Warwick Economics Summer School. Course: International Development Lecture 9: Microfinance. Andreas Menzel

Transcription

1 Warwick Economics Summer School Course: International Development Lecture 9: Microfinance Andreas Menzel

2 Introduction Microfinance one of the most prominent Development Tools Rose to prominence with M. Yunous Grameen Bank from Bangladesh The Banker of the Poor Established in 1976, Nobel Peace Price in 2006 Traditionally, offering banking services to the poor was not considered profitable Savings/credits too small to cover fixed costs Microfinance introduced key innovations which drove down these costs Over 200 Mio. Microfinance customers today worldwide

3 Introduction Share population without access to financial services (accounts, credit, )

4 Introduction

5 Introduction

6 Introduction Key innovations of microfinance, to drive down cost of lending to Poor Key cost factor: minimizing loan default 1. Lend only to women More honest and risk-averse lenders? More embedded in local community (also due to children) Less likely to run away 2. Give out only small loans, and reward repayment with larger subsequent loan 3. Lend to groups Each group member liable for loans of other group members. If default by one, no later loans for anyone Reduces transactions costs for loan officers Concept rooted in traditional rural self-help groups (RoSCAs) Points 1 & 3 relaxed over the years, as MF became more commercial Interest rates down to 10% pa, much less than traditional moneylenders (> 60%) (also a size effect compared to traditional moneylender)

7 Rocky Roads Microfinance was star of development policy in 90 th and 2000 th (2005 UN Year of Microfinance) Make a profit while helping the poor Aligned with pro-market, grassroot development philosophy of western governments and donors Plus focus on women! However, in recent years, image much more sobering No evidence found of microcredit reducing poverty Image tainted by commercial microfinance lenders & large scale MF default crises (often induced by politicians) Attention moved from micro-credit to micro-saving & microinsurance

8 Definition Microfinance Micro-credit Micro-saving Micro-insurance Microcredit most prominent & widespread branch But under fire recently Micro-saving growing strongly now: Next big thing? Micro-insurance still in its infancy Many MF customers see less strict separation between these branches

9 Outline of Lecture 1. Is there a need for Microfinance? 2. Micro-credit: What are the effects 3. Micro-credit: What kind of Micro-credit 4. Micro-saving 5. Micro-insurance

10 Part 1: Is there a Need for Microfinance?

11 Potential Needs for Microfinance 1. Small businesses need credit to grow, to create more jobs and income Original justification for microfinance on large scale M. Younous: Every poor is a hidden enterpreneur 2. Poor people need credit to smooth over income shocks More popular justification recently 3. Poor people cannot save because of lack of secure saving possibilities 4. The poorer people are, the more existential income risks they usually face

12 Are small businesses credit constraint? Theory that became popular in 80 th (DeSoto, ) Majority of poor are subsistence entrepreneurs, own a little business (street vendors, small scale agriculture) They could grow their business and generate more income, if only someone would give them a credit to invest in it. But nobody wants to Lenders find it too risky (borrowers often have no address or legal documents) Borrowers have no valuable collateral Loans involved too small to warrant the fixed costs of giving a credit Ineffective government and courts do not help in case of default small businesses are credit constraint, and therefore its owners stay poor!

13 Are small businesses credit constraint? Most basic theory of capital and credit constraints Cobb-Douglas Production Function Output = C α L 1 α 0 < α < 1 C = Capital, L = Labour: #Worker Each unit of capital C costs interest r

14 Are small businesses credit constraint? Theory of capital and credit constraints Imagine a small business with one worker Output Costs (capital) Endowment of capital of entrepreneur Capital C

15 Are small businesses credit constraint? Theory of capital and credit constraints Imagine a small business with one worker Output Costs (capital) If he could, he would lend this much capital Endowment of capital of entrepreneur Capital C

16 Are small businesses credit constraint? Would small entrepreneurs earn more income if you give them more capital? Even after subtracting the interest rate costs they would have to pay if they lent the capital from a bank? What would be the maximum interest rate on the capital, so that the entrepreneur is still better off? Influential Study by S. DeMel, D. MacKenzie and C. Woodruff (from Warwick!) in Sri Lanka from 2008 Pick 230 random small entrepreneurs and give them $100 or $200 to invest in their business Average capital invested by these businesses: $180 Tailors, carpenters, food shops, Give it to some in cash, or in kind (a good of choice by entrepreneur of that value for business)

17 Are small businesses credit constraint? Results: The grants in kind and in cash both had a return of 5-6% per month that is a return of more than 60% per year. The smaller grants increased profits almost as much as the large ones. Effect was mainly visible among male entrepreneurs, not females! Effect was also mainly visible among enterpreneurs with more cognitive ability (IQ) Effect was much reduced among enterpreneurs with more household wealth Direct evidence for credit contrains!

