M E R 82401 U T BE N NO 12 OCTOBER 2013 FINDEX The Global Findex Database NOTES Islamic Finance and Financial Inclusion Compared with non-Muslims, self-identified Muslims are significantly less likely to own a formal account or save at a formal financial institution, after controlling for other individual and country characteristics, according to Asli Demirguc-Kunt recent research using Global Findex data. But the authors find no evidence Leora Klapper that Muslims are less likely to borrow formally or informally. Data from an additional survey of adults in five North African and Middle Eastern Douglas Randall countries with relatively nascent Islamic finance industries shows very little use of Sharia-compliant banking products, though there is evidence of a hypothetical preference for Sharia-compliant products among a plurality of respondents despite higher costs. The Global Findex database and related research can help Islamic finance industry leaders expand their outreach to Muslim clients, and help policymakers more clearly define their role in expanding financial inclusion—Islamic or otherwise—to Muslim adults. WWW.WORLDBANK.ORG/GLOBALFINDEX Recently, much attention has been given to the tremendous growth of the Islamic finance industry and the potential of Sharia-compliant financial services to expand financial inclu- sion among Muslim adults. Yet there has been a scarcity of empirical research that measures the degree to which Muslims are not accessing conventional financial systems, and how much they demand and use Sharia-compliant financial products, particularly within the realm of household finance. Even less is known about how these usage gaps and preferences vary between various financial products and across regions and countries. This note summarizes recent research by Demirguc-Kunt, Klapper, and Randall (2013) that explores whether Muslims are less likely than non-Muslims to use formal financial services and, in a smaller sample of countries, the usage and awareness of Sharia-compliant banking products. The case for expanding Sharia-compliant products is often based on the relatively low levels of financial inclusion in Muslim-majority economies, yet these claims confound religious identification with other individual- and country-level attributes. Individual-level, demand-side data are required to isolate the relationship between Muslim identity and preferences for, and usage of, formal financial services. The Global Findex database pro- vides such data, measuring how people in 148 economies around the world save, borrow, make payments, and manage risk. These new indicators are constructed with survey data from interviews with more than 150,000 nationally-representative and randomly-selected adults age 15 and above. The survey was carried out over the 2011 calendar year by Gallup, Inc. as part of its Gallup World Poll. This note features Global Findex data based on nearly 65,000 interviews across 64 economies, as well as data from an additional survey in five North African and Middle Eastern countries (map 1). Differences in Account Ownership between Muslims and Non-Muslims The gap in account penetration between Muslims and non-Muslims is large and statistically significant, even when controlling for other individual- and country-level characteristics. In order to examine variation between Muslims and non-Muslims within economies, the analysis excludes countries where less than 1 percent or more than 99 percent of the adult FINDEX NOTES 1 M 1 AP Economies included in the analysis Economies included in Demirguc-Kunt, Klapper, and Randall (2013) Source: Demirguc-Kunt, Klapper, and Randall 2013. population identify themselves as Muslim.1 According to data covering 64 countries—and approximately 75 percent of the world’s adult Muslim population—24 percent of Muslim adults report having an account at a bank or formal financial institution, compared with 44 percent of non-Muslims.2 Multivariate regression analysis suggests that although much of this difference is related to other individual- and country-level characteristics, self- identification as a Muslim is associated with a 6 percent decrease in the probability of hav- ing a formal account.3 The gap in account penetration between Muslims and non-Muslims persists across most regions, except for East Asia and Pacific economies, where there is no significant difference in account penetration between the two populations. In the sample of economies worldwide, just 7 percent of unbanked Muslims and unbanked non-Muslims cite religion as a barrier to account ownership. Similar to non-Muslims, Muslims are most likely to cite cost, distance and documentation as barriers to account ownership. In regression analysis, however, there is some evidence that unbanked Muslims are more likely than unbanked non-Muslims to report not having an account at a formal financial institution because of religious reasons. But this finding does not hold consistently across different regression specifications, and it appears to be driven largely by a handful of economies, particularly those in Sub-Saharan Africa. The