Improving Access to Banking: Evidence from Kenya. F. Allen, E. Carletti, R. Cull, J. Qian, L. Senbet, and P. Valenzuela April, 2013

Transcription

1 Improving Access to Banking: Evidence from Kenya F. Allen, E. Carletti, R. Cull, J. Qian, L. Senbet, and P. Valenzuela April, 2013

2 Motivation Sub-Saharan African countries have undergone extensive economic/financial sector reforms over the last few decades. Yet, many still face a severe financial development gap relative to advanced and other peer developing economies. A key obstacle to financial development is access by the disadvantaged to financial services, which could promote economic growth at the broadest scales. In view of the important role of financial inclusion, we study the expansion strategies of Equity Bank and other banks, and their impact on access to banking services in Kenya.

3 Why Financial Inclusion? Literature Our paper extends the emerging literature on financial access and inclusion in developing economies: Finance and growth Levine (2005); Rajan and Zingales (1998); Levine and various Link between finance and poverty alleviation Beck, Demirguc-Kunt, and Levine (2007) and Clarke, Xu, and Zou (2006). Outreach of banking services (cross-country measures) Beck, Demirguc-Kunt, and Martinez Peria (2007, 2008); Kendall, et al. (2010) Use of banking services in EMs (cross-country analysis) Beck and Brown (2011) Country studies (Mexico s Banco Azteca; Rural Banks in India; microfinance in Southeast Europe) Bruhn and Love (2009); Burgess and Pande (2005); Brown, Guin, Kirschenmann (2013)

4 Micro-evidence from Financial Diaries High-frequency surveys of poor households in Bangladesh, India, and South Africa based on cash flows (Collins, Morduch, Rutherford, and Ruthven, 2009) Every household in our 250-strong sample, even the very poorest, held both savings and debt of some sort. No household used fewer than four types of instrument during the year. Households are coping with incomes that are not just low, but also irregular and unpredictable, and too few financial instruments are available to effectively manage these uneven flows.

5 Why Kenya? Kenya s financial development indicators are among the best in the Sub-Saharan Africa region. Focusing on a specific country avoids the difficulties of cross-country studies regarding omitted (or poorly measured) country-level factors. A laboratory for the study of financial access as it has witnessed a general expansion in bank branching and the emergence of Equity Bank, which targets low income clients and traditionally under-served geographic districts.

6 Why Equity Bank? Equity Bank is a pioneering institution that devised a strategy targeting low income clients and traditionally under-served geographic areas. It is now the third largest listed company in Kenya by market capitalization. It does not have any government ownership. It has 55% of the deposit accounts in Kenya and 112 branches. It also operates in Uganda (42 branches) and South Sudan (3 branches), and is expanding in Tanzania and Rwanda.

7 Equity Bank s Business Model The business model of the bank is based on affordability and accessibility: Affordability: To open a bank account only requires an ID and a photo (which can be taken in the branch itself). By contrast, foreign banks, such as Barclays, require a minimum balance of about US$ 222. Accessibility: Local languages are spoken in its branches. In central Kenya, 30-40% people speak only a minority language.

8 Proximity: Bank is near/not too far Source: FinAccess Kenya % 55% 45% 40% 35% 30% 20% 0% Total N=4214 Currently banked n=721 Previously banked n=225 Never banked n=3268

9 Savings Account: Minimum Balance Source: Beck, Demirguc-Kunt, and Martinez Peria (2007) 80 Minimum Balance To Open Savings Account (% Of Gdppc) Cameroon Nepal Uganda Sierra Leone Kenya Georgia Sample size: 58 countries Lebanon Nigeria Ghana Madagascar Malawi Mozambique Armenia Moldova Philippines Zambia Albania Ethiopia Jordan India Sri Lanka Indonesia Hungary Pakistan Uruguay Lithuania Czech Republic Greece Colombia Croatia South Africa Bangladesh Bulgaria Bolivia Slovak Republic Romania Malta Dominican Republic Mexico Peru Trinidad and Tobago Thailand Brazil Bosnia and Herzegovina Belarus Slovenia Sweden Korea, Rep. Germany Turkey Switzerland Spain Israel Egypt, Arab Rep. Denmark Chile Belgium Australia