18 Are small businesses credit constraint? Summary: If enterpreneurs had lend same amount of capital than the grants, they still would have made a profit even if they had to pay interest rates of 50% pa. Average bank lending rates in Sri Lanka at that time: ca. 20% Other studies in Ghana and Mexico find equally high returns on capital for small businesses. If someone gave them a loan with less than 50% interest rate, it should make these entrepreneurs better off! does it?

19 Part 2: Micro-Credit What are the effects

20 Effects of Micro-credit How does the availability of micro-credit affect Business performance Living Standards (Consumption) Social outcomes (health, education, women empowerment) String of high profile randomized studies in various countries, with various backdrops A very well studied question in development Possibly most well-known one: The Miracle of Microfinance? conducted in Hyderabad, India (2010) From most prominent development economists these days

21 Micro-credit in Hyderabad Banerjee, Duflo, et al. (2010) Authors team up with large MFI that wants to expand into Hyderabad in 2005 MFI identifies 104 neighborhoods in which they want to open presence Authors randomly pick 52 neighborhoods in which MFI opens offices first (others: 1-2 years later)

22 Micro-credit in Hyderabad Large scale design allows studying effect of MF on whole communities BUT: How about other MF Institutions? Study did not affect others MFIs. They were still free to open up shop where they wanted. But: In neighborhoods selected for experiment, 27% of households had a micro-loan, compared to 18% in non-selected neighborhoods. Random selection => Must be caused by improved access to microfinance in selected neighborhoods! Loans were classical micro-loans, handed out to groups of 6-10 women, who were liable for each other s loans Loans ca. $200, repaid weekly over 50 weeks, with 12% interest r.

23 Micro-credit in Hyderabad About population in targeted neighbourhoods at start of study: Typical household: Family of five with ca. $500 average monthly income Universal primary schooling, 85% secondary schooling 30% of households run some business Only 7.5% of these businesses employ someone Monthly Revenue: ca $1000, profits: $325 43% of businesses female headed 40% of households have an outstanding loan from a moneylender, and 25% from family, neighbors or friends 50% of households have a sick member who had to borrow for health related reasons over the last year

24 Micro-credit in Hyderabad: Effects Three years after the MFI opened its offices in the selected neighborhoods: Compared to the non-selected other 52 neighborhoods - Households have no higher income or consumption - The share of households running a business is the same! - BUT: Households that already had a business invested more in it, and diversified activities! - Profits increased only among the 5% most profitable businesses before the study. - New-found businesses have lower profitability, and are more likely to be run by women.

25 Micro-credit in Hyderabad: Effects No effects on health or schooling expenditures No effects on female empowerment However, expenditures on durable goods (electronics, appliances, ) increased while expenditures on temptation goods (sweets, alcohol, cigarettes, parties) decreased. Effects rather sobering: Problem: Hyderabad booming city during time of study. Improvements for everyone masked effects of microfinance? Compare effects with four other studies in Marocco, Mongolia, Bosnia and South Africa

26 Micro-credit studies: Other countries Marocco: Similar project, but rural villages instead of city neighborhoods Again, no effect on consumption. But increase in investment in existing businesses Plus more time in self-employment, and less in outside work Small effect on education, none on women empowerment Bosnia: More business creation and shift from wage-employment to self-employment Again no effect on income & consumption South Africa: Slightly different program: Individual loans and higher interest rates Mainly used as emergency loans (also to pay back other loans) Positive effect on income, mainly due to prevention of job loss

