunities to provide �nancial education in schools need to be explored, as basic principles of �nancial capability such as savings or making provisions for the future should be acquired at a young age. If people learn good habits on how to manage their money from a young age, they are more likely to stick to them throughout their lives. Although, there is mixed international evidence on the effective- ness of school-based �nancial education programs on changing consumer behaviors, there are lessons learned from other countries that have undertaken such programs. Success has been observed when providing education in ways which students �nd relevant to -4- their lives either currently or in the near future, and if it is interactive. Therefore, a high-quality material / textbook developed by experts is required, and teachers need to be well-trained on the content and techniques. Moreover, as existing curricula are al- ready saturated, it is practicable to integrate �nancial education into a variety of existing subjects, such as math, economics, social studies, rather than adding a new subject into the curriculum. If resources to train teachers and to develop and provide teaching and learning materials are limited, it may be best to focus, at least initially, on incorporat- ing �nancial education into one or two subjects (if possible, these should be compulsory subjects so that all students bene�t) over three or four consecutive academic semesters (preferably close to the time when students tend to leave school). 12. In addition to primary target groups, �nancial education programs should also focus on their complements. Thinking broadly about complementary target groups is important. For instance, when not only students but also their parents are targeted with �nancial capability training, the expected outcomes are even larger. Simi- larly, targeting not only remittance receivers but also senders is expected to deliver more promising results, since both the desire to change behavior and the ability to apply new information is much greater if complementary groups are targeted.3 3 Yoko Doi, David McKenzie and Bilal Zia ‘Who You train Matters: Identifying Complementary Effects of Financial Education on Migrant Households’, World Bank Policy Research Working Paper no. WPS6157. -5- 1.3. Financial Inclusion Key Findings 13. Despite the expansion of the �nancial sector in the past decade, large parts of the Tajik population are still �nancially excluded. Compared to all avail- able �nancial products and services, insurance policies, private pension products, and money transfer services are widely used whereas loans are in lower demand. Formal and informal savings mechanisms are hardly used at all. Currency exchange offices and MTOs are by far the main �nancial institutions (FIs) whose products and services Tajiks use. Services offered by any other FI only play a minor role. Tajiks who are out of labor force represent the most underserved part of the population. They do not only use fewer �nancial products, but are also less likely to access services of MTOs and MFOs. Compared to people living in rural areas, urban dwellers tend to hold fewer �nancial products. Although, the number of �nancial products held does not vary with income, low income groups are less likely to have used any services provided by MTOs and currency exchange offices. Banks’ services are in lower demand by people with lower educational attainment, those with more economically dependent children, and those who did not save as a child. 14. Despite many other barriers, access and usage of �nancial products and ser- vices can be constrained by lack of �nancial knowledge. Tajiks who answered more of the quiz-like �nancial literacy questions correctly, also hold more �nancial products. This positive relationship turns stronger and becomes signi�cant if only adults with at least one �nancial product are considered. 15. Likewise, access and usage of �nancial products can be constrained by �- nancial attitudes and behavior. The survey provides strong evidence that people who are achievement-oriented and those who are more inclined to save, to use infor- mation, to plan for unexpected expenses and old age, also tend to use more �nancial products. This may be due to the fact that people who plan for unexpected events and for old age expenses are more likely to use saving and investment products. In addi- tion, people who do better in terms of achievement orientation and getting information and advice, may feel more con�dent about using �nancial products. Recommendations 16. To increase uptake of �nancial products, �nancial knowledge, and capability- enhancing programs could be combined with available �nancial products. Although large parts of the population are �nancially excluded, the survey results may indicate that enhanced �nancial knowledge of those who already use a �nancial product is key for further �nancial sector widening and deepening. Financial education programs could be tied to those �nancial products and services most people can access and use, such as compulsory insurance products (including motor third party liability, �re and renters’ insurance of state-owned property), private pensions products or formal loans. Moreover, survey results suggest that enhancing �nancial capabilities, especially with regard to building savings cushions for unexpected economic shocks and planning for old age expenses, may be another crucial step toward the goal of accelerating �nancial inclusion Thus, it should be considered to require insurance companies, banks and other -6- FIs to offer education programs, which inform customers about the importance of sav- ings, even if only a little, and the need to plan for the future. It should be ensured, however, that any educational materials are truly educational, which means that they need to be informative, clear, impartial, and free from marketing. Given that Dushanbe’s population appears to struggle in putting money aside and planning for unexpected or old age expenses, it could make sense to pilot these programs �rst in the capital and to evaluate their effectiveness via Randomized Controlled Trials before scaling up. 17. Generic �nancial education programs alone may be ineffective, but when combined with other interventions can be valuable. The limitation of tradi- tional �nancial education in changing �nancial behavior is well documented by research. However, coupling the �nancial education intervention with reminders (e.g. to save) or speci�c �nancial capability messages (e.g. to shop around or to make larger remittance transfers) is more likely to induce behavioral change.4 Another advantage of such inter- ventions is that they are also quite cost-effective and could be taken rapidly to national scale. 18. The NBT may consider sharing the results of this survey with commercial banks and MFOs in particular to potentially develop savings products tai- lored to the needs of underserved parts of the population. Consumer choice is rather limited or absent in Tajikistan. It could make good business sense for FIs to develop products which �t the needs of underserved populations and help them to reach personal savings goals. These could not only comprise low cost “no-frills� or basic accounts with less burdensome opening requirements, but also savings products with product design features that affect the extent of individuals’ use of �nancial services, such as commitment savings or labeled accounts. The former consist of accounts where a certain amount is deposited and access to cash is relinquished for a period of time or until a goal has been reached. The latter are accounts created with explicit savings goals such as a car purchase, housing, or education. Credit products could include the design feature of small, initial tester loans that provide information to lenders useful for assessing risk on subsequent larger loans. 19. Removing impediments to �nancial inclusion other than a lack of �nancial knowledge is important. To stimulate demand for formal savings and credit prod- ucts it should be considered to further develop and expand alternative banking delivery channels. Mobile or agent banking can dramatically reduce the cost of providing �nan- cial services in an environment of low population density and low per capita income prevailing in Tajikistan. These are the settings in which mobile banking technologies can have the greatest welfare bene�ts because they offer a commercially viable way of reaching locations and segments of the population that were previously excluded from formal �nancial services due to the prohibitively high costs of providing such services. Mobile banking services could also make it possible for Tajiks, especially those who live on low incomes, to receive remittances from family members or friends abroad. This would be a signi�cant improvement given the vital role remittances play in Tajikistan. 4 Karlan, Dean, et al. Getting to the top of mind: How reminders increase saving . No. w16205. National Bureau of Economic Research, 2010. -7- 1.4. Consumer Protection Key Findings 20. Consumers do not widely report complaints or other conflicts with �nancial service providers. The low level of self-reported �nancial service provider conflicts certainly relates to the low levels of �nancial inclusion but most likely also to concerns of adverse reactions from �nancial providers, and limited awareness and understanding of terms and conditions of �nancial contracts, �nancial consumer rights, or obligations of FIs. 21. Moreover, around a third of those who did not encounter any conflicts would not do anything if they faced a �nancial service provider conflict. This may be due to the fact that they do not know which institutions could help them to solve �nancial service provider conflicts or because they do not trust that formal redress mechanisms would do anything for them. 22. Financial services offered by insurance companies satisfy their customers less than those provided by other FIs. Those parts of the population who have used the services offered by commercial banks, intermediary organizations, MFOs, Non-Bank Financial Institutions (NBFIs), MTOs and currency exchange offices appear to be highly satis�ed with their services. Only insurance companies seem to offer lower quality services which satisfy their customers much less. Given the by international standards the fraction of premiums paid out in claims is rather small, these �ndings are not surprising. Recommendations 23. Delivering �nancial education is not enough, measures to strengthen the con- sumer protection framework also need to be taken to ensure that new and existing consumers bene�t from the �nancial services they use. It needs to be ensured that consumers are provided with sufficient information to enable them to select �nancial products and services that are the most suitable and affordable. Therefore, it should be considered to require banks and MFOs to inform consumers about the exis- tence of the deposit insurance fund, and the nature and limits of insurance provided. In addition, it should be considered to require Key Facts Statements for all basic consumer �nance products and to provide price comparison information on regulator’s websites. For those without internet access, complimentary dissemination mechanisms should be provided. 24. FIs should also be required to provide customers with key information on internal complaints handling focal points and standards. Legal or regulatory provisions should require all FIs to provide customers with contact information of the person or division within the institution that is authorized to receive and respond to complaints. This information should be included not only in advertising or marketing materials but also in materials, provided to consumers when applications are �led and contracts are concluded. Consumers should also be provided with information on the �nancial institution’s complaints handling procedures. 25. Giving adequate recommendations to consumers is especially important in the insurance sector, as insurance products are more complex and therefore -8- less understood by consumers. Despite the inability of most consumers to assess whether the policy offered by a particular insurer is appropriate for their circumstances, all insurance is sold on a ‘non-advised’ basis (i.e. without any obligation on the insurer or its agent to consider whether the policy is appropriate for the consumer). The current system’s reliance on the insurance company (or its agent) to give oral explanations to consumers about policy features is likely to mislead some consumers into believing that they are receiving advice.5 26. The lack of trust in Tajikistan’s �nancial sector has been recognized as an- other impediment to deepening and broadening �nancial markets. In the absence of trust, people are less likely to access and use the formal �nancial system, which in turn deprives the system of the support and funding necessary for its develop- ment and stability. Improvements in customer services and consumer rights protection are, therefore, critical issues. 27. Financial regulators should analyze the statistics on consumer complaints submitted by �nancial institutions and use this information as input to their supervisory and regulatory activities. Based on the analysis of information re- garding consumer complaints and inquiries, each regulator could propose guidelines, instructions, or awareness campaigns that address the common problems identi�ed in such analysis. 