significant and economically meaningful gap in account penetration between Muslims and non-Muslims, combined with the generally insignificant gap in reporting religion to be a barrier to account ownership, suggests that constraints may be supply-driven—possibly by discrimination or the relative scarcity of financial services in predominantly Muslim areas. Unfortunately, the currently available data do not allow formal empirical tests of these hypotheses. Differences in Borrowing between Muslims and Non-Muslims Nine percent of non-Muslims and 7 percent of Muslims report having borrowed money from a bank or another formal financial institution in the past 12 months. The differences in formal borrowing behavior between Muslims and non-Muslims are considerably smaller than those related to account ownership, and are not statistically significant in the complete sample, after controlling for other individual- and country-level characteristics. Differences in borrowing from a range of other sources—family, friends, retail outlets using store or installment credit, an employer, or an informal private lender—are also not found to be 2 FINDEX NOTES statistically significant in regression analysis. When examined on a regional level, however, some differences do emerge. According to mul­ tivariate regression analysis, Muslims are significantly more likely than non-Muslims to have borrowed money from formal financial institutions in high income economies. In contrast, Muslims are significantly less likely to do so in the four East Asian and Pacific countries in the sample.4 An important caveat to the above findings is that Global Findex data does not distinguish between conventional and Sharia-compliant financial products. It is possible that the absence of a gap in borrowing behavior is the result of widespread availability and use of Sharia-compliant products. However, given that less than 1 percent of total global financial assets (and less than 1 percent of microfinance products) are from Islamic financial institu- tions, that is unlikely to be the case. Rather, it seems more plausible that a vast majority of financially included Muslims use conventional banking products and services. Usage of, and Preferences for, Sharia-Compliant Financial Services The analysis supplements Global Findex data with an additional questionnaire on the awareness, use, and preference for Islamic financial products, which was included in the 2012 Gallup World Poll. This module was included in Algeria, Egypt, Morocco, Tunisia, and Yemen with approximately 1,000 respondents per country. In each of those countries, with the exception of Egypt, 100 percent of respondents self-identified as Muslim.5 In Egypt, 95 percent of respondents did so. An important caveat is that these results are not representative of Islamic banking globally; in many countries, use of and attitudes towards Islamic finance are likely markedly different. Across the five countries, 48 percent of respondents report having heard of Sharia-compliant products in their country that offered services to people like them (table 1). This ranges from 35 percent in Algeria to 57 percent in Tunisia. However, just 2 percent of respondents report currently using a Sharia-compliant banking service. In no country does this value exceed 3 percent. Among those who separately report having an account at a formal financial institution or having borrowed from a formal financial institution in the past year, just 8 percent report currently using an Islamic banking service. This is consistent with the low concentration of Islamic banking services in these countries as reported by Beck et al. (2013), which shows that in Egypt and Tunisia, Islamic banks’ share of total bank assets is 4.2 and 1.5 percent, respectively. Islamic banks’ share of total bank assets in Yemen is 51.5 percent, though this is mostly concentrated in non-commercial assets. Regression analysis suggests that income and access to information are strongly and positively associated with awareness and use of Sharia-compliant banking products. LE TA 1 B Sharia-compliant Financial Services Hypothetical loan choice Prefers Currently (more Prefers Has heard uses an expensive) (cheaper) Does not about Islamic loan from loan from have a Sample Islamic banking Islamic conventional preference Don't know/ size banks (%) service (%) bank (%) bank (%) (%) Refuse (%) All 5,071 48 2 45 27 10 17 Algeria 1,022 35 3 49 27 22 3 Egypt, Arab Rep. 1,020 49 3 0 0 0 0 Morocco 1,000 41 1 54 16 13 17 Tunisia 1,029 57 2 31 40 12 17 Yemen, Rep. 1,000 53 1 37 18 22 22 Note: “All” averages are weighted by economy-level adult population. Due to survey execution errors, data from Egypt, Arab Rep. are not included in the hypothetical loan choice summary statistics. FINDEX NOTES 3 The third and final question in the module examines the significantly higher costs, an almost equally large share of robustness of demand for Islamic banking services.6 The adults prefer the cheaper, conventional loan or do not have a question puts forth a hypothetical scenario, in which the preference between the two. This suggests that