Lecture notes on financial access. Rajeev Dehejia

Size: px

Start display at page:

Download "Lecture notes on financial access. Rajeev Dehejia"

Jessie Davidson
6 years ago
Views:

1 Lecture otes o fiacial access Rajeev Dehejia

2 Today The empirical ladscape Iformal fiace Istitutios Mechaisms Demad Market failure Adverse selectio Moral Hazard Makig bak available

3 The empirical ladscape

4 The Baked: Fractio of Households with a Accout i a Fiacial Istitutio Source: World Bak, Databak

5 Half the world is ubaked Half the world is ubaked. Fiacial Access Iitiative.

6 Formal fiacial access icreases with icome. World Bak, Fiace For All. Households usig fiacial services (percet) GDP per capita (thousads of 2000 USD) 6

7 Access versus use Defiig access is ot easy World Bak, Fiace For All. Usage is much easier to measure Need to collect idicators o: Actual use of various services (savigs, paymet, credit) Barriers to access, to idetify boudaries ad causes of exclusio 7

8 Access versus use World Bak, Fiace For All. Populatio Users of fiacial services Volutary selfexclusio No eed Cultural / religious reasos/ /idirect access No- users of fiacial services Ivolutary exclusio Isufficiet icome / high risk Discrimiatio Cotractual / iformatioal framework Access to fiacial services No access to fiace Price / product features 8

9 Barriers to access World Bak, Fiace For All. Through bak surveys idetify barriers to Opeig ad maitaiig a accout Applyig/processig, gettig a loa Payig bills, makig moey trasfers Alog differet dimesios Physical barriers: braches, ATMs, Eligibility: documetatio, paperwork, procedures Affordability: iterest rates, fees, miimum balaces 9

10 Physical access World Bak, Fiace For All. Brach ad ATM peetratio, by icome quitile Number per 100,000 populatio 60 Number of bak braches Number of ATMs Icome quitile 10

11 Eligibility documetatio World Bak, Fiace For All. Number of documets required to ope a checkig accout Greater tha 4 documets, 11% 1 documet, 9% 3 to 4 documets, 24% 1 to 2 documets, 27% 2 to 3 documets, 29% 11

12 Affordability miimum balaces World Bak, Fiace For All. Miimum balace required to ope a checkig accout (percet of GDP/capita) 15-50% required 10% Greater tha 50% required 10% No miimum balace required 32% 5-15% required 12% 1-5% required 17% 0-1% required 19% 13

13 ad fees World Bak, Fiace For All. Costs of trasferrig fuds abroad (percet of $250, a typical remittace) Greater tha 25$ 11% Less tha 5$ 11%, 5 to 12$ 27% 12 to 25$ 51% 14

14 Barriers ad fiacial exclusio World Bak, Fiace For All. Share of populatio uable to afford checkig accout fees Malawi Ugada Sierra Leoe Keya Swazilad Nepal Cameroo Chile Madagascar Ghaa South Africa Percet 15

15 Iformal istitutios servig the poor Iformal commercial leders: Moeyleders, traders, ladlords etc. They have rich iformatio ad effective eforcemet but limited resources Easy access to credit Very high iterest rates e.g. 10%-100%pm Do ot provide other fiacial services e.g. savigs

16 Formal istitutios dealig with the poor State owed Baks: Regioal Rural Baks, Specialized Baks, IRDP, BAAC, foreig aid After WWII most low icome coutries tried to develop agri sector rural fiace became crucial Heavy subsidies to compesate for losses due to high trasactio costs ad iheret risks Subsidized credit, capped roi e.g. Phillipies Siphoed off by local elites, do ot reach poor High default rates, high arrears, large losses Iefficiet ad corrupt

17 A example: IRDP s i Idia Allocated credit accordig to social targets 30% to SC ST, 30% to wome Subsidies betwee : $6b: 25-50% of loa volume Poor istitutioal performace: Repaymet rates fell below 60% Oly 11% borrowers took a secod loa I 2000 repaymet rate was 30% Devastatig criticisms of state-led developmet bakig Credit is ot like seeds ad fertilizers Subsidies made it worse weakeed icetives, created moopolies, udermied poverty reductio.