5 Source: Diagnostic Review of Consumer Protection and Financial Literacy, World Bank, forthcoming. -9- 2. Background 28. This survey instrument has been extensively tested in the context of low- and middle income countries. The survey instrument used is based on an instrument developed with support by the Russia Financial Literacy and Education Trust Fund to measure �nancial capability in the context of low and middle income countries. Extensive qualitative research was conducted to develop this survey instrument, including about 70 focus groups and more than 200 cognitive interviews in eight countries to identify the concepts that are relevant in low- and middle-income settings, and to test and adapt the questions to ensure that they are well understood and meaningful across income and education levels. The instrument is currently used or will be used in 14 countries in Latin America, Africa, Middle East, and Asia Paci�c. 29. The survey instrument allows �nancial capability, �nancial inclusion, and consumer protection issues to be assessed and measured. Financial capability is measured by knowledge of �nancial concepts and products, and by attitudes, skills and behavior related to day-to-day money management, planning for the future, choosing �nancial products and staying informed. In order to jointly analyze �nancial capability and inclusion, the survey instrument captures information on access to different kinds of �nancial products and service providers. The consumer protection section gathers infor- mation on incidence of conflicts with �nancial service providers and levels of satisfaction with �nancial products offered by different FIs. The survey instrument has been further customized to the policy priorities of the NBT by through adding Tajikistan-speci�c questions (i.e. deposit insurance awareness or levels of satisfaction with currency ex- change offices and MTOs). The survey instrument was translated into Tajik, as well as Russian and Uzbek, the two other major languages spoken in Tajikistan. For validation purposes it has also been back-translated into English. 30. The Tajik survey is designed to be nationally representative of the adult population aged 18 years and older, and it comprises a total sample of 1,000 individuals. The sample was designed to allow reliable estimation of main variables of interest on the national level. The most recent 2010 population census data, kindly pro- vided by the Statistical Agency under the President of the RT, was used as a sampling frame. The population was divided into 9 strata: 5 regions (Dushanbe, DRS, Khat- lon, Sogd, and GBAO) and each region was further divided into urban/rural strata, except the capital Dushanbe. To ful�ll the requirement of a scienti�cally sound survey a probabilistic sample selection mode was applied. The sample was selected through a two stage cluster sampling method, with census enumeration areas as primary sam- pling units (PSUs) and households as secondary sampling units. The effective sample size was established with 1,000 interviews, but in order to account for non-contact and non-response a gross sample of 1,486 households was selected. In total, 143 PSUs were drawn with probability proportional to the square root of the number of households in each stratum. Within each PSU, 5 households (+3 reserve households) were randomly selected in urban areas. In rural environments, on the other hand, in an effort to mini- mize transportation cost, 10 households (+4 reserve households) were randomly drawn. Finally, within each selected household one respondent was randomly drawn by means of the Kish grid out of the pool of eligible household members. To be eligible for being interviewed, household members needed to be 18 years or older and to have some role in managing either personal or household �nances. - 10 - 31. The �eldwork took place between August and October 2012. To implement the survey, the WB contracted the Center of Sociological Research ‘Zerkalo’ (survey �rm). After 5 days of intense training and piloting the data was collected by a team of 36 enumerators and 4 supervisors. To achieve 1,000 completed interviews, 1,402 (of the gross sample of 1,486) adults needed to contacted. Out of these 25 percent (346 adults) were ineligible for various reasons (e.g. dwelling was destroyed, dwelling was not found, etc.). Of those considered as being eligible, only 5 percent (56 adults) refused to participate in the survey which yielded an overall response rate of 95 percent. 32. All graphs and regression analysis in this report are based on weighted data. To adjust for varying probabilities of selection and non-response and to reflect gender and age distributions in the entire population aged 18 years and over, proper weights were calculated for each respondent, and then applied to the data set. 33. The sample characteristics of the FCCP survey are largely comparable to the 2009 Living Standards Measurement Study (LSMS). The households dis- tribute roughly across the same regions (see appendix, �g. 70). The higher fraction of respondents living in the Kathlon region in the 2009 LSMS than in the FCCP sample can be explained by the fact that the Kathlon population was intentionally oversampled in the LSMS. However, the share of urban and rural household is essentially the same in both studies (see appendix, �g. 68). This is important for the validity of the results as the sampling strategy takes the household’s environment into account. Even though the main focus of the FCCP survey is not to measure living standards, the representation of different household income groups is also fairly similar in both surveys. The Sample Characteristics in Comparison with other Data Sources 34. The sample characteristics of the FCCP survey are largely comparable to the 2009 Living Standards Measurement Study (LSMS). The households distribute roughly across the same regions (see app. �g. 67). The higher fraction of respondents living in the Kathlon region in the 2009 LSMS than in the FCCP sample can be explained by the fact that the Kathlon population was intentionally oversampled in the LSMS. However, the share of urban and rural household is essentially the same in both studies (see appendix, �g. 68). This is important for the validity of the results as the sampling strategy takes the household’s environment into account. Even though the main focus of the FCCP survey is not to measure living standards, the representation of different household income groups is also fairly similar in both surveys (see app. �g. 69). 35. The household composition is largely comparable, too. The intra-household composition of males and females is the identical (see app. �g. 70). Concerning age, the LSMS includes on average ‘older’ household members, whereas the FCCP survey includes younger Tajiks (see app. �g. 71). 36. The sample’s key demographics resemble the ones of the 2010 census data. The FCCP survey’s sampling design is based on the 2010 census data, which state that 73.5% of all Tajik house- holds are in an urban environment and 26.5% are in rural areas. Concerning the gender distribution, according to the 2010 census in Tajikistan, 50.4% of the population is males and 49.6% females.6 Both the LSMS and the FCCP survey are close to these values (see app. �gs. 68 and 70). In order to achieve this the data are weighted. Hence, relatively underrepresented groups receive higher weight in below analyses in order to mirror the Tajik population as precisely as possible. In return, relatively over-represented sub-populations receive less weight. - 11 - 3. Financial Capability Background Information: Financial Capability Financial Capability is the internal capacity to act in one’s best �nancial interest, given socioeconomic environmental conditions. It therefore encompasses the knowledge, attitudes, skills, and behaviors of consumers with regard to managing their resources and understanding, selecting, and making use of �nancial services that �t their needs. 3.1. Knowledge of �nancial Concepts and Products 3.1.1. Concepts 37. A certain level of knowledge of fundamental �nancial concepts and products and basic numeracy skills are needed in order to effectively participate in �nancial markets. To evaluate people’s awareness of �nancial concepts and their basic numeracy skills all respondents were exposed to a battery of 8 quiz-like questions with true or false choices. These include questions where respondents were required to do simple calculations, to show their understanding of inflation, interest rates, compound interest, risk diversi�cation, and the main purpose of insurance products. Knowledge of inflation, interest rates and compound interest is crucial for informed savings and borrowing decisions, whereas the last two questions are vital for being able to take informed investment decisions and to protect against risks. 4.3 − 4.5 4.1 − 4.3 3.9 − 4.1 3.7 − 3.9 3.5 − 3.7 3.3 − 3.5 0 − 3.3 Figure 1: Average �nancial literacy score in regions. Lowest possible score: 0, highest possible score: 7, see �g. 3. - 12 - 30 Simple Division Compare Bargain Purpose of Insurance 20 Percent Compound Interest Risk Diversification 10 Simple Interest Inflation Bank Bailout 0 0 20 40 60 80 100 0 1 2 3 4 5 6 7 Figure 2: Percentage of respondents who correctly answered Figure 3: Discrete frequency distribution of �nancial �nancial literacy questions. literacy scores. 38. On average, Tajiks were able to correctly answer 4 out of 7 �nancial literacy related questions. Slightly more than half of the population was able to provide correct answers to at least 4 questions. Nevertheless, only around 12 percent of the population was able to provide correct answers to at least 6, even less than 2 percent answered all 7 �nancial literacy questions correctly (see �g. 3). A high fraction of all adults, around 13 percent, was not even able to provide more than 2 correct answers. 39. The majority of the population possesses the necessary skills to perform basic �nancial calculations, but lacks the speci�c knowledge required to make sensible savings and investment decisions. A closer look at the type of �nancial knowledge questions people were able to answer correctly reveals that almost all Tajiks are able to perform simple divisions and about three quarters are comfortable with simple calculations in order to identify better bargains (see �g. 2). Roughly the same fraction of the population understands the main purpose of insurance policies (71.5 percent). About 3 in 5 Tajiks correctly answered the compound interest question. More worrisome, about almost half of the population is not aware of the bene�ts of diversifying risks by holding stocks from different companies instead of only. The most concerning fact is that only 2 out of 5 Tajiks demonstrate basic understanding of simple interest rates and only a �fth of the population knows how inflation affects their savings. 40. Moreover, awareness of deposit insurance is almost inexistent. In August of 2011, a law was passed creating a deposit insurance scheme for losses by individuals that result from the bankruptcy or other liquidation of a bank or micro�nance deposit organization. Banks do not yet appear to provide appropriate information to their customers about the existence of the Deposit Insurance Fund and the nature and extent of insurance on interest-bearing accounts7 . Therefore, it is not surprising that when asked about the amount which would be covered by the Deposit Insurance Fund in case a bank goes bankrupt, only around 4 percent of the population was able to state the correct answer (Tajikistani somoni (TJS) 7,000). 7 Source: ibid. - 13 - Background Information: Financial Literacy Quiz Question 1 Imagine that �ve brothers are given a gift of TJS 1,000. If the brothers have to share the money equally, how much does each one get? Question 2 Now, imagine that the �ve brothers have to wait for one year to get their share of the TJS 1,000 and inflation stays at 10%. In one year’s time will they be able to buy: • More with their share of money than they could today • The same amount • Less than they could buy today • It depends on the types of things that they want to buy (do not read out) Question 3 Suppose you put TJS 100 into a savings account with a guaranteed interest rate of 2% per year. You don’t make any further payments into this account and you don’t withdraw any money. How much would be in the account at the end of the �rst year, once the interest payment is made? Question 4 How much would be in the account at the end of �ve years? Would it be: • More than TJS 110 • Exactly TJS 110 • Less than TJS 110 • It is impossible to tell from the information given (do not read out ) Question 5 Let’s assume that you saw a TV-set of the same model on sales in two different shops. The initial retail price of it was TJS 1,000. One shop offered a discount of TJS 150, while the other one offered a 10% discount. Which one is a better bargain, a discount of TJS 150 or 10%? • A discount of TJS 150 • They are the same • A 10% discount Question 6 Which of the following statements best describes the primary purpose of insurance products? • To accumulate savings • To protect against risks • To make payments or send money • Other (If other, specify) Question 6 Suppose you have money to invest. Is it safer to buy stocks of just one company or to buy stocks of many companies? • Buy stocks of one company • Buy stocks of many companies Question 7 If a citizen has a deposit in any bank of Tajikistan and the bank goes bankrupt, do you know how much is the biggest payout from the government. • The Government did not pay any compensation for bank deposits. • TJS 7,000 • TJS 10,000 • TJS 15,000 • The Government will pay the compensation for bank deposits - 14 - 41. Financial knowledge and awareness levels are a signi�cant challenge in Tajik- istan, as for many countries at a similar income level. Tab. 1 reveals that Tajiks belong in many areas to the least �nancially knowledgeable population in Europe and Central Asia. Although the fraction of people with basic understanding of compound interest is strikingly high compared to other countries in the region, Tajiks demonstrate not only in absolute, but also in relative terms substantial knowledge gaps, in particular with regard to inflation and consumer protection laws (deposit insurance). Table 1: International comparison of knowledge of basic �nancial concepts (in % of adults). Country8 Inflation Simple Interest Compound Interest Deposit insurance Bosnia & Herzegovina (2011) 58 64 22 7 Bulgaria (2010) 46 38 39 20 Azerbaijan (2009) 46 52 46 6 Russia (2009) 54 25 40 14 Tajikistan (2012) 17 35 56 4 Romania (2010) 43 32 24 12 4 4 Financial Literacy Score Financial Literacy Score 3 3 2 2 1 1 3.9 4.2 3.9 4.2 0 0 Not Head, n=492 Is Head, n=508 Female, n=576 Male, n=424 Figure 4: Average �nancial literacy score, by respondent Figure 5: Average �nancial literacy score, by respondents’ being head of household. gender. 