27 Effects of Micro-credit: Take Away Microcredit looks far from panacea as people believed 10 years ago! Little transformative effect on life of poor as proponents believed. Not all poor are entrepreneurs which only need a little credit to succeed. A number of studies have achieved remarkable consensus on effect of access to microcredit: 1. Low take-up, typically 15-30% 2. No detectable effect on average income 3. No detectable effect on health or female empowerment 4. Positive effects in investment into existing businesses a. Including more labour input by owner b. But no additional jobs for others 5. Increased profits among largest targeted businesses 6. Helps to cover acute negative shocks to income source

28 Indebtedness and Microcredit-crisis Recently, the potential danger of microcredit to lure vulnerable people into indebtedness has been highlighted. Economics studies have very little to say on indebtedness Even without microcredit, debt and indebtedness to employers, moneylenders and family/friends is a constant factor in lives of majority of poor throughout history (80% of Hyderabad sample had some form of debt) Interest rates lower than traditional moneylenders Cruel group lending and joint liability?

29 Indebtedness and Microcredit-crisis Image of microcredit has changed, from benign NGO-style lenders in 90 th to commercial lenders today. Commercial lenders have ambitious growth and profit targets, offer credit to people already indebted with other MFIs. Often made loan conditions not clear to lenders Andra Pradesh Micro-credit crisis, 2010 State with most microloans in India Anectodal reports about microcredit induced suicides State Government wanted to push own MF scheme Toughened regulation, and tacit encouragement to not repay microloans Collapse of microcredit, slow recovery, end of Indian microcredit boom.

30 Indebtedness and Microcredit-crisis Reddy Subrahmanyam, Official in Andhra Pradesh: The industry has become no better than the loan sharks it intended to replace... But the money lender lives in the community. At least you can burn down his house. With these companies, it is loot and scoot Andra Pradesh Micro-credit crisis, 2010 State with most microloans in India Anectodal reports about microcredit induced suicides State Government wanted to push own MF scheme Toughened regulation, and tacit encouragement to not repay microloans Collapse of microcredit, slow recovery, end of Indian microcredit boom.

31 Indebtedness and Microcredit-crisis

32 Effects of Micro-credit: Take Away But why so little effects? When we saw that returns on additional capital into existing small businesses are so large Maybe credit not binding constraint? Other problems constrain business growth? Insecure property rights? Risk aversion/mental constraints Less human capital? Don t want to be entrepreneurs -> wish child gets govt. job Current microfinance arrangements not optimal? Next Part!

33 Part 3: What kind of Micro-Credit?

34 Typical Features of Microcredit Could the low effects of microcredit, especially on business creation, be caused by its typical design? Group lending Joint liability Immediate start of repayment (from first week on) Frequent payments (usually weekly) Zero-tolerance policy to late payment

35 Theory of Microcredit Which purpose do these features have? Every Lender/Bank in the world is faced with three basic problems 1. Adverse Selection How do I screen out bad types of borrowers 2. Ex-Ante Moral Hazard How do I screen out borrower with risky/bad investment ideas 3. Ex-Post Moral Hazard How do I make sure borrower puts effort into his project, so that it succeeds, and he does not default and walk away Problems of Asymmetric Information Cost of recovering loan in case of default is so high, especially in developing countries, that lender want to prevent default from the beginning

36 Theory of Microcredit If lender has asymmetric information, exploit information from people that know borrower better, e.g. people from his village Make them guarantee each other s loan. A microfinance group will only form if all member are convinced that risk of default by other group members is low Addresses all three sources of asymmetric information However, devil is in the detail Joint liability could have some perverse effects ( more in later slides)

37 Theory of Microcredit Problem of Self Control Small weekly repayment instalments are easier to repayment. Everyone has always a but of money around to pay. Bigger chunks might be trickier: Mohammed Younus (2006) It is hard to take a huge wad of bills out of one s pocket and pay the lender. There is enormous temptation from one s family to use that money to meet immediate consumption needs.... Borrowers find this incremental process easier than having to accumulate money to pay a lump sum because their lives are always under strain, always difficult

38 Theory of Microcredit Setting Expectations and Norms Strict enforcement and immediate start or repayment from week one communicates that repayment of loan must have highest priority No delay in starting project for which money was borrowed Counteracts self-control problems

39 Theory of Microcredit But could strict rules stifle potential of loans? If other group members are liable for my loan, I might chose only the least risky investment idea, with little potential It might discourage most reliable borrowers (as people want to be grouped only with equally reliable other borrowers). With repayment typically starting from week one after disbursement of loan, borrowers often report to use a share of the loan only to repay the first instalments.