10 Annual Fees on Savings Account Source: Beck, Demirguc-Kunt, Martinez Peria (2007) 5 Nepal Annual Fee For Savings Account (% Of Gdppc) Malawi Uganda Kenya Lebanon Bolivia Sample size: 56 countries Cameroon Uruguay South Africa Indonesia Ghana Colombia Peru Chile Albania Bosnia and Herzegovina Georgia Mozambique Romania Mexico India Turkey Australia Egypt, Arab Rep. Spain Brazil Greece Slovak Republic Trinidad and Tobago Switzerland Sweden Sri Lanka Slovenia Sierra Leone Philippines Pakistan Nigeria Moldova Malta Madagascar Lithuania Korea, Rep. Jordan Israel Hungary Germany Ethiopia Dominican Republic Denmark Czech Republic Croatia Bulgaria Belgium Belarus Bangladesh Armenia

11 # Documents to Open Checking Account Source: Beck, Demirguc-Kunt, Martinez Peria (2007) Number Of Documents Needed To Open Checking Account Bangladesh Chile Trinidad and Tobago Zambia Nepal Sierra Leone Uganda Cameroon Kenya Ethiopia Nigeria Malawi Ghana South Africa Uruguay Turkey Indonesia Philippines Malta Colombia Australia Madagascar Armenia Mexico India Brazil Dominican Republic Pakistan Sri Lanka Lebanon Bolivia Greece Peru Moldova Croatia Jordan Korea, Rep. Slovenia Belgium Bosnia and Herzegovina Bulgaria Georgia Lithuania Hungary Slovak Republic Belarus Denmark Romania Thailand Israel Switzerland Sweden Spain Mozambique Czech Republic Albania Sample size: 56 countries

12 Research Questions Does Equity Bank pursue different branching strategies from other commercial banks? If so, do these expansion strategies lead to different effects on households access to banking services? And do such effects differ by type of household?

13 Main Findings While all banks (including Equity Bank) open a greater number of branches in urban, highly populated and English speaking districts, Equity Bank was more likely to expand to underdeveloped districts than other types of banks. Controlling for the total number of branches in a district, Equity Bank presence is strongly positively associated with residents likelihood of having access to both bank accounts and loans. This effect is particularly strong for individuals with low income and less education, from households that do not own a permanent house and have no member with a salaried job.

14 3500 A Snapshot of Financial Inclusion: Commercial Bank Accounts per 1000 Adults (Kendall et al., 2010)

15 Financial Inclusion: RCTs Credit: Mixed/weak results Banerjee et al. (2010) MFI branch expansion in India; Karlan and Zinman (2010, 2011) high-interest loans in the Philippines. Savings: More positive results Ashraf, Karlan, and Yin (2006) commitment savings in Philippines; Dupas and Robinson (2013) no-interest savings accounts, female shopkeepers in Kenya; Brune et al. (2011) commitment savings, farmers in Malawi. Limitations: Sample sizes are small, contexts differ, small variations in experimental design affect results (external validity). [Ravallion, 2009.]

16 RCT: Kenya Dupas, Green, Keats, and Robinson (2011) Villages surrounding three rural market centers in Western Kenya Paid account opening fees and provided minimum balance for random sub-sample of unbanked 63% opened account, but only 18% actively used it. Reasons for not saving in account: distrust of bank; unreliable service, withdrawal fees. Context matters: Scandal, risk of embezzlement listed as main reason for not having/using account. Equity not in these villages (KCB and Village Bank)

17 Data This paper combines a new dataset on bank penetration at the district level with household level survey data: Bank penetration dataset Based on branch-level information aggregated as a districtlevel panel on the number of branches by bank. Covers 45 commercial banks that operated in 65 Kenyan districts in the period between 2006 and FinAccess household surveys Conducted by Financial Sector Deepening Trust Kenya. 4,420 completed interviews in 2006 and 6,598 in Data on access to financial services and individual/household characteristics.