4 4 Financial Literacy Score Financial Literacy Score 3 3 2 2 1 1 3.9 4.3 3.2 3.9 4.3 0 0 Rural, n=570 Urban, n=430 Primary, n=57 Secondary, n=539 Tertiary, n=404 Figure 6: Average �nancial literacy score, by respondents’ Figure 7: Average �nancial literacy score, by respondents’ urbanization. education. 42. Some parts of the population are less �nancially knowledgeable than others. As would be expected, being head of the household with decision making competence is associated with better knowledge of �nancial concepts (see �g. 4). More than two thirds of all households heads are men (73 percent). The gender gap revealed in �g. 5 is therefore not surprising. Married respondents or those who live in an informal union also 8 Source: World Bank FCCP Survey, sorted by average position of all categories. In contrast to the number for Tajikistan, the results of the other surveys mentioned in table 1 are not weighted. However, this is expected to have only little influence on the results. - 15 - demonstrate better knowledge of �nancial concepts than those who are neither married nor living.with a partner. Accounting for the effects of several socio-demographic factors at the same time – by means of regressions – it turns out that the gender gap is not statistically signi�cant while being the head of the household has a signi�cant effect on being �nancially literate. Similarly, neither age nor income seem to have a signi�cant effect on being knowledgeable about �nancial concepts. 43. Lower �nancial knowledge is in particular associated with rural areas and lower educational attainment. Those who live in the Kathlon region (see �g. 1) ap- pear to belong to the group within the population, which lacks most �nancial knowledge. However, the regional differences concerning the average �nancial knowledge scores are not statistically signi�cant. The knowledge gap which can be observed between rural and urban populations (�g. 6) is, however, statistically signi�cant. Rural dwellers are much less �nancially knowledgeable than those living in urban areas. As would be expected, being �nancially knowledgeable strongly correlates with higher educational attainment. Those parts of the population whose highest educational attainment is completed pri- mary school (see �g. 7) have economically and statistically signi�cant lower �nancial literacy scores than those with completed tertiary education. 80 Simple Interest Compound Interest Risk Diversification Inflation 60 40 20 38 56 49 16 28 59 60 25 36 65 63 19 20 23 73 23 0 None, n=723 < 1, n=62 2−12, n=115 >12, n=34 Figure 8: Selected average �nancial literacy score, by approximate household debt measured in number of monthly incomes. 44. Adults with high levels of debt lack speci�c �nancial knowledge. As 4 shown in �g. 9, the average scores on Financial Literacy Score the �nancial literacy quiz-type questions 3 for populations with outstanding debts, which account for more than their an- nual income hardly differ from those with 2 lower levels of indebtedness. Respon- dents with outstanding debt greater than 1 their annual income are, however, much less likely to have a basic understanding 4.0 3.9 4.1 4.0 of simple and compound interest rates 0 None, n=751 < 1, n=64 2−12, n=115 >12, n=37 Figure 9: Average �nancial literacy score, by approximate (see �g. 8). Although it is not clear to household debt measured in number of monthly incomes. what extent their high level of indebted- ness is due to their limited understanding of loan costs, this knowledge gap makes them - 16 - more vulnerable to economic shocks, such as interest rate increases, and ultimately ad- versely affects their ability to repay their outstanding debts. 3.1.2. Products 4.3 − 4.5 4.1 − 4.3 3.9 − 4.1 3.7 − 3.9 3.5 − 3.7 3.3 − 3.5 0 − 3.3 Figure 10: Average number of �nancial products known by respondents in different regions. Lowest possible number: 0, highest 7, see �g. 12. 20 None Exchange Office 15 Money Transfer Percent Microfinance Organizations 10 Commercial Bank Insurance 5 Other Non−Banking Broker 0 0 20 40 60 80 0 1 2 3 4 5 6 7 Figure 11: Percentage of respondents who know about the Figure 12: Discrete frequency distribution of knowledge services offered by �nancial institutions. about services offered by the �nancial institutions. 45. MTOs and currency exchange offices are widely known, but awareness of broader �nancial services provided by commercial banks, MFOs, and other FIs is much more limited. More than 80 percent of the population is familiar with services provided by currency exchange offices. Around three quarters of the popula- tion are familiar with the services offered by MTOs. Services provided by MFOs and commercial banks are known by a bit more than half of the population. Similarly in- surance companies and the related services are known by around half of the population (see �g. 11). Only a third of the respondents indicates that they are familiar with the services offered by other NBFIs. Around 30 percent of the population are familiar with the products and services that brokerage houses provide. - 17 - 46. On average, citizens are familiar with services provided by 3.8 different types of FIs. As shown in �g. 12, a quarter of the population claims to know the services offered by all types of FIs. Although this number could be an indication that some parts of the population think they know more than they actually do, a signi�cant fraction – around a third of the population - admits that they are not familiar with the services offered by more than two different types of �nancial service providers. Of particular concern is that one out of ten Tajiks is not familiar with the services offered by any �nancial service providers. 47. Knowledge of �nancial products and services varies noticeably with location. Signi�cant knowledge gaps become apparent between Dushanbe’s population and resi- dents of other regions. Interestingly, residents of DRS, Kathlon, Sogd, and in particular GBAO appear to know on average more �nancial products and services than those who live in the capital. These regional differences do not only exist with respect to the num- ber of �nancial products, but also the type of products. Whereas Dushanbe’s population seems to be signi�cantly more knowledgeable with regard to currency exchange offices than GBAO’s population, residents of other regions indicate they are more familiar with services offered by commercial banks, MFOs, and MTOs (see �g. 10). 48. Having more �nancially dependent children and being out of the labor mar- ket is strongly associated with being less familiar with services provided by different types of FIs. Knowledge about �nancial services offered by different types of FIs is negatively correlated with having more economically dependent children (see �g. 13). This difference remains signi�cant even after controlling for other factors, such as household income. A simple explanation could be that the time adults spend on caring for their children limits available time for any other activities, including those which could raise their awareness of �nancial products. Fig. 14 shows that there are also knowledge disparities between groups with different employment status. Among all segments with different employment status, Tajiks who are out of the labor force are the least knowledgeable, in particular regarding MFOs and their services. Interestingly, even after accounting for the effects of income and other characteristics, the unemployed appear to be relatively more knowledgeable about currency exchange offices, MTOs, MFOs, and banks, which might reflect their higher dependence on remittances and the need to smooth consumption. 49. Another pattern which emerges is that better knowledge of �nancial products and services correlates with being able to read and write in Tajik and the regular use of different types of media. Not surprisingly, those who were able to read and write in Tajik demonstrated a greater knowledge of the different �nancial service providers and their offered products than illiterate Tajiks (see �g. 16). In particular, they showed signi�cantly better knowledge about products and services provided by banks, insurance companies, and brokerage houses. Likewise, people who try to stay informed by using different types of media regularly, such as TV, radio, internet, and national or local newspapers, are also more likely to show greater familiarity with products offered by different types of �nancial service providers, except currency exchange offices and MTOs (see �g. 17). - 18 - Exchange Office Money Transfer Operator MFOs Banks Insurance Other NBFIs Broker 80 Percent 60 40 20 78 74 53 56 54 32 24 88 87 66 57 54 37 29 88 73 61 53 45 33 24 84 69 48 51 41 29 28 83 66 42 42 35 23 18 77 78 44 64 29 24 15 0 0, n=252 1, n=183 2, n=216 3, n=158 4, n=121 5 or more, n=70 Figure 13: Percentage of respondents knowing about services offered by the �nancial institutions, by the number of economically dependent children within the household. Exchange Office Money Transfer Operator MFOs Banks Insurance Other NBFIs Broker 100 - 19 - 80 Percent 60 40 20 91 87 71 65 60 43 31 89 75 63 61 42 28 22 82 77 60 52 51 37 31 77 66 41 46 39 23 18 0 Formal, n=257 Informal, n=112 Self Empl., n=153 Not Work, n=439 Figure 14: Percentage of respondents knowing about services offered by the �nancial institutions, by employment status. Exch. Off. Money Transf. MFOs Banks 100 100 Exch. Off. Money Transf. MFOs Banks Insurance Other NBFIs Broker Insurance Other NBFIs Broker 80 80 Percent Percent 60 60 40 40 20 20 79724946392520 90776365584131 71664544332319 89785958533527 0 0 - 20 - Primary and Secondary, n=596 Tertiary, n=404 Low: 0−4, n=301 High: 5−8, n=699 Figure 15: Percentage of respondents knowing about services offered by the �nancial Figure 16: Percentage of respondents knowing about services offered by the �nancial institutions, by respondents’ education. institutions, by respondents’ �nancial literacy. Exchange Office Money Transfer Operator MFOs Banks Insurance Other NBFIs Broker 100 80 Percent 60 40 20 61 65 29 34 28 22 25 69 64 37 42 29 21 17 86 69 52 41 39 21 17 90 82 55 48 43 22 12 85 79 67 74 65 47 41 98 81 70 74 65 59 39 0 None, n=27 1, n=188 2, n=206 3, n=244 4, n=273 5, n=62 Figure 17: Percentage of respondents knowing about services offered by the �nancial institutions, by respondents’ regular media consumption. 3.2. Financial Attitudes and Behavior 3.2.1. Dimensions of Financial Capability 50. Knowledge and awareness of ba- sic �nancial concepts and products Not Overspending does not necessarily translate into Non−Impulsiveness sound �nancial behavior. To iden- Achievement Orientation tify the role attitudes play in individu- Using Information and Advice als’ �nancial decisions and to see how Foresightedness attitudes translate into �nancial behav- Living within Means ior, the survey instrument used contains Controlled Budgeting questions on different aspects (compo- Saving nents) of �nancial capability, including Planning for one’s own Future attitudes/motivations and behaviors. 0 20 40 60 80 51. In the Tajiks dataset, 9 main com- 100 Figure 18: Average �nancial capability scores. ponents of �nancial capability can be identi�ed, some of which refer to attitudes or motivations, and others others to behaviors. Each component is measured through a combination of the relevant questions, which are identi�ed by using a statistical method called factor analysis. Factor analysis is a data reduction method that �nds a small number of linear combinations of the variables that also explain most of the variance in the data. The method is used to aggregate the variables that measure different nuances of the same component, in order to obtain a single indicator (or score) for that component. Each component score ranges between 0 (least capable) to 100 (most capable). 