18 Iformal fiace: Mechaisms

19 Stuart Rutherford: Savig up ad savig dow

20 Stuart Rutherford From The Poor ad their Moey Mary s ROtatig Savigs ad Credit Associatio from a Nairobi slum MERRY GO ROUND Mary is a sigle mother who sells vegetables. Her stock is worth about 1500 shilligs. Whe she has to ru her stock dow because of some eed for a lump sum such as medicie for her so she ca re-stock withi 15 days from her ROSCA. 15 members each cotribute100 Keya shilligs/day. Each day oe takes 1,500 shillig pot

21 Stuart Rutherford From The Poor ad their Moey Jyothi a deposit collector from Idia Jyothi s fee Slum wome are very kee to use a deposit collector like Jyothi. Where else ca they save up the usefully large sums they so ofte eed? Jyothi collects a set amout from her cliets every day for 220 days. At ed, cliets get back value of 200 days savigs.

22 A urba moey leder from Idia sequece of small repaymets resembles the deposits made by Mary s ROSCA members ad Jyothi s cliets. Loa repaymet is based o the ability to save, ad loas are essetially advaces agaist future savigs. Stuart Rutherford From The Poor ad their Moey The iterest deductio URBAN MONEYLENDER Weekly repaymets for 10 weeks. Moeyleder deducts fee (or iterest) up frot

23 Iformal fiace: pet-up demad What proportio of respodets applied for a loa? Brazil Colombia Mexico Idia 15% 9% 14% 3% Did the baks award them loas? Brazil Colombia Mexico Idia (formal ad iformal) Approved 68% 72% 75% 20% Source: Ajali Kumar et al Access to Fiacial Services - What Do we Kow Across Coutries? Prelimiary Comparisos. Presetatio at World Bak Fiace Forum 2004

24 Idoesia MASS 2002 Survey Survey covered a radomized sample of 1438 Idoesia households i six provices, completed by Bak Rakyat Idoesia i July/August Coverage of six provices: West Java, East Java, West Kalimata, East Kalimata, North Sulawesi, ad Papua. Provices icluded 20.6 millio households ad 85 millio people. Eumerators were BRI loa officers ad other professioals, offerig the chace to assess the creditworthiess of both customers ad o-customers usig the stadard procedures applied by the bak. Study used the household survey to revisit claims made largely o the basis of admiistrative data ad aecdote.

25 Loa Uses Busiess: workig capital of existig veture, Diversify icome, Startig a ew busiess, Purchasig ew equipmet, New busiess ifrastructure (e.g., store or warehouse), Busiess ifrastructure improvemet. Household: Home improvemet, No-busiess lad or buildig purchase, School tuitio, Medical treatmet, Loa repaymet, Meetig daily eeds or retiremet eeds, Vehicle purchase, Buyig household goods, Ceremoy or social expediture, Holiday eeds Jewelry purchase.

26 Uses of loas (percet) withi icome groups. Borrowers from ay source Loa use Below poverty lie Per capita icome is 1 to 3 times the poverty lie Per capita icome is more tha 3 times the poverty lie Busiess Household Other Observatios

27 Loa Sources BRI Uit Borrowers have take loas from the microfiace arm of Bak Rakyat Idoesia. Other formal bak borrowers have take loas from other formal sources icludig the BRI brach offices, Bak Cetral Asia (BCA), Bak BNI, a local developmet bak, Bak Daamo, Bak Madiri, Bak Bukopi, a Sharia commercial bak, other private commercial bak, Bak Perkredita Rakyat (BPR), or a Sharia rural bak. Microbak borrowers have borrowed from a rural credit agecy (BKD/TPSP/LDKP), credit uio/cooperative, rural uit cooperative (KUD), BMT/BMM Islamic istitutio, market bak, local fiacig istitutio, or govermet bureau. Iformal sources iclude pawshop service, joit veture, a selfmaaged istitutio, professioal moeyleder, family/relative/frieds, or other iformal source.