18 Branch Expansion: Urban districts Urban Districts Foreign banks Govt banks Local private banks

19 Branch Expansion: Low Population Density Districts 300 Low Population Density Districts Foreign banks Govt. banks Local private banks

20 Branch Expansion: High Minority Language Districts High Minority Language Districts Foreign banks Govt. banks Local private banks

21 Methodology Bank Expansion Strategies: Regression Analysis Branches dt = μ t + αx d + ε dt Use of Bank Accounts: Probit Model Pr(Banked idt = 1) = F(μ d + αequity dt + βbranches dt + γz idt ) Access to Credit: Ordered Probit Model The dependent variable takes the value: 0 = the individual does not have a bank account/loan, 1 = the individual has a bank account, 2 = the individual has a loan from a bank.

22 Concerns about Endogeneity Benchmarking good for comparisons with other banks But, bank expansion does not occur at random Greater entry into better developed districts with more growth opportunities Could be hard to identify well Equity s Contribution to the take-up of accounts

23 Additional Tests Instrumental variables Based on Equity s shift in banking strategy, business model % of people speaking a minority language in a district Difference-in-differences Look at districts with no Equity presence in 2006 Compare change in likelihood of account for (a) Districts Equity had entered by 2009 [treatment] (b) Districts without Equity presence 2006,2009 [control]

24 Branch Expansion Strategies Dep. variable: Number of branches over number of banks Foreign Banks Govt. Banks Private Domestic Equity Bank Rural ** *** ** Arid and Semi-Arid ** *** ** * log( Density population) * ** Language: Minority language ** *** ** *** Language: Swahili ** *** ** *** Dummy (2009=1) 0.118** 0.213*** 0.102** 1.045*** Constant *** *** 7.711*** *** Observations Adjusted R-squared

25 Branch Expansion Strategies Dep. variable: Number of branches over number of banks Foreign Banks Govt. Banks Private Domestic Equity Bank Rural ** *** ** Arid and Semi-Arid ** *** ** * log( Density population) * ** Language: Minority language ** *** ** *** Language: Swahili ** *** ** *** Dummy (2009=1) 0.118** 0.213*** 0.102** 1.045*** Constant *** *** 7.711*** *** Observations Adjusted R-squared

26 Branch Expansion Strategies Dep. variable: Number of branches over number of banks Foreign Banks Govt. Banks Private Domestic Equity Bank Rural ** *** ** Arid and Semi-Arid ** *** ** * log( Density population) * ** Language: Minority language ** *** ** *** Language: Swahili ** *** ** *** Dummy (2009=1) 0.118** 0.213*** 0.102** 1.045*** Constant *** *** 7.711*** *** Observations Adjusted R-squared

27 Branch Expansion Strategies Dep. variable: Number of branches over number of banks Foreign Banks Govt. Banks Private Domestic Equity Bank Rural ** *** ** Arid and Semi-Arid ** *** ** * log( Density population) * ** Language: Minority language ** *** ** *** Language: Swahili ** *** ** *** Dummy (2009=1) 0.118** 0.213*** 0.102** 1.045*** Constant *** *** 7.711*** *** Observations Adjusted R-squared

28 Branch Expansion Strategies Dep. variable: Number of branches over number of banks Foreign Banks Govt. Banks Private Domestic Equity Bank Rural ** *** ** Arid and Semi-Arid ** *** ** * log( Density population) * ** Language: Minority language ** *** ** *** Language: Swahili ** *** ** *** Language: Minority language x d *** Language: Swahili x d * 0.305** 0.144* 1.129** Constant *** *** 7.761*** *** Observations Adjusted R-squared

29 Branch Expansion Strategies Dep. variable: Number of branches over number of banks Foreign Banks Govt. Banks Private Domestic Equity Bank Rural ** *** ** Arid and Semi-Arid ** *** ** * log( Density population) * ** Language: Minority language ** *** ** *** Language: Swahili ** *** ** *** Language: Minority language x d *** Language: Swahili x d * 0.305** 0.144* 1.129** Constant *** *** 7.761*** *** Observations Adjusted R-squared

30 Use of Banking Services and Bank Presence Dep. variable: Bank account (1) (2) (3) (4) (5) Bank branches *** Equity Bank presence *** ** *** Foreign bank presence Gov. influenced banks presence ** * Local Private banks presence ** Foreign bank branches *** Gov. influenced bank branches Local private bank branches *** Observations Pseudo R-squared District Fixed Effects YES YES YES YES YES Control Variables YES YES YES YES YES