52. Five components measure �nancial capability related to behaviors such as budgeting, not overspending, living within means, saving, planning for un- expected and old age expenses. More speci�cally, ‘budgeting’ measures the extent to which people plan how to use their money and whether they stick to the plan. ‘Not overspending’ assesses whether people refrain from spending their income on things that are not necessary or on things they cannot afford. ‘Living within means’ measures the level of borrowing, speci�cally whether people borrow to buy food and other essentials, and whether people who have money left over after buying essentials save it – instead of only spending it on non-essentials. ‘Saving’ measures whether people see themselves as trying to save for the future, trying to save for emergencies, and trying to save even if a small amount. Eventually, ‘planning for unexpected and old age expenses’ indicates whether people could cover an unexpected expense equivalent to a month’s income and whether they worry about it, and if they have provisions that could cover their old age expenses in full. 53. Four �nancial capability components refer to attitudes and motivations such as attitudes towards the future, (non-)impulsiveness, attitudes towards infor- mation seeking, and achievement orientation. In particular, ‘attitudes towards the future’ measure whether people agree or disagree with statements such as ‘I live more for the present day than for tomorrow’, ‘The future will take care of itself’, and ‘I only focus on the short term’; the measure of ‘non-impulsiveness’ is determined by whether people agree or disagree with statements about impulsivity (being impulsive, saying things without giving them too much thought, and doing things without thinking - 21 - them through); ‘attitudes towards information’ is a combination of getting information and advice before making �nancial decisions, learning from others, and being disciplined; and ‘achievement orientation’ measures to what extent people agree with statements on having aspirations, working hard to be the best, and always looking for opportunities to improve one’s own situation. 54. Tajiks demonstrate high levels of �nancial capability in most areas, although they score higher with regard to their attitudes and motivations than their �nancial behaviors. Fig. 18 shows all �nancial capability scores in increasing order. As can be seen, the highest average score (94) is obtained for �nancial behavior which is characterized by not overspending. The next 4 highest scores on the other hand measure attitudes and motivations (in decreasing order: non-impulsiveness, achievement orienta- tion, using information and advice, attitudes towards the future, living within means, and budgeting). Compared to these �nancial capability aspects, Tajiks score slightly worse with regard to �nancial behaviors such as living within means and budgeting Tajiks score slightly worse. By far the lowest average scores are found for �nancial behaviors that relate to planning for one’s own future: saving (66) and planning for unexpected and old age expenses (47.5). These low scores are worrying given their implications for long-term wellbeing. 55. The following subsections present detailed results for each �nancial capability area and analyze differences across key socio-demographic indicators. The analysis is presented by focusing �rst on components that measure attitudes/motivations and then on components that measure �nancial behavior. 3.2.2. Financial Capability Related to ‘not overspending’ Behavior 90 − 100 80 − 90 70 − 80 60 − 70 50 − 60 40 − 50 0 − 40 Figure 19: Average �nancial capability score ‘not overspending’, in different regions. 56. Compared to other aspects of �nancial capability, Tajiks are most capable of not overspending and prioritizing their expenses on essentials and things they can afford. Only a very small fraction of the population (between 2 and 4 percent) - 22 - reports that they sometimes or regularly spend their income on non-essentials before they have bought food and other necessary items, or that they buy non-essential things that they cannot afford. 57. Compared to Dushanbe’s population residents from other regions are less likely to overspend and more capable of prioritizing their expenses on es- sentials. As can be seen in �g. 19, people who live in Dushanbe are less capable of prioritizing their spending on essentials than residents from other regions, in particular those from Kathlon and GBAO. These differences in the not overspending score appear to be statistically signi�cant even after controlling for other characteristics. 100 80 ’Do Not Overspend’ Average Score 60 40 20 96 94 93 0 TJS 1,500, n=264 n=561 n=168 Figure 20: Average �nancial capability score ‘not overspending’, by approximate monthly household income. 58. Interestingly, income does not matter for Tajiks’ ability to resist temptations of spending their income on things that are not necessary or on things they cannot afford. Fig. 20 suggests that those groups of the population which earn the highest incomes compared to lower income groups are the least capable of not over- spending and in prioritizing spending on their needs rather than on their wants. These differences are, however, not statistically signi�cant. Moreover, household size and the number of dependent children appear to matter more in determining residents’ ability to refrain from overspending. In other words, Tajiks in larger households and with more �nancially dependent children are less likely to overspend and to allocate their income to non-essential expenses. - 23 - 3.2.3. Financial Capability Related to ‘living within means’ Behavior 90 − 100 80 − 90 70 − 80 60 − 70 50 − 60 40 − 50 0 − 40 Figure 21: Average �nancial capability score ‘living within means’, in different regions. 100 80 ’Live with means’ Average Score 60 40 20 81 81 91 0 TJS 1,500, n=264 n=549 n=167 Figure 22: Average �nancial capability score ‘living within means’, by approximate monthly household income. 59. The high score in ‘living within means’ reflects the fact that a large majority of Tajiks could borrow more than they currently do, or have not borrowed at all (83 percent). However, 69 percent run short of money for food and other essentials (either regularly or sometimes), 22 percent report borrowing money or using credit to buy these essential items. 60. As would be expected, people who live on lower incomes tend to struggle more with living within their means. Faced with small and erratic income streams low income populations are less likely to live within their means than individuals living in high income households (�g. 22). Likewise, Tajiks with a greater number of household members also experience signi�cant difficulties living within their means. - 24 - 3.2.4. Financial Capability Related to ‘budgeting’ Behavior 90 − 100 80 − 90 70 − 80 60 − 70 50 − 60 40 − 50 0 − 40 Figure 23: Average �nancial capability score ‘budgeting’, in different regions. 80 60 Average Score ’Budgedting’ 40 20 85 85 76 84 77 79 0 Formal Informal Self−Employed Retired Unemployed Out of Labor n=233 n=107 n=128 n=111 n=57 n=201 Figure 24: Average �nancial capability score ‘budgeting’, by respondent’s employment status. 61. Tajiks are also quite capable of mak- ing and keeping plans for how to 80 spend their money and in keeping to the plan. In Tajikistan, 82 percent 60 of the population plan how to spend the Average Score money they receive, either sometimes (23 ’’ percent) or always (59 percent). Almost 40 all of the people who plan (98 percent) also say they stick to the plan at least 20 sometimes. Around half of those who 85 81 plan always keep to the plan (51 percent). 0 Other, n=159 Married/Cohabit, n=677 62. In terms of budgeting, married res- Figure 25: Average �nancial capability score ‘budgeting’, by respondent’s cohabiting status. idents, as well as those who are less educated, are the least �nancially - 25 - capable segments of the population. The difference between married and single respondents in terms of their ability to set up a budget and to stick to it, is statistically signi�cant even after controlling for other characteristics (see �g. 25). Similarly, people with completed tertiary education are also signi�cantly more capable of budgeting than those who have completed primary education at the most. 63. Residents living at the lowest incomes levels are more challenged in terms of budgeting than high income groups. Poor people’s income is not only small but also irregular, making it much more difficult for them to adhere to current of future �nancial plans than for people with high and more predictable income streams. Even though the fraction of highest income earners is larger in Dushanbe, residents living in DRS, Kathlon, Sogd, and in particular in GBAO are relatively more capable of budgeting. 3.2.5. Financial Capability Related to ‘saving’ Behavior 90 − 100 80 − 90 70 − 80 60 − 70 50 − 60 40 − 50 0 − 40 Figure 26: Average �nancial capability score ‘saving’, in different regions. 80 80 60 60 Average Score Average Score ’Saving’ ’Saving’ 40 40 20 20 60 67 71 64 76 0 TJS 1,500, 0 n=264 n=557 n=162 Not Save, n=790 Did Save, n=195 Figure 27: Average �nancial capability score ‘saving’, by Figure 28: Average �nancial capability score ‘saving’, by approximate monthly household income. respondent’s childhood saving behavior. 64. Compared to other �nancial capability areas, Tajiks are least capable in terms of saving. Between 48 and 53 percent of Tajiks strongly agree with statements such as “I try to save money for the future�, “I try to save money regularly, even if it is only - 26 - a little�, and “I try to have some provision for emergencies or unexpected expenses�. However, between 16 and 21 percent disagree with these statements. 65. Low income groups and those with fewer �nancially dependent children tend to be less capable of saving money. As would be expected, the ability to save is constrained by available resources. Those who live on the highest incomes are also the ones who are signi�cantly more inclined to put money aside regularly and to save money for the future, in comparison to the lowest income group. The difference remains signi�cant when other factors are taken into account (see �g. 27). Moreover, even after controlling for income and other characteristics, those with more �nancially dependent children also appear to be more capable of saving money. 66. Those who learned savings practices as children appear to be more capable of putting money aside as adults. As shown in �g. 28, starting to save at an early age has value, as these respondents score where those who already saved as a child score on average signi�cantly higher with respect to their savings behavior today as their counterparts who did not save in their childhood. 67. Dushanbe’s residents demonstrate less capable savings behavior than popu- lations of any other regions. Fig. 26 illustrates that residents from DRS, Katholon, Sogd, and in particular GBAO outperform people who live in the capital with regard to their ability to save. These differences in the saving scores are statistically signi�cant even after controlling for the effect of income and other characteristics. 3.2.6. Financial Capability Related to ‘planning for unexpected and old age expenses’ Behavior 90 − 100 80 − 90 70 − 80 60 − 70 50 − 60 40 − 50 0 − 40 Figure 29: Average �nancial capability score ‘planning for unexpected and old age expenses’, in different regions. - 27 - ’Plan for Oldage / Unexpected Expenses’ 60 ’Plan for Oldage / Unexpected Expenses’ 50 40 Average Score 40 Average Score 30 20 20 10 41 50 60 46 53 0 Age<35, 3555, 0 n=213 n=520 n=228 Not Save, n=768 Did Save, n=192 Figure 30: Average �nancial capability score ‘planning for Figure 31: Average �nancial capability score ‘planning for unexpected and old age expenses’, by unexpected and old age expenses’, by respondent’s age. respondent’s childhood saving behavior. 68. A worrying �nding is that only slightly more than a third of the population (36 percent) could cover an unexpected expense equivalent to one month’s income without borrowing. 54 percent of those who could not cover such an expense without borrowing say they are ‘very worried’ about it, while 43 percent report to be ‘a bit worried’, and 3 precent are not worried at all. Similarly concerning is the �nding that only 13 percent have made provisions that fully cover or will cover expenses in old age. Moreover 26 percent have made provisions that provide partial coverage, and 61 percent have no provision at all. 69. Unsurprisingly, retirees and higher income groups score better in planning for unexpected and old age expenses. Although the capability to plan for unexpected and old age expenses seems to strongly increase with age (see �g. 30), this effect turns insigni�cant once other characteristics are taken into account. One’s retirement status and income are more signi�cant indicators than age. Those who already retired are, as one would expect, more likely to have made provisions that cover their old age expenses than those who are out of labor force (and not retired). Similarly, low income seems to signi�cantly constrain people’s ability to plan for unexpected and old age expenses. 70. Starting early pays off again since having saved as a child strongly correlates with a higher ability to plan for unexpected and retirement age expenses. As shown in �g. 31, having saved as a child is associated with a higher ‘planning for unexpected and old age expenses’ score. This relationship remains signi�cant after controlling for other characteristics. 71. Interestingly, Dushanbe’s residents are not only less capable of saving but are also less capable of planning for unexpected and old age expenses. In particular, people who live in GBAO and Sogd score signi�cantly higher in this �nancial capability area than Dushanbe’s population, see �g. 29. - 28 - 3.2.7. Financial Capability Related to ‘(non-)impulsiveness’ 90 − 100 80 − 90 70 − 80 60 − 70 50 − 60 40 − 50 0 − 40 Figure 32: Average �nancial capability score ‘(non-)impulsiveness’, in different regions. 