40 Group Lending & Joint Liability Group Lending is core characteristic of microcredit Significantly reduces lending costs Strains communities and lets borrowers invest loans too conservatively Group Lending affects costs through many possible channels. Understanding which specifically would allow to make it more targeted and less draconian How do we get most of the green box with least of the red box?

41 Group Lending & Joint Liability Possible Effects of Group Lending on preventing default Group Lending with Joint Liability Group Lending without Joint liability Pre- Loan disbursement Post- Loan disbursement Selection of good borrowers/projets Ensure effort in implementation and non-default Concerns of loosing face in front of group if one does not repay. Make one less trustworthy in community for future cooperation If one of these is the dominant mechanism, we can relax some features of group lending & joint liability

42 Group Lending & Joint Liability Study by Karlan et al. (Science 2014) Which of these three mechanisms are important? Two experiments with microfinance groups in Philippines: 1) Switch existing microcredit groups surpisingly from Joint Liability to individual liability Clever: Distinguish if joint liability works through selection before loan disbursement, or enforcement after. 2) Offer joint- and individual liability groups in different, randomly selected villages Is fear of loosing face in front of group enough to enforce repayment?

43 Group Lending & Joint Liability Results experiment 1: Groups that were formed under joint liability but then converted to individual liability exhibited 1) no drop in defaults 2) higher retention in subsequent loan round 3) took only smaller loans in subsequent rounds 4) BUT: people with less social connections showed an increase in default after conversion Looks like joint liability works primarily through selection of borrowers, not so much later enforcement However, anecdotes about great relief when conversion to individual liability was announced.

44 Group Lending & Joint Liability Results experiment 2 Among newly formed groups, those formed under individual liability show no different default rates So is it all about peer effects of group lending ( not loosing face ), not joint liability at all? BUT: Credit officers were reported being reluctant to form individual liability groups? Conservatism of officers <- or -> Things experiment missed?

45 Group Lending & Joint Liability Unintended effects of joint liability? Imagine being in a group of five and guaranteeing the loans of all group members, and it becomes obvious that one or two of them won t repay. What are you going to do? If repaying for other s loans costs more than the returns from your loan? Declaring default as well? Jump the queue? Joint Liability can cause a default run => cause more defaults Happened in reality several times!

46 Group Lending & Joint Liability Take aways: Complex picture Scientific evidence does not show joint liability being necessary Crucial aspect seems to be Group Lending: Everyone is only liable for own credit, but pay-out of loan and collection of repayments is done in groups. Saves transaction costs for MFI, and maintains powerful peerpressure within group to service loan (default cannot be kept hidden from community) Another study showed that individual lability group borrower in India with weekly meetings has better repayment rates than those with monthly repayment rates (but self control effect?) but they also reported closer ties among them later In reality, many MFI have moved away from joint liability to individual liability, especially since Indian microfinance crisis in 2010

47 Grace Period, yes or no? Allmost all MFIs insist on loan repayment starting one week after pay-out Many borrowers use a share of loan only to repay first rates => very inefficient! Experimental study in India: Among 170 loan groups of 5 borrowers each: After formation of group, half of them is granted a two month grace period.

48 Grace Period, yes or no? Results 6% more of total loan amount was invested into business activity Likelihood loan being used to open new business doubled Three years later on, businesses of grace period loan holders have 20% higher businesses 80% more capital invested Effect most pronounced among risk-averse lenders and those with low liquidity However: Modest increase in default rates

49 Grace Period, yes or no? How did a grace period lead to these effects? 1. Allows buying larger amount of inputs for lower prices 2. Allows to experiment with new varieties of products (e.g. tailor: different kind of clothes) Why do most MFIs not offer these grace periods? - No obvious answer - Waiting 2 months for first interest payment is a cost - Evidence of higher default rates might push interest rates even higher - Some kind of vicious cycle? - Measure against self-control problems?