28 Would the eumerator make a loa to this household? % Poor Bottom 50% 49 Top 50 % 43 Not poor 1-2 times z times z times z times z 80 > 7 times z 88

29 The advatages of moeyleders Moeyleders (vs. formal baks) have good iformatio o who is who ad who did what i the eighborhood. Moeyleders ted to live ad work amog the people to whom they led. They build up histories over time. They are able to eforce cotracts through social, ad sometimes physical, meas, eve whe cotracts have ot bee writte dow ad whe a court would refuse to hear a case. The lack of collateral is ot a problem; it may eve provide a profitable busiess opportuity. Ad their cliets, who ofte have owhere else to tur, accept high charges willigly if ot happily. But moeyleders have oly limited amouts of moey, ad, because their iformatio is rich but very localized, moeyleders ca oly led i small amouts to a arrow bad of cliets.

30 Why do credit markets fail? Cotext Limited liability due to lack of collateral Weak eforcemet of cotracts Limited property rights These problems the facilitate: Adverse selectio: hidde iformatio Moral hazard: hidde actio

31 Adverse selectio ad moral hazard Basic problem: Raisig iterest rates reduces average loa repaymet rates. Adverse selectio: safe borrowers are screeed out, risky borrowers optio i. Moral hazard: beefit of safe behavior falls, less eed to exert effort for success of project Implicatio: May ot be able to fid profit-makig iterest rate

32 Figure 2.1: Adverse selectio example (a). At iterest rates betwee k+a ad y the bak ears a profit ad both safe ad risky types wat to borrow. Safe types leave the market oce iterest rates rise above y, ad the bak loses moey. Oce iterest rates are pushed up to k/p, the bak ca agai ear profits, while servig oly risky borrowers. At iterest rates above y-bar eve the risky borrowers leave the market. Expect ed Pro fit Bak profits. Safe ad risky borrowers gai access Bak profits. Oly risky borrowers served Thik of k as the risk-free rate. A is the premium eeded above the risk-free rate to break eve with all borrowers. (+) 0 ( -) Loss. Iterest rate too low Loss 0 k k+ A y k/p / y Iterest rat e p With oly risky borrowers to break eve you eed p r + (1-p) 0 = k where p is the prob of risky borrowers repayig ad r is the risky iterest rate or r = k/p

33 / p Figure 2.2: Adverse selectio example (b). Here, the risky types are riskier tha i example (a) i Figure 2.1. Now the safe types ca ever be served by a bak aimig to break eve (sice profits are egative eve at iterest rate y). The bak must raise rates to k/p to ear profits, at which price the bak will oly attract risky borrowers. At iterest rates above y-bar, the risky borrowers leave the market. Exp ected Pr ofit (+ ) No iterest rate at which safe borrowers ca be served while the bak ears expected profit 0 (-) 0 k y k+a k/p I tere st rate y

34 Adverse selectio Expected profit Worseig expected quality of customers Iterest reveue up Is this profit level high eough to make busiess sese? Iterest rate

35 Eough collateral: riskless world 1. Post sufficiet collateral 2. Take loa 3. Work hard. Ear profit/loss 4. Repay loa, keep et Assets Debt Icome

36 Eough collateral: risky world (+) 0 (-) Proposed project Alterative project 1. Compare expected gross icome from 2 risky projects (with expected value of first project > secod project). 2. Repay loa, keep et icome

37 Isufficiet collateral (+) 0 Agecy ret Agets lackig collateral to credibly pledge to absorb losses i bad times must istead ear agecy rets that makes them more risky ad costly to cotract with. 1. Compare expected gross icome from 2 risky projects. 2. Repay loa, keep et icome (-) Proposed project Alterative project Upaid debt

38 Isight from moder corporate fiace 1. Ski i the game. Agets who ca affect the value of the firm through costly uverifiable actios eed to share i both fortues ad losses to have icetives to maitai the value of the firm ad the claims it sells to outsiders. 2. Iefficiecies. Agets that do ot have additioal (or collateral) resources to offer as guaratee to credibly pledge to absorb losses i bad times must istead ear agecy rets that makes them more risky ad costly to cotract with.

39 Reproductio of poverty Thus, all else equal, idividuals with low et worth: More likely to be excluded from capital markets Or to obtai fiace o less favorable terms tha others.