72. Tajiks on average describe them- 100 selves as not being impulsive. Be- tween 89 and 94 percent disagree with 80 statements such as “I am impulsive�, “I ’Not Being Impulsive’ Average Score say things before I have thought them 60 through� and “I do things without giv- ing them much thought�. 40 73. Younger people are tend to be im- pulsive. Fig. 33 does not reveal any 20 major differences between different age 93 94 93 groups. Nevertheless, once other char- 0 Age<35, 3555, n=225 n=541 n=234 acteristics are taken into account, in- Figure 33: Average �nancial capability score ‘(non-)impulsiveness’, by respondent’s age. creasing age is associated with signi�- cantly lower levels of impulsiveness, but the more age advances the more impul- sive people seem to be. - 29 - 3.2.8. Financial Capability Related to ‘achievement orientation’ 90 − 100 80 − 90 70 − 80 60 − 70 50 − 60 40 − 50 0 − 40 Figure 34: Average �nancial capability score ‘achievement orientation’, in different regions. 74. Achievement orientation is another 100 strong area for the Tajik popula- tion. Almost all Tajiks (between 94 and 80 ’Achievement Orientation’ 96 percent) agree with statements such as “I have many aspirations�, “I work hard Average Score 60 to be among the best at what I do�, “I al- ways look for opportunities for improving 40 my situation�. 75. People with more economically de- 20 pendent children tend to be more 88 96 achievement-oriented. Neither in- 0 Not Save, n=803 Did Save, n=196 come, age, or occupation appear to have Figure 35: Average �nancial capability score ‘achievement orientation’, by respondent’s childhood saving a signi�cant effect on how achievement- behavior. oriented people are. In contrast, what seems to matter is whether someone be- gan saving as a child (see �g. 35). In terms of one’s level of achievement-orientation, those who saved as children, outper- form their counterparts who did not save in childhood. Likewise, a higher achievement- orientation score signi�cantly correlates with an increasing number of economically de- pendent children. These differences remain statistically signi�cant even after accounting for the effects of other characteristics. - 30 - 3.2.9. Financial Capability Related to ‘seeking information and learning from others’ Attitudes 90 − 100 80 − 90 70 − 80 60 − 70 50 − 60 40 − 50 0 − 40 Figure 36: Average �nancial capability score ‘seeking information and learning from others’, in different regions. 80 80 ’Learning and Information’ 60 Average Score 60 Average Score ’’ 40 40 20 20 86 89 87 86 87 0 Age<35, 3555, 0 n=225 n=541 n=234 Other, n=200 Married/Cohabit, n=799 Figure 37: Average �nancial capability score Figure 38: Average �nancial capability score ‘foresightedness’, by respondent’s cohabiting ‘foresightedness’, by respondent’s educational status. attainment. 76. Tajiks are also inclined to use information and learn from others. 95.3 percent agree with the statement that they always get information or advice when they have an important �nancial decision to make. On the other hand, 79 percent mention that they learn from the mistakes other people make in managing their money. In addition, 95 percent indicate that the statement “I am very disciplined when it comes to managing money� describes them well. 77. The overall score for seeking information and advice is lower for people living in Dushanbe and for those who use fewer types of media regularly. Similar to results for other �nancial capability areas, Dushanbe’s population seems to lag behind residents from other regions with regard to their attitude towards seeking information and learning from others. As would be expected, a greater regular use of different types of media appears to be related to better attitudes towards seeking information and advice. 78. Whether one is inclined to seek information also depends on their age and if the person saved as a child. Scores for ‘seeking information and advice before taking - 31 - a �nancial decision’ increase as age advances (see �g. 37). However, the older people are, the less inclined they are to seek information on �nancial matters. In comparison to those who saved in their childhood, those who did not are also less inclined to seek information before taking �nancial decisions (see �g. 38). 3.2.10. Financial Capability Related to ‘foresightedness’ 90 − 100 80 − 90 70 − 80 60 − 70 50 − 60 40 − 50 0 − 40 Figure 39: Average �nancial capability score ‘foresightedness’, in different regions. 80 80 60 ’Foresightedness’ 60 Average Score Average Score ’’ 40 40 20 20 78 86 87 84 85 0 0 Other, n=200 Married/Cohabit, n=799 Primary, n=57 Secondary, n=539Tertiary, n=404 Figure 40: Average �nancial capability score Figure 41: Average �nancial capability score ‘foresightedness’, by respondent’s cohabiting ‘foresightedness’, by respondent’s educational status. attainment. 79. The vast majority of Tajiks is quite farsighted. In terms of attitudes towards the future, a large majority of Tajiks states that they do not live more for the present day than for tomorrow� (86.6 percent), nor do they “focus on the short term� (76.4 percent). Moreover 71.3 percent disagree that ‘The future will take care of itself’. 80. Unmarried Tajiks and those belonging to lower income groups think and care less about the future. Fig. 40 reveals a gap between single and married people in their propensity to think about the future. The difference remains signi�cant once other factors are accounted for. Likewise, high income groups tend to be signi�cantly more farsighted in comparison to low income populations. This may be an indication that - 32 - successfully managing day-to-day �nances and solving immediate problems requires full attention, and does not leave much room for longer term considerations. 3.3. Relationship between Financial Knowledge and Financial Attitudes and Behavior 81. Lack of �nancial knowledge can be a constraint for those �nancial capability aspects which are related with long term planning and borrowing. Tajiks who answered more �nancial literacy related questions correctly tend to be more capable with respect to living within their means, saving, and planning for unexpected and old age expenses than those with lower scores. However, those with higher �nancial literacy scores appear to be less capable of budgeting and less inclined to refrain from overspending. This result may at least partly be explained by the fact that knowledge of �nancial concepts such as inflation, simple and compound interest is more relevant for long term planning than it is for day-to-day-money management. - 33 - 4. Financial Inclusion 82. The �nancial sector in Tajikistan has been growing rapidly for several years, before slowing its pace of expansion in 2008, coinciding with the global �nan- cial crisis.9 To get a better understanding of the current state of �nancial inclusion and the ways Tajiks save, borrow, make transactions, and manage risks, the FCCP survey also asked about the type of products/services and �nancial service providers, which Tajiks access to or use. 4.1. Access and Use of Financial Services 4.1.1. Products 1 − 1.1 .9 − 1 .8 − .9 .7 − .8 .6 − .7 0 − .6 Figure 42: Average number of �nancial products used in regions, out of ‘Insurance’, ‘Formal Credit’, ‘Informal Credit’, ‘Formal Saving’, ‘Informal Saving’, ‘Checking/Money transfers’, and ‘Investments’. 40 Insurance Investments 30 Checking / Money Transfer Percent Formal Credit 20 Informal Credit 10 Informal Saving Formal Saving 0 0 5 10 15 20 25 0 1 2 3 4 5 6 7 8 9 Figure 43: Percentage of respondents with different forms of Figure 44: Discrete frequency distribution of �nancial �nancial products. product usage. 9 Source: Diagnostic Review of Consumer Protection and Financial Literacy, World Bank, forthcoming. - 34 - 83. Despite the expansion of the �nancial sector in the past decade, large parts of the Tajik population are still �nancially excluded and do not have access to any �nancial products. Compared to all available types of �nancial products, Tajiks have greatest access to insurance policies. Around a quarter of the population (26 percent) has access to / uses insurance products (see �g. 43). The insurance category thereby covers two types of insurance policies: (i) private health, life or income replace- ment insurance, as well as (ii) general insurance, such as car insurance or household content insurance. The insurance penetration rate is mostly driven by the latter type of insurance, which is currently held by around 25 percent of the population. This rather high insurance penetration rate can be explained by the existence of several types of compulsory insurance, including motor third party liability, �re and renters’ insurance of state-owned property. The other type of insurance which is critical for mitigating risks that relate to personal health or to one’s livelihood is only used by approximately 2 percent of the population.10 This information is consistent with the Global Findex data, which showed that between mid 2011 and mid 2012 only around 1 percent of adults personally paid for health insurance.11 84. Investment products which help to mitigate old age risks are slightly less relevant than insurance policies. This category of �nancial products mostly com- prises private pension products. Only one in �ve adults (19.2 percent) is currently able or willing to manage risks related to old age through accessing this type of �nancial products. 85. Compared to other �nancial services, money transfer services are more fre- quently used. Migrant remittances are an important factor for the Tajik economy. In 2011, Tajikistan received remittances equivalent to 31 percent of GDP.12 Consequently, 15 percent of adults currently use basic �nancial services related to sending money and, more important, to receiving it, such as those offered by Western Union and other MTOs. This category also includes checking accounts. However, given that Global Findex data suggest that only 2.5 percent have access to an account, no matter which type, it is legitimate to assume that they only play an insigni�cant role. 86. Only around a tenth of the population (12.2 percent) has an outstanding loan from a formal �nancial institution. The formal credit category comprises loans from a commercial bank, mortgages, credit cards, and credit from MFOs (e.g Imon, Humo, etc.). According to data from Global Findex, however, approximately 5 percent of the Tajik population has borrowed from a formal �nancial institution. The comparability of these indicators is limited, especially due to the fact that the 2 surveys used different reference periods. Whereas the FCCP survey asked for outstanding loans from different formal FIs at the time the interviews took place, Global Findex asked if a loan from a formal FI was taken out between mid 2010 and mid 2011. It is notable, however, that the FCCP estimate is higher than the one from Global Findex, since it also includes loans which were taken out prior to mid 2010, but have not been repaid yet. 87. Around one out of ten Tajiks (8.8 percent) has borrowed informally from either informal money lenders or relatives/friends. On the other hand, Global Findex data show that around a quarter of the adult population borrowed from friends or family and around 3 percent from private lenders. Despite the different reference 10 Note that some individuals hold both types of insurance products. Thereby the sum for single insurance products differs from the percentage of holding any insurance. 11 Source: Global Findex data 12 Source: Diagnostic Review of Consumer Protection and Financial Literacy, World Bank, forthcoming. - 35 - periods, the discrepancy between these speci�c numbers may be due to the wording of the questions. Unlike Global Findex, the FCCP survey only asked for money borrowed from moneylenders, friends or family which needs to be repaid. The relatively quite high number of loans from family and friends suggested by Global Findex might be an indication that money borrowed from friends or family may not always have to be repaid. 88. Neither formal nor informal savings mechanisms are widely used in Tajik- istan. Formal savings include deposit accounts and savings at micro deposit institutions. (e.g. Bovari va Hamkori, etc.). Even less than 1 percent (0.7 percent) of adults reported to have formal savings. Informal savings mechanisms such as savings clubs or keeping money under the mattress are used by slightly more than 1 percent (1.2 percent). Keep- ing the different reference periods in mind, these savings numbers match well with the ones suggested by Global Findex data. According to Global Findex around 0.3 percent saved at a formal �nancial institution, while 1.9 percent saved by using a savings club between mid 2010 and mid 2011. 50 40 All Products (weighted mean=.8) Net of Insur. (weighted mean=.6) 40 30 30 Percent Percent 20 20 10 10 0 0 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 Figure 45: Discrete frequency distribution of �nancial Figure 46: Discrete frequency distribution of �nancial product usage. product usage. 89. On average Tajiks do not even hold 1 �nancial product (0.8). Without taking (compulsory) insurance policies into account, this number reduces to an average of just 0.5 products (see �gs. 45 and 46). In terms of insurance policies slightly more than one out of �ve Tajiks uses more than 1 �nancial product. Keeping in mind that the bulk of insurance policies is most likely induced by mandatory schemes, without accounting for insurance products, only slightly more than 10 percent of the adult population holds more than one �nancial product. However, the observed pattern does not change once insurance policies are taken into account or not. - 36 - 1.5 Number of Financial Products 1 .5 0.93 0.65 0.98 1.38 0.79 0.61 0 Formal Informal Self−Employed Retired Unemployed Out of Labor n=267 n=119 n=156 n=146 n=66 n=246 Figure 47: Average number of �nancial products used, by respondents’ educational attainment. 