Building Better Products through Randomized Evaluations Dean Karlan Yale University Innovations for Poverty Action IPA Grupo de Análisis para el Desarrollo GRADE Financial Access Initiative FAI ID: 496817
Download Presentation The PPT/PDF document "Microfinance: " is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Microfinance: Building Better Products through Randomized Evaluations
Dean KarlanYale University,Innovations for Poverty Action (IPA), Grupo de Análisis para el Desarrollo (GRADE),Financial Access Initiative (FAI)
1Slide2
How do we Reach 1 Billion?Is it flexibility?Is it price?
Is it institutional?What impact can we hope for?Massive untapped market: Why are so many yet to be reached?Slide3
3
prolifically The Dream SlideSlide4
1. Impact of finance (versus no finance)2. How to do microfinance better
4Two types of studiesSlide5
Key questions:Flexibility and priceMarket failuresFour examples
Group versus individual liabilityCredit with educationCredit scoringSavings product designTheme: Making the research speak to the practitioner.Plan5Slide6
Why despite our best efforts are so many not reached?Is it flexibility?Is it price?Is it institutional (organization structure, financing, human resource policy, etc.)?Can we rule out lack of impact?
Flexibility? Price?6Slide7
Cash flows do not match cash flow needs in many cases.FarmersFluctuation in incomeIs it fear?
Afraid of not having money to repay (irony: this is a GREAT client! she is so trustworthy that she won’t even take out a loan due to her fear of going into default!)Lack entrepreneurial skills to expand businessFears peer punishmentWhat product designs can alleviate these concerns? Savings? Insurance?Is It Flexibility?7Slide8
Strikingly little evidence on price.The “old” line: price does not matterRecent work challenges this:Dehejia, Montgomery & Morduch from BangladeshKarlan and Zinman from South Africa
Clearly, elasticity of demand depends on a lot:CompetitionBusiness opportunitiesFinancial literacyFraming of offer (this can matter more than price)Is It Price?8Slide9
Market FailuresThree basic questions for understanding credit markets and formulating policy:
Are there market failures, and can we specify more precisely what is happening (adverse selection, moral hazard, etc.)?“Observing Unobservables” with Zinman finds evidence of both What innovations can solve these market failures?Joint liability?“Group versus individual liability” with Xavier Gine finds no differenceCredit bureaus?work by deJanvry, McIntosh and Sadoulet in Guatemala
Dynamic incentives? Working, and in progress
Better screening? Credit scoring?
Can other non-credit interventions matter?
Teaching Entrepreneurship, with Martin Valdivia, GRADE-Peru
What is the welfare improvement from solving these market failures?
Little data on this… earlier work from Bangladesh (Pitt &
Khandker
) and India (Burgess &
Pande
) say yes, but identification and micro-data always a severe challenge.Slide10
Group/Individual Liability, Philippines and BoliviaJoint work with Xavier Giné, World Bank
Credit with Education, PeruJoint work with Martin Valdivia, GRADECredit Scoring, South Africa and PhilippinesJoint with Jonathan ZinmanSavings Product DesignPeru, India and the PhilippinesJoint with Sendhil Mullainathan and Jonathan ZinmanFour Examples10Slide11
Microfinance is typically seen as a solution to credit market failures faced by the poorGroup liability, a feature found in many micro loans, is perceived as a key innovation that has contributed to this successe.g.: Grameen, FINCA,
AccionGroup versus Individual Liability11Slide12
Yet, in recent years, many micro-lenders have expanded rapidly using individual liabilityIn turn, this has motivated other lenders that were using group liability to shift to individual liability
Motivation12Slide13
Green Bank of Caraga in the Philippines170 joint liability Grameen-style centers80 randomly assigned to convert to individual liability centers, but weekly meetings remained intact (“treatment”)90 randomly assigned to remain as-is, under joint liability
Group versus Individual Liability13Slide14
Screeningfuture study (a little here)Monitoring + Enforcementthis study removed these peer incentivesNote: we have not eliminated shame or reputation protection from process
Group versus Individual Liability14Slide15
Outcomes:No change in repaymentNo change in savingsNo change in allocation of time by credit officersHigher client retentionHigher number of new members joined
Current paper has one year results2.5 year results showing same Main Results15Slide16
Evidence of monitoring effects
Both baseline and new clients in converted centers remember less about other members’ defaultsEvidence of selection effects New clients in converted centers are less likely to predict defaults of other members correctly.