31 Timing the Relationship Between Bank Presence and Account Usage Dep. variable: Bank account (1) (2) (3) (4) 2009 Dummy *** Bank branches 0.067*** Equity Bank presence ** Foreign bank presence Gov. banks presence Local Private banks presence Foreign bank branches ** Gov. bank branches Local private bank branches Bank branches* *** Equity Bank presence* *** ** *** Foreign bank presence* Gov. banks presence* Local Private banks presence* Foreign bank branches* Gov. bank branches* Local private banks branches* Observations Pseudo R-squared District Fixed Effects YES YES YES YES Time Fixed Effects YES YES YES YES Control Variables YES YES YES YES

32 Rural, Semi-Arid, and Arid Districts Dep. variable: Bank account (1) (2) (3) (4) (5) Bank branches *** Equity Bank presence *** *** *** Foreign bank presence Gov. influenced bank presence *** ** Local private bank presence ** * Foreign bank branches *** *** Gov. influenced bank branches Local private bank branches Observations Pseudo R-squared District Fixed Effects YES YES YES YES YES Control Variables YES YES YES YES YES

33 IV, Diff-in-Diff Results IV GMM IV Probit Difference in Differences Dep. variable: Bank account (1) (2) (3) (4) (5) First stage Minority language district *** *** Minority language district* *** *** Second stage Bank branches *** *** Equity Bank presence *** *** ** ** ** Foreign bank presence Gov. banks presence ** Local Private banks presence Foreign bank branches Gov. bank branches Local private bank branches Observations Centered R-squared Partial R-square of excluded IVs p-value of excluded instruments p-value of Hansen J statistic Pseudo R-squared District Fixed Effects YES YES YES YES YES Control Variables YES YES YES YES YES

34 Who Benefits from Equity s Services? Dep. variable: Bank account (1) (2) (3) (4) (5) Equity Bank presence *** *** *** *** *** Asset score x Equity bank presence ** * Own house x Equity bank presence *** ** Sec.-Ter. x Equity bank presence *** ** Waged x Equity bank presence ** ** Observations Pseudo R-squared District Fixed Effects YES YES YES YES YES Control Variables YES YES YES YES YES

35 Access to Credit Ordered Probit Average Marginal Effects Financial Financial Financial Access=0 Access=1 Access=2 Bank branches *** *** *** *** Equity Bank presence *** *** *** *** Observations Pseudo R-squared District Fixed Effects YES Control Variables YES

36 But is it profitable?

37

38

39 Concerns and caveats (Have we painted too rosy a picture?) Over half of savings accounts in one bank Small amounts, but vital to owners Implications for deposit insurance system Expansion beyond Kenya Would more competition for this segment help or hurt? Expanding loans likely to be more lucrative going forward. Risky?

40

41 Conclusions We study the expansion strategies of Equity Bank and other banks, and their impact on access to banking services in Kenya in view of the vital role of financial inclusion in developing economies. Kenya offers a laboratory for the study of financial access, especially in light of the emergence of Equity Bank. The presence of domestic private banks (including Equity), government or government-influenced banks, or foreign banks is associated with increased financial access at the district level.

42 Conclusions 2 Controlling for branching by other banks, Equity Bank presence is strongly associated with greater usage of bank accounts and bank loans. Especially true for individuals with low income and less education, from households that do not own a permanent home and that have no member with a salaried job. Jury is still out on sustainability of Equity strategy. Replicable in other countries? Is Kenya special?

43 Variable Obs Mean Std. Dev. Min Max Obs Mean Std. Dev. Min Max Bank account Bank Loan Household size Education: Primary Education: Secondary Education: Tertiary Female Asset score Own permanent house Income type: Agriculture Income type: Waged Income type: Business Age group: Age group: Age group: Age group: Age group: Language: Swahili Language: Minority language Urban Rural Arid and Semi-Arid Total bank branches Bank branches (excluding Equity Bank) Equity ban branches Equity bank presence