1.5 1 Number of Financial Products Number of Financial Products .8 1 .6 .4 .5 .2 0.62 1.01 1.15 0.91 0.70 0 0 Age<35, n=225 3555, n=234 Rural, n=570 Urban, n=430 Figure 48: Average number of �nancial products used, by Figure 49: Average number of �nancial products used, by respondents’ age. urbanization. 90. Compared to the rest of the population, individuals who are not in the la- bor force, as well as urban dwellers hold the smallest number of �nancial products (see �gs. 47, 48, and 49). After controlling for other characteristics, the number of �nancial products held by populations outside the labor force are statistically indistinguishable from people who are unemployed, informally, or formally employed. However, the difference between people outside the labor force versus retired and self- employed persons appears to be signi�cant. Concerning the urban living environment the differences persist after controlling for all the other characteristics included in the analysis. The key to this may lie, at least in part, in the aforementioned lower capability of putting money aside and planning for the long-term found among the urban popu- lation (Dushanbe’s inhabitants). Surprisingly, the average number of �nancial products used, does not vary with income and any differences are statistically insigni�cant. On average, the lowest income group holds 0.85 products and the highest income group holds 0.97 products. The largest difference in the average number of �nancial products held can be seen across groups with different levels of indebtedness. Those who do not have any outstanding debts hold only 0.7 products on average, which compares to an average of 2 products for adults with outstanding debts higher than their annual income. - 37 - 4.1.2. Financial Institutions 1.9 − 2 1.8 − 1.9 1.7 − 1.8 1.6 − 1.7 1.5 − 1.6 1.4 − 1.5 1.3 − 1.4 1.2 − 1.3 0 − 1.2 Figure 50: Average number of �nancial services used in regions, out of ‘Commercial banks’, ‘Insurance companies’, ‘Intermediary organizations’, ‘Micro�nance organizations’, ‘Other non-banking �nancial institutions’, ‘exchange office’, and ‘money transfer operators’. 91. Currency exchange offices (58 per- 100 Exch. Off. Money Transf. MFOs Banks Insurance Other NBFIs Broker cent) and MTOs (48.9 percent) are by far the main FIs whose products 80 and services Tajiks use. Services of- fered by any other FI only play a mi- nor role. Fig. 51 reveals that Tajikistan’s 60 banking, micro�nance and insurance sec- tors can be characterized by low usage 40 rates of their services among the adult population. Services offered by commer- 20 cial banks are only used by 8 percent, 58 49 11 8 12 1 0 those of insurance companies by 12.4 per- 0 Figure 51: Percentage of respondents using �nancial cent and MFOs only reach about 10.7 services. percent of the adult population. Hardly any Tajiks ever used services from other NBFIs or brokerage houses. 92. Few people living on low incomes and few residents from regions other than Dushanbe ever have used the services of currency exchange offices. High income earners are signi�cantly more likely to report having used any services of cur- rency exchange offices than those living on lower incomes. Similarly, people who live in Dushanbe are signi�cantly more likely to have used services from currency exchange offices than residents from DRS, Kathlon, and GBAO (see �g. 52). 93. Access to MTOs appears to be the most limited for low income groups and those who are out of the labor force. Signi�cantly fewer people living on lowest incomes use the services of MTOs, compared to high income populations (see �g. 58). - 38 - Compared to Dushanbe’s population, residents from Sogd are also signi�cantly more likely to ever have used any services provided by MTOs (see �g. 53). 94. Similarly, few individuals outside the labor force ever used the services offered by MFOs. Informally employed and self-employed Tajiks are signi�cantly more likely to have used any micro�nance services than those out of the labor market (see �g. 57). 95. Few singles and few people who are not the heads the of household use products provided by insurance companies. Given the fact that mainly general insurance products, such as motor or household content insurance, are available on the insurance market in Tajikistan, it is not surprising that household heads are in comparison to those who are not heads of household signi�cantly more likely to use these products (see �g. 55). Unmarried Tajiks, on the other hand, are less likely to ever have used any services offered by insurance companies compared to those who are married or living with a partner (see �g. 56). 96. People with high educational attainment and those who already saved as a child are more likely to report having used any bank services than those who do not have these de�ning characteristics. In comparison to those whose highest educational attainment is primary school or less, those with tertiary education are signif- icantly more likely to have used any products and services offered by commercial banks (see �g. 59). Likewise, those who began savings habits as a child are signi�cantly more likely to have used any bank services than those who did not save in their childhood (see �g. 60). In a similar vein, having more economically dependent children is associated with having used any services offered by commercial banks, potentially indicating differ- ent savings behavior in favor of the children. All three effects are statistically signi�cant even after other characteristics are taken into account. - 39 - 65 − 70 60 − 65 55 − 60 65 − 70 60 − 65 55 − 60 50 − 55 45 − 50 40 − 45 50 − 55 45 − 50 40 − 45 35 − 40 35 − 40 Figure 52: Percentage of respondents using services of currency exchange offices in Figure 53: Percentage of respondents using services of money transfer operators in regions. regions. 33 − 37 29 − 33 25 − 29 21 − 25 17 − 21 13 − 17 9 − 13 5−9 0−5 - 40 - Figure 54: Percentage of respondents using services of micro�nance organizations in regions. Exch. Off. Money Transf. MFOs Banks Exch. Off. Money Transf. MFOs Banks 60 60 Insurance Other NBFIs Broker Insurance Other NBFIs Broker 40 40 Percent Percent 20 20 56 46 10 6 11 1 0 61 55 11 13 16 2 0 49 38 10 5 5 2 0 60 52 11 9 14 1 0 0 0 Not Head, n=492 Is Head, n=508 Other, n=200 Married/Cohabit, n=800 Figure 55: Percentage of respondents using �nancial service, by respondent being head of Figure 56: Percentage of respondents using �nancial service, by respondent cohabiting stats. household. 80 Exchange Office Money Transfer Operator MFOs Banks Insurance Other NBFIs Broker - 41 - 60 Percent 40 20 6557141119 2 0 6347151114 2 0 544615 5 14 2 0 4746 6 4 5 1 0 7671111813 1 0 5241 5 6 7 1 0 0 Formal, n=267 Informal n=119 Self−Employed, n=156 Retired, n=146 Unemployed, n=66 Out of Laborforce, n=246 Figure 57: Percentage of respondents using �nancial service, by respondents’ employment status. 80 Exchange Office Money Transfer Operator MFOs Banks Insurance Other NBFIs Broker 60 Percent 40 20 48 42 8 6 11 1 0 56 49 11 9 13 1 0 73 56 13 8 13 2 0 0 < TJS 500, n=264 TJS 500−1,500, n=562 > TJS 1,500, n=168 Figure 58: Percentage of respondents using �nancial service, by approximate monthly household income. 80 80 Exch. Off. Money Transf. MFOs Banks Exch. Off. Money Transf. MFOs Banks - 42 - Insurance Other NBFIs Broker Insurance Other NBFIs Broker 60 60 Percent Percent 40 40 20 20 52 47 8 5 10 1 0 67 53 14 13 17 2 0 56 48 9 7 11 1 0 67 55 18 15 19 3 0 0 0 Primary and Secondary, n=596 Tertiary, n=404 Not Save, n=803 Did Save, n=196 Figure 59: Percentage of respondents using �nancial service, by respondents’ Figure 60: Percentage of respondents using �nancial service, by respondent’s educational attainment. childhood saving behavior. 4.2. Relationship between Financial Inclusion and Financial Capability 4.2.1. Relationship between Financial Inclusion and Financial Knowledge and Awareness 97. Access and usage of �nancial products and services can be constrained by lack of awareness about �nancial concepts. It cannot be assumed that someone who does not use any �nancial products and services faces speci�c barriers in partici- pating in the �nancial sector. Some parts of the population might voluntarily exclude themselves. However, even if non-participation in �nancial markets is based on voluntary decision-making it could be that this decision is the outcome of lack of required �nan- cial knowledge. As shown in �g. 61 a positive relationship between better knowledge of �nancial concepts and the number of �nancial products used exists. For example Tajiks who answered more of the quiz-like �nancial literacy questions correctly, also hold more �nancial products. Interestingly, this positive relationship turns stronger and becomes signi�cant if adults possessing with at least one �nancial product are considered. These results may indicate that enhanced �nancial knowledge of those who already use a �- nancial product is key for further �nancial sector widening and deepening. Before that, removing impediments to �nancial inclusion other than lack of �nancial knowledge (e.g. high cost, physical distance, lack of trust in FI, etc.) may be more important. Clearly, these statements are indicative at best and cannot be interpreted as causal. It is equally plausible that those with higher exposure to �nancial products are more likely to become �nancially knowledgeable. 98. In addition, being more �nancially literate relates positively with increased usage of products provided by different FIs. In comparison to those with lower scores, those who scored better with regard to the �nancial knowledge questions are more likely to use products offered by different types of �nancial service providers, such as currency exchange offices, MTOs, MFOs, and commercial banks (see �g. 62). Given that many insurance policies are compulsory, it is not surprising that no major differences emerge in the usage of insurance products across the groups with different levels of �nancial knowledge (see �g. 62). 4.2.2. Relationship between Financial Inclusion and Financial Attitudes and Behavior 99. Similarly, access and usage of �nancial products and services can be further constrained by �nancial attitudes and behaviors. However, in Tajikistan only speci�c aspects of �nancial attitudes and behavior seem to be associated with usage of a larger number of �nancial products and services. For example, people with higher scores for budgeting, saving, planning for unexpected and old age expenses, for using information and for achievement orientation tend to use more �nancial products. This may be due to the fact that people who plan for unexpected events and for old age expenses are more likely to use saving and investment products. In addition, people who tend to use advice and information may feel more con�dent about using �nancial products. 100. On the other hand, people who are better in budgeting and living within their means tend to use fewer �nancial products. A potential explanation could be that people who plan their expenses well and do not need to borrow to buy food and other essentials are less in need of using credit products. - 43 - 6 Jittered Scatter, J=15 Number of Financial Products Unjittered Scatter Fit (All Participants): b=.05 (not sig.) Fit (At least 1 Prod.): b=.07** 0 2 4 0 1 2 3 4 5 6 7 Financial Literacy Score Linear fit without further controls. Significance levels: p < .01***; p < .05**; p < .1*. J(15): Jitter=15 Figure 61: Relationship between �nancial inclusion vs. �nancial literacy. - 44 - 40 Exchange Office Money Transfer Operator MFOs Banks Insurance Other NBFIs Broker 30 Percent 20 10 20 15 12 6 10 0 1 26 19 16 12 8 1 1 33 28 9 18 10 3 1 0 0−2, n=124 3−5, n=741 6−8, n=135 Figure 62: Percentages of respondents currently using the �nancial product, given their �nancial literacy score. 5. Consumer Protection 101. Consumer protection is critical in ensuring con�dence in the �nancial sys- tem and increasing �nancial inclusion. A high incidence of conflicts with �nancial service providers or low levels of satisfaction with �nancial products offered by FIs could undermine the trust in the �nancial system and impede �nancial inclusion. Also, weak consumer protection can result in consumers being harmed by �nancial products rather than bene�tting from them. 102. Consumers do not widely report complaints or other conflicts with �nancial service providers. Even less than 1 percent of the adult population reported that they experienced conflicts with FIs in the past 3 years. This low level of self-reported �nancial service provider conflicts certainly relates to the low levels of �nancial inclusion, but most likely also to difficulties many consumers have in understanding which �nancial products would �t their needs best. Due to this low number of incidences, no reliable conclusions can be drawn from the actual behavior of those who encountered a conflict. 103. A signi�cant fraction of those who did not encounter any conflicts would not do anything if they faced a conflict with a �nancial service provider. Around half of the people who did not encounter a conflict reported that in case of a conflict they would choose to solve it formally through a system for redress. A third indicated that they would not do anything at all. This may be due to the fact that they do not know which institutions could help them to solve �nancial service provider conflicts or because they do not trust that formal redress mechanisms would do anything for them. Interestingly, of those adults who would seek either formal redress or would try to solve the conflict informally by directly discussing with bank representatives, almost everyone expresses con�dence that the issue would be solved. 