50 Micro-credit + Business training? Low share of microloans invested into business, especially new businesses Effect on profits only among already most profitable businesses Improved by combining micro-loan with business training (accounting, writing business plans)? Quite a few studies written on this topic Most found little to no effect Those that did found some positive effects were those with the most rudimental training (simple rules of thump )

51 Micro-credit: Wrap-up Celebrated development tool until few years ago But little evidence on transformative effect on lives of poor. Some benefits among people that were already entrepreneurial before supports them with more reliable & cheaper access to finance But few poor are natural born entrepreneurs Environment in which MFI operates limits optimal credit designs (e.g. longer grace periods) Debate moves on: Micro-Saving (Next Part!)

52 Part 4: Micro-Saving

53 Micro-saving: Realistic? Do the poor have anything to save? Banerjee & Duflo (2007): Even those who live on 1$ per day spend only 2/3 of income on food. And what they spend on food, they spend it not just to maximize calories and nutriants, but also on taste and treats 10% often spend on sugar, alcohol and tobacco Another 10% on festivals and parties People that report going hungry every now and then still own a (old) TV Rural Indians today, despite higher income, eat less calories today than 20 years ago. Even the very poor have disposable income from which they could save And they do! Rich literature on saving behaviour of poor Problem: Many don t have access to good saving devices (e.g. saving accounts)

54 Micro-saving: Obstacles Poor face at least six obstacles to effective saving 1. Regular saving accounts at banks too expensive Sums involved are too small to warrant opening costs at regular banks Storing money at home is risky 2. Rural poor live remote, far away from many banks Only 22% or world population have access to formal saving account 3. Lack of Trust Poor Institutional Environment and Regulation of banks 4. Regulation itself problematic Anti-terrorism and anti-money laundering regulation often requires official IDs and documents to open accounts Poor don t have it

55 Micro-saving: Obstacles 5. Psychological/ Behavioral Constraints Present Bias/Self control Banerjee & Mullainathan (2012): Temptations are often small (a cigarette, a sweet), relatively less problematic for richer people Much research now: more on this in Anandi s part next week 6. Social Constraints Many poor embedded in rich kin/friendship networks with implicit duty to help each other Due to lack of formal insurance M. Younus: Live is always under strain If you have some money, always pressure to share Focus of much recent research

56 Traditional Microsaving There was always big demand and ingenious solutions to these obstacles, especially Nbr 5 & 6 1. RoSCAs (Rotating Saving and Credit Associations) Group of people (mostly women) meets regularly Each meeting each member gives a fixed amount of money to one member. Member who receives money rotates from meeting to meeting For giving a small amount every week or so, one receives a large amount at regular times Helps against presence bias and demands from others (especially husbands, apparently )

57 Traditional Microsaving 2. Taking Credit against Savings Intriguing Research by Rutherford (2001) and others (mainly in Africa) Even though people have savings available at larger credit and saving associations, they still take credit from a lender Credit signals lack of any money to share to others But comes at a big cost, paid interest rates instead of earned ones Many poor will accept negative nominal interest rates for a safe (and anonymous) place to save money

58 Effects of Microsaving Formal Microfinance has recently expanded into micro-saving Especially as evidence mounted that micro-credit not as effective as hoped Plus reports on micro-borrowers driven into indebtedness Many MFIs now demand that a micro-loan borrowers open a micro-saving account and pay some money in on top of each repay installment Ca. 80 Mio micro-savers by now But is it effective?

59 Effects of Microsaving Two main randomized evaluation of effects of micro-saving 1) Dupas and Robinson (2013) 160 self employed market vendors and bicycle taxi driver in Western Kenya offered free saving account (usual opening fee, $6.40) 87% accepted offer 41% becoming active users (more than two deposits in first six months) Among market vendors, access to saving account increased investment into business by 40-50% after half a year No effect among bicycle taxi drivers Interesting: Vendors mostly female, taxi drivers mostly men

60 Effects of Microsaving Two main randomized evaluation of effects of micro-saving 2) Prina, (2013) 560 randomly selected female household heads in Nepal were offered saving account without any costs Interest rate 6% (but inflation > 10%) 84% took up offer 80% became active users (more than two deposits in first six months) After one year, women offered account had 25% more monetary assets 12% overall more assets No evidence of crowding out of other assets (lifestock, durable goods) 20% more expenditures on education (35% among those with schoolage children) 15-20% more expenditures on meat & fish Among households hit by health shock, smaller reduction of weekly income levels

61 Effects of Micro-saving Taking Stock: Large effect on business investment and social investments (education, better food) Mainly among women Many follow-up evaluations of micro-saving in different countries under way now Important emerging variant of micro-saving: Commitment saving devices To beat lack of self control!