there is likely respondent has been approved for a one-year loan from a to be demand for both conventional and Islamic banking conventional bank and an Islamic bank. The value of the services, and that preferences for Sharia-compliant products hypothetical loan is equivalent to 15 percent of the GDP per are influenced by price. capita of the respondent’s home country. However, the loan from the Islamic bank comes with an effective 20 percent Conclusion interest rate, while the loan from the conventional bank There is wide scope for future research on this topic. Additional comes with an effective 15 percent interest rate. The price research is needed to investigate whether the differences difference is meant to test the price sensitivity of respondent’s that exist between Muslims and non-Muslims in the use of potential preference for Islamic products. In reality, the price financial products are demand- or supply-driven. The results difference between conventional and Islamic loans is often from Demirguc-Kunt, Klapper, and Randall (2013) also raise less than 5 percentage points. The interest rate is presented the question of whether a gap exists between Muslims and in terms of the value of the monthly payment, since explicit non-Muslims in the ownership of formal accounts but not in rates are not Sharia-compliant. The respondent is asked to formal credit products due to divergent “urgencies of need,” choose which loan he/she preferred. with respect to savings and payments versus borrowing. Additional cross-country, demand-side data (particularly in In the sample, 45 percent of respondents reported a preference countries where the Islamic finance industry is more developed) for the more expensive Islamic bank loan, while 27 percent of on the use of, and preferences for, Sharia-compliant finance respondents reported a preference for the cheaper, conventional products would also be valuable in better understanding the bank loan. Respondents in Morocco were most likely to opt variation in the demand for Sharia-compliant finance products for the Islamic bank loan (54 percent), while respondents among Muslim adults. Survey instruments that can vary in Tunisia were most likely to choose the conventional loan price and other hypothetical product features would allow (40 percent). There was relatively little variation in choice by researchers to determine elasticities of demand for certain within-country income levels. Men were significantly more financial products, Sharia-compliant and otherwise. Finally, likely than women to choose the Islamic option (48 percent time-series data that can track the development of Islamic vs. 43 percent), with the largest gender gap in Egypt. finance industries across countries and the accompanying Taken together, the results from the Islamic finance module shifts in demand-side usage of, and attitudes toward, Sharia- suggest that individual-level use of Islamic banking services compliant financial products would provide insight into is extremely low in these countries. This is consistent with the relationship between Islamic finance and the broader supply-side data. And while a plurality of respondents do patterns of financial inclusion. report a preference for Islamic banking services despite WWW.WORLDBANK.ORG/GLOBALFINDEX 1. See Demirguc-Kunt, Klapper, and Randall (2013) for more details. References 2. Thirty-six percent of respondents in the sample self-identify as a Beck, T., A. Demirguc-Kunt, and O. Merrouche. 2013. “Islamic vs. Muslim. 17 percent of respondents in the complete worldwide conventional banking: Business model, efficiency and stability.” Global Findex dataset report the same. Journal of Banking & Finance 37(2): 433-447. 3. Even after controlling for other individual-level and country-level dif- Demirguc-Kunt, A., and L. Klapper. Forthcoming. “Measuring Financial ferences, the authors do not interpret the observed correlations as Inclusion: The Global Findex Database.” Brookings Papers on causal effects. Of particular concern is the omitted variable critique Economic Activity. in which an omitted variable causes individuals both to identify as a Muslim and to behave in a certain way. The omitted variable Demirguc-Kunt, A., L. Klapper, and D. Randall. 2013. “Islamic Finance can be transmitted from parents to children (either genetically or and Financial Inclusion: Measuring Use of and Demand for through education). Formal Financial Services among Muslim Adults.” Policy Research Working Paper 6642, World Bank, Washington, DC. 4. Indonesia, Malaysia, Philippines, and Thailand. 5. For this reason, Algeria, Morocco, and Yemen are not included in the main analysis. Tunisia was not surveyed by Gallup in 2011. The reference citation for the Global Findex data is as follows: 6. Due to the omission of the “No preference” option in the Egypt survey, Demirguc-Kunt, A., and L. Klapper. Forthcoming. “Measuring Financial we do not include data from Egypt in the analysis of this question. Inclusion: The Global Findex Database.” Brookings Papers on Economic The omission was the result of a clerical error. Activity. 4 FINDEX NOTES