104. The levels of satisfaction with �nancial services offered by FIs other than insurance companies are very high (see �g. 63). Interestingly, those parts of the population who have used the services offered by commercial banks, intermediary organizations, MFOs, NBFIs, money transfer and exchange offices are highly satis�ed with their services. Only insurance companies seem to offer lower quality services, which satisfy their customers much less. Given that by international standards the fraction of premiums paid out in claims is quite small, these �ndings are not surprising. 105. Some customers are even less satis�ed with the services that insurance com- panies provide. Whereas 95 percent of younger aged populations indicated that they are satis�ed with the services offered by insurance companies, the level of satisfaction clearly decreases with higher age. Only two thirds of those aged 55 or more reported to be satis�ed with services provided by insurance companies (see �g. 65). A plausible explanation could be that seniors more likely faced a situation where they submitted a claim to an insurance company and either received low or no compensation at all. Fig. 64 reveals that women are another group of the population who are the least satis�ed with the services provided by insurance companies. Whereas 88 percent of all men who ever held insurance policies reported to be satis�ed, the corresponding number for women is only 68 percent. Compared to formally, informally and self employed Tajiks, those who are out of work also indicated to be much less satis�ed with insurance companies and their services (see �g. 66). - 45 - Exch. Off. Money Transf. MFOs Banks Insurance Other NBFIs Broker Exchange Office (N=492) 100 Money Transfer (N=590) 80 Microfinance Org. (N=129) Percent 60 Commercial Bank (N=118) Insurance (N=117) 40 Other Non−Banking (N=32) 20 Broker (N=3) 98 99 91 98 79 100 98 98 99 98 92 100 100 0 0 20 40 60 80 100 Female Male Figure 63: Percentage of respondents being satis�ed with services offered by �nancial Figure 64: Percentage of individuals being satis�ed with services offered by �nancial institutions given respondents use it; number of cases in parentheses. institution, by respondents’ gender. Exchange Office Money Transfer Operators MFOs Banks Insurance Other NBFIs Broker 100 - 46 - 80 Percent 60 40 20 97 98 89 100 98 100 99 98 100 97 76 100 100 99 100 96 100 85 100 100 0 Age<35 3555 Figure 65: Percentage of individuals being satis�ed with services offered by �nancial institution, by respondents’ age. Exchange Office Money Transfer Operators MFOs Banks Insurance Other NBFIs Broker 100 80 Percent 60 40 20 97 96 100100 86 100100 98 98 100100 99 100100 100100 77 91 98 100100 98 100100 98 72 100 0 Formal Informal Self Empl. Not Work Figure 66: Percentage of individuals being satis�ed with services offered by �nancial institution, by respondents’ employment status. - 47 - A. Background Tajik Survey Table 2: Information on sampling design. Strata Urban / Number of Cluster Number of N of expected N of selected Rural households size Clusters interviews households Dushanbe Urban 138,588 5 26 130 208 DRS Urban 40,731 5 14 70 112 DRS Rural 196,095 10 15 150 210 Kathlon Urban 75,341 5 19 95 152 Kathlon Rural 280,966 10 18 180 252 Sogd Urban 110,001 5 22 110 176 Sogd Rural 259,937 10 18 180 252 GBAO Urban 5,111 5 5 25 40 GBAO Rural 25,851 10 6 60 84 Total 1,132 143 1,000 1,486 Dushanbe (12.7%) Dushanbe (11.9%) Sogd (33.1%) Sogd (30.2%) Khatlon (32.3%) Khatlon (26.5%) RRP (18.7%) RRP (28.8%) GBAO (3.3%) GBAO (2.6%) Figure 67: Comparison of regional coverage of all households in sample. Rural (67.5%) Rural (70%) Urban (32.5%) Urban (30%) Figure 68: Comparison of household environment of all households in sample. < TJS 500 (23.4%) < TJS 500 (21.3%) TJS 500−1,500 (45.8%) TJS 500−1,500 (57.1%) > TJS 1,500 (30.8%) > TJS 1,500 (21.6%) Figure 69: Comparison of total inflation corrected household income of all households in sample. Inflation data retrieved from World Bank open data panel. - 48 - Female (53%) Female (53%) Male (47%) Male (47%) Figure 70: Comparison of intra-household gender distribution of all households in sample and of individuals 18 years or older. Younger 35 (27.5%) Younger 35 (36.5%) 35 To 55 (30.7%) 35 To 55 (32.7%) Older 55 (41.7%) Older 55 (30.8%) Figure 71: Comparison of intra-household age distribution of all households in sample and of individuals 18 years or older. - 49 - B. Financial Capability B.1. Knowledge of Financial Concepts and Products Background Information: Methods and Dependent Variables in Table 3 Literacy Score based on questions from sec. 3.1.1. The score sums all correct answers per respondent. Each question receives the same weight, i.e. knowing more difficult questions is not “rewarded�. The analysis applies a basic linear probability model, extended for the correction of the survey design. Financial Products Knowledge Score The respondent is confronted with a list of �nancial institutions: Commercial banks, Insurance companies, Broker (Intermediary organizations), Micro�nance orga- nizations, Other non-banking �nancial institutions, Exchange office, and money transfer. For each institution the respondent answers ‘Yes’ or ‘No’ to the following question: ‘Do you know about the services offered by the following institutions? ’. The variable counts how any institutions are know by the respondent. For the analysis a poisson model is �tted, extended for the correction of the survey design. Table 3: Financial Knowledge Financial Literacy Financial Products Score Knowledge Score Age −0.06 0.02 [−0.21, 0.09] [−0.09, 0.13] Age Square 0.00 −0.00 [−0.00, 0.00] [−0.00, 0.00] Age Cube −0.00 −0.00 [−0.00, 0.00] [−0.00, 0.00] Male −0.00 0.03 [−0.31, 0.31] [−0.14, 0.20] Secondary Educ. 0.41 −0.00 [−0.25, 1.07] [−0.32, 0.32] Tertiary Educ. 0.64∗ 0.05 [−0.08, 1.36] [−0.32, 0.41] Literate (Tajik) 0.15 0.27∗∗ [−0.26, 0.56] [0.04, 0.49] Household Size 0.00 0.02∗ [−0.06, 0.07] [−0.00, 0.05] Econ. Dep. Kids −0.05 −0.07∗∗∗ [−0.13, 0.03] [−0.12, −0.02] Married/Cohabit 0.58∗∗∗ 0.17 [0.22, 0.95] [−0.04, 0.37] Household Head 0.41∗∗∗ 0.11 [0.15, 0.67] [−0.05, 0.28] Formal Sector −0.25 0.14 [−0.60, 0.10] [−0.04, 0.31] Informal Sector −0.06 0.11 [−0.59, 0.46] [−0.13, 0.36] Self Employed 0.15 0.14 - 50 - [−0.31, 0.61] [−0.13, 0.40] Retired −0.06 −0.06 [−0.49, 0.37] [−0.27, 0.16] Unemployed 0.12 0.27∗∗ [−0.33, 0.58] [0.03, 0.51] Urban Environment 0.35∗ −0.00 [−0.01, 0.72] [−0.15, 0.15] 500 to 1,500 −0.06 −0.16∗∗ [−0.33, 0.21] [−0.31, −0.00] Higher 1,500 0.13 −0.07 [−0.22, 0.47] [−0.23, 0.09] DRS 0.11 0.46∗∗ [−0.59, 0.81] [0.10, 0.82] Khatlon −0.19 0.44∗∗∗ [−0.84, 0.46] [0.12, 0.76] Sogd 0.44 0.54∗∗∗ [−0.18, 1.06] [0.24, 0.83] Gbao 0.38 0.48∗∗ [−0.32, 1.08] [0.11, 0.84] Media Consumption 0.21∗∗∗ 0.11∗∗∗ [0.09, 0.33] [0.04, 0.18] Save as Child 0.18 0.16∗∗ [−0.10, 0.46] [0.03, 0.30] Observations 993 993 Signi�cance: ∗ (p < 0.10), ∗∗ (p < 0.05), ∗∗∗ (p < 0.01); 95% Con�dence Intervals in brackets. Reference Categories: female, cannot read or write Tajik, neither married nor living with partner, maximum primary education, not head of household, out of labor force, rural environment, income lower TJS 500, Dushanbe Region, not saved as a child - 51 - B.2. Knowledge of �nancial Products Background Information: Methods and Dependent Variables in Table 4 Table 4 considers the same list of institutions as analyzed for table 3. The analyses in table 4 explore the results from column ‘Knowledge’ in table 3 in greater detail. Each institution is considered separately. For each institution a zero-one dummy is created, equal to one if the respondent answers ‘Yes ’ and zero if the reply is ‘No ’. The results are obtained by �tting probit models, extended for the correction of the survey setup. The following abbreviations are applied: • Banks Commercial banks • Insur Insurance Companies • Broker Intermediary organizations • MFO Micro�nance Organizations • ONFI Other Non-Banking Financial Institutions • Exch. Exchange Office • MTO Money Transfer Operator - 52 - Table 4: Knowledge of Speci�c Financial Products Banks Insur Broker MFO ONFI Exch. MTO Age 0.05 0.17 0.12 −0.00 0.12 0.06 −0.18 [−0.19, 0.29] [−0.04, 0.38] [−0.20, 0.44] [−0.22, 0.21] [−0.13, 0.37] [−0.24, 0.36] [−0.45, 0.09] Age Square −0.00 −0.00 −0.00 0.00 −0.00 −0.00 0.00 [−0.01, 0.00] [−0.01, 0.00] [−0.01, 0.00] [−0.00, 0.00] [−0.01, 0.00] [−0.01, 0.01] [−0.00, 0.01] Age Cube 0.00 0.00 0.00 −0.00 0.00 −0.00 −0.00∗ [−0.00, 0.00] [−0.00, 0.00] [−0.00, 0.00] [−0.00, 0.00] [−0.00, 0.00] [−0.00, 0.00] [−0.00, 0.00] Male −0.03 0.01 0.19 −0.12 0.19 0.18 0.04 [−0.46, 0.39] [−0.38, 0.39] [−0.29, 0.66] [−0.46, 0.22] [−0.24, 0.62] [−0.34, 0.71] [−0.44, 0.53] Secondary Educ. 0.21 0.11 −0.29 −0.33 −0.06 −0.03 0.30 [−0.28, 0.71] [−0.51, 0.73] [−1.09, 0.51] [−0.85, 0.19] [−0.71, 0.59] [−0.66, 0.60] [−0.23, 0.83] Tertiary Educ. 0.42 0.24 −0.17 −0.26 0.07 0.14 0.11 [−0.21, 1.05] [−0.47, 0.95] [−0.98, 0.65] [−0.85, 0.33] [−0.64, 0.78] [−0.57, 0.84] [−0.60, 0.82] Literate (Tajik) 0.61∗∗ 0.59∗∗ 0.71∗∗∗ 0.36 0.24 0.53 0.01 [0.10, 1.12] [0.05, 1.12] [0.17, 1.25] [−0.13, 0.85] [−0.40, 0.89] [−0.33, 1.38] [−0.52, 0.55] Household Size 0.01 0.06 0.08∗∗ 0.04 0.06 0.04 −0.00 [−0.04, 0.07] [−0.02, 0.14] [0.01, 0.14] [−0.03, 0.10] [−0.02, 0.13] [−0.04, 0.13] [−0.08, 0.07] Econ. Dep. Kids −0.08 −0.23∗∗∗ −0.14∗ −0.12∗∗ −0.10∗ −0.07 −0.09 [−0.18, 0.02] [−0.35, −0.11] [−0.30, 0.01] [−0.23, −0.01] [−0.23, 0.02] [−0.19, 0.05] [−0.19, 0.02] Married/Cohabit 0.36 0.15 0.35 0.41∗ 0.09 0.31 0.30 [−0.08, 0.80] [−0.29, 0.58] [−0.14, 0.83] [−0.05, 0.88] [−0.40, 0.57] [−0.24, 0.85] [−0.15, 0.74] Household Head 0.24 0.22 0.36 0.22 0.06 0.23 0.22 - 53 - [−0.16, 0.64] [−0.15, 0.59] [−0.09, 0.80] [−0.15, 0.58] [−0.33, 0.45] [−0.23, 0.69] [−0.21, 0.65] Formal Sector 0.07 −0.05 −0.06 0.61∗∗∗ 0.24 0.21 0.55∗∗ [−0.35, 0.48] [−0.42, 0.31] [−0.44, 0.33] [0.20, 1.02] [−0.21, 0.70] [−0.40, 0.82] [0.08, 1.03] Informal Sector 0.31 −0.22 −0.21 0.63∗∗ 0.01 0.35 0.18 [−0.24, 0.85] [−0.74, 0.30] [−0.78, 0.35] [0.11, 1.15] [−0.55, 0.57] [−0.36, 1.06] [−0.41, 0.78] Self Employed −0.01 0.11 0.18 0.52∗∗ 0.31 0.07 0.23 [−0.56, 0.53] [−0.43, 0.66] [−0.45, 0.81] [0.00, 1.03] [−0.28, 0.89] [−0.57, 0.70] [−0.28, 0.75] Retired −0.27 −0.32 −0.26 −0.02 −0.22 0.26 −0.09 [−0.84, 0.29] [−0.82, 0.18] [−0.76, 0.24] [−0.52, 0.47] [−0.65, 0.20] [−0.29, 0.81] [−0.66, 0.47] Unemployed 0.53∗∗ 0.07 0.16 0.68∗∗ 0.47 0.56 1.00∗∗∗ [0.04, 1.01] [−0.59, 0.73] [−0.65, 0.97] [0.06, 1.31] [−0.14, 1.07] [−0.69, 1.81] [0.47, 1.53] Urban Environment −0.19 0.03 −0.04 0.07 0.02 −0.01 0.03 [−0.58, 0.20] [−0.36, 0.42] [−0.43, 0.35] [−0.30, 0.44] [−0.35, 0.39] [−0.44, 0.41] [−0.36, 0.42] 500 to 1,500 −0.16 −0.24 −0.31 −0.41∗∗ −0.38∗∗ −0.24 −0.13 [−0.54, 0.23] [−0.57, 0.09] [−0.71, 0.10] [−0.72, −0.10] [−0.74, −0.02] [−0.67, 0.18] [−0.57, 0.32] Higher 1,500 −0.05 −0.14 −0.30 −0.33 −0.27 0.16 0.18 [−0.53, 0.43] [−0.55, 0.28] [−0.75, 0.15] [−0.75, 0.08] [−0.68, 0.14] [−0.50, 0.82] [−0.37, 0.74] DRS 0.93∗∗∗ 0.38 0.66∗ 0.93∗∗∗ 0.64∗ −0.31 0.89∗∗ [0.24, 1.61] [−0.32, 1.09] [−0.02, 1.35] [0.28, 1.58] [−0.02, 1.30] [−1.03, 0.41] [0.20, 1.59] Khatlon 0.66∗∗ 0.50 0.58∗ 0.83∗∗ 0.17 0.04 1.24∗∗∗ [0.03, 1.29] [−0.16, 1.16] [−0.10, 1.25] [0.19, 1.48] [−0.50, 0.84] [−0.71, 0.79] [0.59, 1.88] Sogd 0.51∗ 0.60∗∗ 0.93∗∗∗ 1.03∗∗∗ 0.60∗∗ 0.59∗ 1.45∗∗∗ [−0.04, 1.07] [0.02, 1.18] [0.37, 1.49] [0.45, 1.62] [0.03, 1.17] [−0.05, 1.23] [0.85, 2.04] Gbao 1.03∗∗∗ 0.71∗ 0.71 1.17∗∗∗ 0.74∗ −0.73∗ 0.50 [0.31, 1.75] [−0.11, 1.54] [−0.15, 1.58] [0.43, 1.91] [−0.05, 1.52] [−1.46, 0.00] [−0.16, 1.15] Media Consumption 0.21∗∗ 0.24∗∗∗ 0.20∗∗ 0.17∗∗ 0.18∗∗ 0.13∗ 0.13∗ [0.05, 0.38] [0.10, 0.38] [0.04, 0.36] [0.03, 0.31] [0.02, 0.35] [−0.02, 0.29] [−0.02, 0.28] Save as Child 0.30∗ 0.31∗ 0.13 0.77∗∗∗ 0.21 0.38 0.30 [−0.03, 0.64] [−0.06, 0.68] [−0.26, 0.52] [0.38, 1.15] [−0.17, 0.58] [−0.19, 0.96] [−0.09, 0.69] Observations 993 993 993 993 993 993 993 Signi�cance: ∗ (p < 0.10), ∗∗ (p < 0.05), ∗∗∗ (p < 0.01); 95% Con�dence Intervals in brackets. Reference Categories: female, cannot read or write Tajik, neither married nor living with partner, maximum primary education, not head of household, out of labor force, rural environment, income lower TJS 500, Dushanbe Region, not saved as a child - 54 - B.3. Financial Attitudes and Behavior Background Information: Methods and Dependent Variables in Tabs. 5 and 6 Based on factor analysis nine �nancial capability components are derived. The components are scores ranging from 0 (lowest capability) to 100 (highest capability). Below the �rst �ve measures relate to behaviors, the other four to attitudes and motivations. All analyses in tabs. 5 and 6 apply the same linear probability model, correcting for the survey design. The following �nancial capability scores and abbreviations are applied: Budgeting measures to which extent people plan how to use their money and whether they stick to the plan. Living w/ means The outcome living within means measures the level of borrowing, whether people borrow to buy food and other essentials, and whether people who have money left over after buying essentials save it (instead of only spending it on non-essentials). Not Overspend indicates whether people tend to spend their income on things that are not necessary or on things they cannot afford. Plan old/unexp. The score for planning for unexpected and old age expenses: indicates whether people could cover an