40 Makig bakig available: strategies Problem: lack of collateral impedes credit access Solutios: Secure property rights Public sector baks Social busiess (microfiace --ext time)

41 Securig property rights If the problem is that the poor lack collateral, the why ot give them collateral? How? Grat property rights to lad they occupy. Also icreases the icetives to make improvemets. Also icreases labor supply (willig to leave the house / do t eed to stay home to protect their claim).

42 Burgess ad Pade Examie a bak expasio i Idia. Some issues similar but they ca also look at effect of bak expasios o poverty. Idia atioalized baks i 1969 ad the required them to expad ito uder-baked areas: Baks were restricted to selectig ubaked locatios for brach expasio from a list circulated by the cetral bak. This list idetified all ubaked locatios with a populatio above a certai umber, ad was updated (with a lower populatio cut-off) every three years. The same populatio cut-off was used across Idia, ad therefore the list featured relatively more locatios from states with a lower iitial stock of bak braches per capita. Havig a commo criteria for idetifyig ubaked locatios also implied that, withi a state, more locatios were targeted i districts with fewer bak braches per capita pre-program.

43 Bakig policy 1:4 policy: 4 braches i uder-baked area for oe i a area with baks already. To esure that rural brach expasio traslated ito icreased credit ad savigs opportuities for the rural populatio the cetral bak regulated baks s deposittakig ad ledig policies. Policy was i place , ad cacelled after that. Use this a atural experimet for bak expasio.

44 First stage Chart shows that rural bak locatios positively correlated with the iitial umber braches, the durig policy correlatio turs egative.

45 Reduced form TABLE 7:BANK EXPANSION PROGRAM AND POVERTY: REDUCED FORM EVIDENCE Head cout ratio Rural head cout Urba head cout Rural-Urba head Log real agri. ratio ratio cout diff wage (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) Number baked locatios *** *** *** *** ** *** *** i 1961 * Tred [43.44] [51.28] [51.42] [59.29] [36.19] [47.32] [54.00] [65.28] [1.26] [1.15] Number baked locatios *** *** *** *** *** *** -4.55*** -6.44*** i 1961*Post-77 Tred break [52.63] [58.47] [63.80] [68.24] [48.69] [65.54] [72.54] [84.91] [1.52] [1.55] Number baked locatios *** *** *** *** *** ** *** 15.98*** i 1961*Post-90 Tred break [105.18] [107.40] [126.75] [131.91] [81.02] [79.42] [129.24] [144.43] [2.42] [2.49] State dummies Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Year dummies Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Cotrol variables No Yes No Yes No Yes No Yes No Yes Adjusted R-squared Number observatios Locatios gettig less baks as a result of the policy (i.e., those that that had more iitial baks i 1961) show greater poverty. Notes: Robust stadard errors i paretheses. The depedet variable is umber of baked locatios i a state. Locatios gettig more baks as a result of the policy (i.e., those that that had more iitial baks i 1961) show greater poverty.

46 Dehejia ad Gupta Look at cross-sectioal locatio of braches i 1991 ad ask whether locatio of braches i 1991 affects service sector ad educatio choices from They fid: Reduced probability of beig household employer or worker. More likely to be a formal employee. Wages icrease i large firms. More likely to be a studet.

47 Dupas, Karla, et al Bakig the ubaked : Gave access to basic bak accouts i Ugada, Malawi, ad Chile. Icremetal deposits were 17, 10, ad 3%. No impact o dowstream outcomes. Suggests that access to basic accouts has little micro impact. Does t preclude macro chaels.

48 Takeaways A vast umber of poor are ubaked. The poor use loas for more tha busiesses. Raisig iterest rates to cover costs of bakig to poor exacerbates moral hazard / adverse selectio. Poverty = lack of collateral = precludes usual solutio to MH/AS = exacerbates MH/AS Baks ca be part of the solutio whe their profit motive is retaied but rechaeled.

Development Economics. Lecture 22: Credit Markets Professor Anant Nyshadham ECON 2273

Development Economics. Lecture 22: Credit Markets Professor Anant Nyshadham ECON 2273 Developmet Ecoomics Lecture 22: Credit Markets Professor Aat Nyshadham ECON 2273 This lecture 1. Lumpiess ad time 1. Savigs 2. Isurace 3. Credit 2. Characteristics of credit markets for the poor 2 Lumpiess