Social network
Mostly no change. Some small evidence of fewer side-loans (insurance?) in converted centers.
Less money spent on parties (but no change in prob of a party)
Auxiliary Results
16Slide17
Evidence of mechanisms of screening & monitoring.But they do not add up and lead to default!Why?Perhaps not enough time (2.5 yrs
showing the same)Perhaps simply not economically significantConclusion17Slide18
Design of New Areas (ongoing)Next Steps in Philippines
Stay
New areas
Group
Individual lending
Converted to individual liability after 1st cycle
Stay
Control
Existing groups
Converted to individual liability
18Slide19
FINCA Peru: Clients wanted trainingFreedom from Hunger and Atinchik developed materials239 village banks in Lima and Ayacucho
138 randomly assigned to receive credit with education (“treatment”)101 randomly assigned to remain as-is, receiving credit only (“control”)Study lasted ~two yearsFINCA Peru19Slide20
Impact on MFIRepayment increasedClient retention increased 10%Reciprocity? or improved business outcome?
Business processesInvested profits back in businessKeeps records from businessImplemented innovations in their businessClient outcomesIncrease in average sales up 16%Increase in worst-month sales up 28%No increase in employmentFemale children more likely to attend schoolFINCA Peru Outcomes20Slide21
Win-winMore efficient screening, arguably better decision-making (needs testing)Study impact on marginal borrowers!Implementing this in Philippines, and potentially Peru
Looking for more places: replication criticalCredit Scoring21Slide22
Is screening too rigorous?What is the impact of lending to those not being reached currently?Lender uses credit scoring + subjective decision-making by branch managerUnrejected in real-time rejected
clientsSurveyed them 6-12 months laterCredit reports collected 2 years laterProfitable for lender to lend to themSouth Africa Experiment22Slide23
Positive impact on employment, wages and hunger7 percentage point reduction in poverty levelRemember: CONSUMER lending
Impact23Slide24
Example of replicationPrior study in Philippines (SEED)Commitment to not withdraw can help increase savingsNew studies: How do we get people to deposit!Peru, India and the Philippines
Behavioral Savings24Slide25
Series of ideas from psychology & economics:Attention (reminders)Mental accounting (puzzles, photos, framing, goals)Gains versus losses
Incentives (i-rate, bonuses)Habit formation (timing of deposits/reminders)Goal: Tease out crucial mechanismsGenerate evidence from multiple settingsBehavioral Savings25Slide26
Price (interest rates)Credit bureausde Janvry, McIntosh & Sadoulet
in GuatemalaLoan terms and frequencyField and Pande in IndiaLoan sizeMexico, PhilippinesLinks to formal insurance Hospitalization, health, life, rainfall, cattle insuranceReturns to capitalMcKenzie and Woodruff in Sri Lanka and MexicoLaundry List26Slide27
Ethics and Resources (some more serious thoughts…)Are all interventions unambiguously good? If not, is it ethical to intervene in the lives of the poor without knowing your impact?
Is it ethical to spend resources on treatments that are not proven (and that could have been spent elsewhere)?Let’s not underestimate the power, and thus the responsibility, we have. We must know what works and what does not.Slide28
What proportion of a budget should one spend on monitoring and evaluation?MUST separate monitoring from evaluationDifferent denominators.Remember why you do both:Monitoring: check operational and institutional efficiencyEvaluation: know how to spend future money
Monitoring: Think % of project expensesEvaluation: Think % of future expenses28Core Impact Studies:Resource QuestionSlide29
Thank you!www.poverty-action.org
www.financialaccess.orgwww.povertyactionlab.orgdean.karlan@yale.edu29