unexpected expense equivalent to a month’s income and whether they worry about it, and if they have provisions that cover/will cover their old age expenses in full. Saving measures whether people see themselves as trying to save for the future, trying to save for emergen- cies, and trying to save even if a small amount. Info/Learn The measure for attitudes towards information is a combination of getting information and advice before making �nancial decisions, learning from others, and being disciplined. Foresightedness conveys attitudes towards the future. It measure whether people agree or disagree with statements such as ‘I live more for the present day than for tomorrow’, ‘The future will take care of itself’, ‘I only focus on the short term’. Not Impulsive stands for non-impulsiveness and is determined by whether people agree or disagree with statements about impulsivity (being impulsive, saying things without giving them too much thought, doing things without thinking them through). Achievement designates achievement orientation and assesses to what extent people agree with statements on having aspirations, working hard to be the best, always looking for opportunities to improve one’s own situation. - 55 - Table 5: Financial Capability Scores – Part 1 Budgeting Living w/ means Not Overspend Info/Learn Plan old/unexp. Age −0.01 −1.44 0.78 2.76∗∗ 0.69 [−2.39, 2.37] [−3.58, 0.71] [−1.93, 3.50] [0.64, 4.88] [−2.25, 3.63] Age Square 0.00 0.02 −0.02 −0.06∗∗ −0.01 [−0.04, 0.05] [−0.02, 0.07] [−0.08, 0.03] [−0.10, −0.01] [−0.07, 0.06] Age Cube −0.00 −0.00 0.00 0.00∗∗ 0.00 [−0.00, 0.00] [−0.00, 0.00] [−0.00, 0.00] [0.00, 0.00] [−0.00, 0.00] Male 1.56 3.95∗ −1.24 −0.47 0.64 [−2.55, 5.68] [−0.21, 8.12] [−4.61, 2.13] [−3.90, 2.96] [−3.85, 5.14] Secondary Educ. 5.21 −0.95 2.00 −3.61 3.10 [−2.52, 12.94] [−5.31, 3.42] [−3.95, 7.95] [−8.08, 0.85] [−5.03, 11.23] Tertiary Educ. 8.59∗∗ 0.11 −1.34 −4.09 4.87 [0.66, 16.52] [−5.05, 5.27] [−7.64, 4.96] [−9.13, 0.96] [−5.39, 15.14] Literate (Tajik) 1.91 −2.75 4.78 3.66 −5.71 [−6.20, 10.02] [−7.88, 2.39] [−1.97, 11.53] [−1.45, 8.77] [−13.29, 1.87] - 56 - Household Size −0.20 −0.80∗∗ 0.50∗∗ 0.35 0.08 [−0.93, 0.53] [−1.42, −0.17] [0.08, 0.92] [−0.16, 0.86] [−0.73, 0.89] Econ. Dep. Kids 0.32 −0.29 0.73∗∗ 0.39 −0.67 [−0.83, 1.47] [−1.34, 0.76] [0.01, 1.46] [−0.54, 1.32] [−2.14, 0.80] Married/Cohabit −5.19∗∗ −0.77 1.74 −1.94 −1.21 [−10.20, −0.19] [−4.73, 3.19] [−2.51, 6.00] [−5.63, 1.74] [−6.45, 4.03] Household Head −2.16 −4.98∗∗ 0.55 0.59 3.42 [−5.63, 1.30] [−9.74, −0.22] [−2.20, 3.30] [−3.01, 4.19] [−1.92, 8.76] Formal Sector 0.03 −0.42 1.61 0.59 1.91 [−5.12, 5.18] [−5.50, 4.66] [−2.64, 5.85] [−4.17, 5.36] [−4.52, 8.34] Informal Sector 2.84 1.46 2.92 2.92 −0.04 [−3.82, 9.51] [−4.59, 7.51] [−1.07, 6.91] [−2.04, 7.88] [−8.17, 8.08] Self Employed −5.18 1.27 −0.68 3.53 0.64 [−11.61, 1.25] [−4.90, 7.45] [−5.33, 3.97] [−1.56, 8.61] [−7.45, 8.73] Retired 0.09 −2.69 −0.13 2.01 8.13∗∗ [−5.10, 5.28] [−9.96, 4.57] [−4.86, 4.59] [−2.94, 6.96] [1.31, 14.95] Unemployed −5.26 −2.26 −1.10 1.80 −4.60 [−11.86, 1.33] [−8.99, 4.47] [−6.34, 4.14] [−4.20, 7.79] [−11.20, 2.00] Urban Environment 4.08∗ −3.05 1.64 0.02 3.66 [−0.24, 8.41] [−7.40, 1.30] [−1.06, 4.33] [−3.87, 3.90] [−1.16, 8.48] 500 to 1,500 −1.07 −0.76 −1.94 −1.28 −1.60 [−5.27, 3.13] [−4.98, 3.46] [−4.93, 1.06] [−4.74, 2.17] [−6.29, 3.09] Higher 1,500 5.44∗ 9.49∗∗∗ −2.32 −4.26∗ 9.27∗∗∗ [−0.29, 11.17] [4.48, 14.50] [−6.75, 2.10] [−8.90, 0.38] [2.84, 15.70] DRS 13.68∗∗∗ 0.62 7.72∗∗ 19.34∗∗∗ 4.13 [6.09, 21.27] [−6.49, 7.73] [1.63, 13.81] [11.90, 26.78] [−4.69, 12.96] Khatlon 21.82∗∗∗ −1.57 8.92∗∗∗ 18.78∗∗∗ 5.39 [15.19, 28.45] [−10.53, 7.39] [3.14, 14.71] [11.17, 26.38] [−2.01, 12.79] Sogd 17.26∗∗∗ 2.71 5.91∗∗ 18.16∗∗∗ 7.68∗∗ [11.01, 23.50] [−4.98, 10.39] [0.63, 11.19] [11.23, 25.09] [0.95, 14.41] Gbao 26.37∗∗∗ −5.49 11.19∗∗∗ 19.52∗∗∗ 15.62∗∗∗ - 57 - [15.53, 37.20] [−15.29, 4.31] [5.46, 16.92] [10.38, 28.67] [6.30, 24.93] Media Consumption 0.11 0.26 −0.90 1.37∗ 1.03 [−1.37, 1.58] [−1.30, 1.81] [−2.00, 0.19] [−0.02, 2.77] [−1.18, 3.25] Save as Child 2.90 −0.95 3.65∗∗ 4.17∗∗∗ 3.86 [−1.29, 7.08] [−6.24, 4.34] [0.73, 6.58] [1.12, 7.22] [−1.16, 8.88] Observations 831 979 992 993 954 Signi�cance: ∗ (p < 0.10), ∗∗ (p < 0.05), ∗∗∗ (p < 0.01); Reference Categories: female, cannot read or write Tajik, neither married nor living with partner, maximum primary education, not head of household, out of labor force, rural environment, income lower TJS 500, Dushanbe Region, not saved as a child Table 6: Financial Capability Scores – Part 2 Saving Foresightedness Not Impulsive Achievement Age −0.17 0.88 3.12∗∗ −1.29 [−4.94, 4.60] [−2.46, 4.22] [0.22, 6.02] [−4.03, 1.46] Age Square 0.01 −0.02 −0.06∗∗ 0.04 [−0.09, 0.11] [−0.09, 0.05] [−0.12, −0.00] [−0.02, 0.09] Age Cube −0.00 0.00 0.00∗∗ −0.00 [−0.00, 0.00] [−0.00, 0.00] [0.00, 0.00] [−0.00, 0.00] Male −2.19 −3.94 1.51 −1.42 [−11.14, 6.77] [−9.99, 2.10] [−2.04, 5.07] [−6.21, 3.36] Secondary Educ. −9.42 −5.33 −1.60 −6.46∗∗ [−21.85, 3.01] [−13.48, 2.81] [−5.80, 2.60] [−12.82, −0.10] Tertiary Educ. −10.12 −6.28 −3.78 −1.22 [−22.32, 2.08] [−15.08, 2.52] [−8.84, 1.29] [−8.12, 5.68] Literate (Tajik) −6.97 2.95 2.22 −1.29 [−18.96, 5.01] [−5.45, 11.36] [−2.73, 7.17] [−6.85, 4.26] Household Size −0.20 0.22 0.03 −0.77 [−1.71, 1.30] [−0.69, 1.12] [−0.58, 0.64] [−1.78, 0.23] Econ. Dep. Kids 3.00∗∗ 0.80 0.52 2.19∗∗∗ [0.27, 5.73] [−0.78, 2.39] [−0.46, 1.49] [0.55, 3.83] Married/Cohabit 1.68 6.11∗ −1.44 0.64 - 58 - [−8.57, 11.94] [−0.25, 12.47] [−5.69, 2.81] [−4.71, 5.99] Household Head 3.52 0.92 −0.29 3.81 [−3.99, 11.03] [−4.58, 6.43] [−3.36, 2.78] [−0.96, 8.59] Formal Sector 0.66 3.81 2.29 1.92 [−9.46, 10.78] [−3.36, 10.97] [−2.40, 6.99] [−2.99, 6.82] Informal Sector −2.48 2.09 −0.72 4.32 [−15.28, 10.31] [−6.15, 10.33] [−5.70, 4.27] [−3.14, 11.79] Self Employed −2.05 −3.90 0.39 1.12 [−13.72, 9.61] [−13.58, 5.78] [−3.86, 4.64] [−7.50, 9.73] Retired 0.01 6.41 1.29 −1.54 [−11.12, 11.14] [−2.04, 14.86] [−3.70, 6.28] [−8.84, 5.76] Unemployed 0.94 3.88 1.53 1.68 [−12.55, 14.42] [−3.94, 11.70] [−3.17, 6.24] [−4.74, 8.09] Urban Environment −2.65 −1.65 1.84 −1.85 [−13.02, 7.72] [−8.11, 4.81] [−1.63, 5.31] [−7.09, 3.39] 500 to 1,500 5.96 3.57 1.50 0.97 [−2.67, 14.59] [−2.40, 9.54] [−3.04, 6.03] [−4.98, 6.92] Higher 1,500 10.38∗ 9.11∗∗ 1.62 4.27 [−0.97, 21.73] [1.53, 16.70] [−4.10, 7.34] [−1.17, 9.71] DRS 31.50∗∗∗ 4.09 3.29 11.34∗∗ [16.39, 46.60] [−7.54, 15.72] [−3.43, 10.01] [1.84, 20.84] Khatlon 21.80∗∗∗ 10.90∗ 4.02 8.53∗ [5.55, 38.05] [−0.05, 21.85] [−2.12, 10.15] [−1.19, 18.25] Sogd 29.54∗∗∗ 8.72 2.69 8.03∗ [15.58, 43.50] [−2.13, 19.58] [−4.10, 9.49] [−1.10, 17.17] Gbao 54.96∗∗∗ −0.10 3.82 15.47∗∗∗ [39.69, 70.24] [−11.83, 11.62] [−2.51, 10.14] [5.95, 24.98] Media Consumption 5.18∗∗∗ 1.69 −0.23 −0.02 [2.15, 8.20] [−0.77, 4.14] [−1.57, 1.10] [−2.61, 2.58] Save as Child 9.72∗∗ −0.65 −1.86 8.27∗∗∗ [0.88, 18.56] [−6.13, 4.83] [−6.57, 2.85] [4.01, 12.53] Observations 982 993 993 993 - 59 - Signi�cance: ∗ (p < 0.10), ∗∗ (p < 0.05), ∗∗∗ (p < 0.01); 95% Con�dence Intervals in brackets. Reference Categories: female, cannot read or write Tajik, neither married nor living with partner, maximum primary education, not head of household, out of labor force, rural environment, income lower TJS 500, Dushanbe Region, not saved as a child C. Financial Inclusion: Access and Usage of Financial Services C.1. Products Background Information: Methods and Dependent Variables in Table 3 Products Similar to Knowledge, but asking; ‘Regarding some �nancial products and services that some people use, do you currently have any of the following? ’. The products and services are: Insurance (health, life or income replacement insurance), Mortgage, Formal credit (loan from a commercial bank, credit card), General insurance (car insurance, household contents insurance, building insurance), Formal savings / deposit account in a commercial bank, Checking account / money transfer services (internet banking, Mobile banking, Western Union, debit card), Credit from micro �nance institution (e.g Imon, Humo etc), Saving at micro deposit institution (e.g Bovari va Hamkori etc), Informal credit (loan from informal money lenders, loan from family or friends that need to be repaid), Informal savings (at home: keeping money under the mattress), and Other. The same model is �tted. Table 7: Number of Financial Products Products Age 0.11 [−0.11, 0.34] Age Square −0.00 [−0.01, 0.00] Age Cube 0.00 [−0.00, 0.00] Male −0.21 [−0.56, 0.13] Secondary Educ. 0.02 [−0.50, 0.53] Tertiary Educ. 0.05 [−0.47, 0.57] Literate (Tajik) −0.04 [−0.53, 0.44] Household Size 0.02 [−0.04, 0.08] Econ. Dep. Kids 0.02 [−0.05, 0.09] Married/Cohabit 0.11 [−0.15, 0.38] Household Head 0.08 [−0.15, 0.32] Formal Sector 0.26 [−0.12, 0.65] Informal Sector −0.05 [−0.56, 0.46] Self Employed 0.45∗∗ - 60 - [0.01, 0.89] Retired 0.58∗∗∗ [0.19, 0.97] Unemployed 0.20 [−0.24, 0.63] Urban Environment −0.21∗ [−0.45, 0.02] 500 to 1,500 0.04 [−0.20, 0.29] Higher 1,500 −0.12 [−0.50, 0.26] DRS −0.07 [−0.58, 0.43] Khatlon 0.14 [−0.31, 0.59] Sogd 0.20 [−0.22, 0.62] Gbao −0.12 [−0.73, 0.48] Media Consumption 0.07 [−0.03, 0.18] Save as Child 0.14 [−0.15, 0.43] Observations 993 Signi�cance: ∗ (p < 0.10), ∗∗ (p < 0.05), ∗∗∗ (p < 0.01); 95% Con�dence Intervals in brackets. Reference Categories: female, cannot read or write Tajik, neither married nor living with partner, maximum primary education, not head of household, out of labor force, rural environment, income lower TJS 500, Dushanbe Region, not saved as a child - 61 - C.2. Financial Institutions Background Information: Methods and Dependent Variables in Table 8 Table 8 extends the analysis of table 7. Following the same list and given that an individual knows the services offered by the institution (see tab. 4) he or she is asked: ‘If Yes, have you ever used it? ’. Again zero-one dummies are generated. Agents who did not know the services offered by the institution where assessed as never having used these services. Thereby the results include the entire sample and the analyses become more meaningful. Like above probit models are estimated, extended for the correction of the survey setup. Again, the following abbreviations are applied: • Banks Commercial banks • Insur Insurance Companies • MFO Micro�nance Organizations • Exch. Exchange Office • MTO Money Transfer Operator Analyses for Broker and ONFIs are excluded since too few people use these products, which makes a more thorough statistical investigation not applicable. - 62 - Table 8: Usage of Speci�c Financial Products Banks Insur MFO Exch. MTO Age −0.15 0.25 0.12 0.08 −0.16 [−0.71, 0.41] [−0.20, 0.70] [−0.19, 0.44] [−0.13, 0.30] [−0.41, 0.09] Age Square 0.00 −0.01 −0.00 −0.00 0.00 [−0.01, 0.02] [−0.01, 0.00] [−0.01, 0.01] [−0.01, 0.00] [−0.00, 0.01] Age Cube −0.00 0.00 0.00 0.00 −0.00 [−0.00, 0.00] [−0.00, 0.00] [−0.00, 0.00] [−0.00, 0.00] [−0.00, 0.00] Male −0.02 −0.01 −0.42 0.15 −0.06 [−0.50, 0.45] [−0.59, 0.57] [−0.92, 0.09] [−0.17, 0.48] [−0.45, 0.32] Secondary Educ. 0.25 0.04 −0.46 0.08 0.29 [−0.22, 0.71] [−0.77, 0.85] [−1.36, 0.43] [−0.58, 0.73] [−0.26, 0.84] Tertiary Educ. 0.80∗∗∗ 0.06 −0.18 0.35 0.33 [0.32, 1.29] [−0.75, 0.87] [−1.20, 0.84] [−0.37, 1.06] [−0.29, 0.95] Literate (Tajik) −0.26 −0.29 −0.06 0.43 0.11 - 63 - [−1.10, 0.57] [−1.31, 0.73] [−0.93, 0.82] [−0.12, 0.97] [−0.39, 0.61] Household Size −0.08∗∗ 0.14∗ −0.02 −0.01 −0.01 [−0.15, −0.01] [−0.01, 0.30] [−0.11, 0.07] [−0.06, 0.04] [−0.07, 0.05] Econ. Dep. Kids 0.15∗∗ −0.11 0.04 0.02 0.02 [0.03, 0.26] [−0.24, 0.03] [−0.07, 0.15] [−0.09, 0.13] [−0.09, 0.14] Married/Cohabit 0.19 0.51∗ 0.05 0.15 0.30 [−0.38, 0.75] [−0.04, 1.06] [−0.45, 0.55] [−0.23, 0.54] [−0.10, 0.70] Household Head 0.10 0.51∗∗∗ 0.13 0.02 0.18 [−0.36, 0.56] [0.14, 0.87] [−0.26, 0.52] [−0.33, 0.37] [−0.20, 0.57] Formal Sector 0.15 0.16 0.49 0.26 0.43∗ [−0.40, 0.69] [−0.31, 0.63] [−0.24, 1.23] [−0.18, 0.71] [−0.03, 0.90] Informal Sector 0.34 −0.09 0.86∗∗ 0.37 0.23 [−0.33, 1.02] [−0.91, 0.72] [0.08, 1.64] [−0.12, 0.85] [−0.25, 0.72] Self Employed −0.09 0.17 0.77∗∗ −0.04 0.02 [−0.60, 0.42] [−0.61, 0.94] [0.12, 1.42] [−0.47, 0.39] [−0.51, 0.54] Retired −0.06 −0.51∗ 0.60 0.20 0.12 [−0.63, 0.50] [−1.10, 0.08] [−0.18, 1.39] [−0.26, 0.65] [−0.36, 0.60] Unemployed 0.66 −0.03 0.58 0.61∗ 0.79∗∗∗ [−0.15, 1.48] [−0.70, 0.64] [−0.16, 1.32] [−0.02, 1.25] [0.30, 1.29] Urban Environment 0.21 −0.13 0.20 −0.12 −0.10 [−0.23, 0.65] [−0.58, 0.32] [−0.22, 0.62] [−0.49, 0.25] [−0.45, 0.26] 500 to 1,500 0.32 −0.08 0.17 0.13 0.16 [−0.23, 0.87] [−0.52, 0.37] [−0.40, 0.74] [−0.20, 0.45] [−0.14, 0.46] Higher 1,500 0.29 −0.18 0.25 0.77∗∗∗ 0.44∗ [−0.37, 0.94] [−0.94, 0.58] [−0.51, 1.01] [0.33, 1.21] [−0.04, 0.91] DRS 0.25 −0.55 0.50 −0.84∗∗∗ 0.43 [−0.55, 1.04] [−1.51, 0.41] [−0.54, 1.54] [−1.37, −0.32] [−0.21, 1.08] Khatlon 0.35 −0.27 0.18 −0.72∗∗ 0.35 [−0.37, 1.07] [−1.09, 0.56] [−1.01, 1.37] [−1.29, −0.15] [−0.24, 0.94] Sogd −0.47 −0.38 0.39 −0.14 0.69∗∗∗ [−1.13, 0.19] [−1.22, 0.45] [−0.72, 1.51] [−0.64, 0.35] [0.17, 1.21] - 64 - Gbao 0.77 −0.35 0.83 −0.91∗∗ 0.20 [−0.16, 1.71] [−1.14, 0.45] [−0.37, 2.04] [−1.67, −0.15] [−0.58, 0.98] Media Consumption 0.00 0.22∗∗ 0.09 −0.05 −0.05 [−0.18, 0.18] [0.04, 0.40] [−0.05, 0.23] [−0.18, 0.08] [−0.20, 0.10] Save as Child 0.53∗∗ 0.16 0.38 0.03 0.06 [0.07, 0.99] [−0.46, 0.77] [−0.13, 0.88] [−0.32, 0.38] [−0.26, 0.39] Observations 993 993 993 993 993 Signi�cance: ∗ (p < 0.10), ∗∗ (p < 0.05), ∗∗∗ (p < 0.01); 95% Con�dence Intervals in brackets. Reference Categories: female, cannot read or write Tajik, neither married nor living with partner, maximum primary education, not head of household, out of labor force, rural environment, income lower TJS 500, Dushanbe Region, not saved as a child