62 Theory: Time-consistent preferences Humans discount the utility of receiving a good in the next day/month/year by a factor δ < 1. If I get a dollar tomorrow, it gives me utility now of δ x $1 If I get a dollar in two days, it gives me tomorrow a utility of δ x $1. So today it gives me a utility of δ x δ x $1 = δ 2 x $1 (< δ x $1) Valuation 1 Dollar Today Days

63 Theory: Time-inconsistent preferences Now consider time-inconsistent preferences I will start saving tomorrow, because tomorrow I will value the saving more than today But come tomorrow, tomorrow will be the new today, and I will procrastinate by another day! Valuation 1 Dollar Today Days

64 Time inconsistent preferences Time inconsistent (or present-biased) preferences simple way of capturing phenomenon of procrastination sophisticated present-biased people Are aware of this problem, but find it hard to resist temptations today Would be happy if they could get a commitment device something that binds their hands against spending money on temptations today E.g. saving account which only lets you withdraw savings after a certain day, or after a certain amount has been reached Studied in Philippines and in Kenya

65 Commitment saving accounts Project joint with Green Bank in Phillpinnes Offer randomly chose customers with saving accounts an additional commitment saving account If accepting, customers could chose between an account from which they could only withdraw savings once a pre-specified date was reached (e.g. the date of a wedding) sum was saved up to (e.g. for improvement of roof) Interest rate did not vary from ordinary savings account No compensation for self-imposed lack of liquidity

66 Commitment saving accounts 28% of people offered account accepted it After one year, those offered had 82% more savings But those 28% who actually used it had almost four times as much savings! Effects on saving strongest among those in which a simple psychological test identified the strongest timeinconsistency After three years, those offered had still 33% more savings (though difference is not significant anymore) Saved sums might have already been withdrawn and invested

67 Commitment saving accounts Dupas and Robinson (2013) Offered randomly selected RoSCA participants saving boxes, clearly labelled as health expenditures. Some of them had the key with a saving bank officer, and with the other the key was given to the saver. Found similar saving rates, regardless of whether key was with saver or officer Is it all about mental accounting, not physically locking away savings? But a saving box which to open requires calling in a saving bank officer might not be most attractive device

68 The Equivalence of Micro-credit & -saving We tend to think of credit and saving as two distinct, even opposite kind of activities However, among many poor, this distinction is much less clear Micro-credit & saving often seen as two sides of the same coin There is a high demand simply for products which force you to make small contributions regularly, and give large pay-outs at one point of time Helps as commitment device Helps as excuse to keep demands from acquaintances away Pay-out can be before or after payment of instalments, and need not be larger than sum of payments (forced payments considered valuable service worth paying for)

69 The Equivalence of Micro-credit & -saving Study in Pakistan Members of a state-run saving association were repeatedly offered, each week, the following offer, running over 6 days: Day 1 Day 2 Day 3 Day 4 Day 5 Day 6 Pay Receive Day of payout randomly selected, and payout randomized between 900, 1000, & 1100 Rs Character or product randomly varied between credit and saving product

70 The Equivalence of Micro-credit & -saving Results Demand for product was not affected by the day at which pay-out was done It was also not strongly affected by the interest rate (pay-out of 900 vs vs Rs) People were offered random variations of the product repeatedly. Those with high demand for product took it up regardless at which day pay-out was Interest rate of minor importance Safe, strict saving device in demand

71 Part 5: Micro-Insurance

72 Micro-insurance Poor people are exposed to a high degree of uncertainty Many insurances we take for granted do not exist for them Little to no health insurance Most self-employed or working as day-labourer Income uncertain Public services bad e.g. bad housing Not just life inherently risky People also report high levels of depression and anxieties Lack of focus in income generation => Lower income Implicit obligations to help each other => Lower capital accumulation Adoption of riskless but low yield agricultural production

73 Micro insurance Micro-insurance possibly addressing huge demand But micro-insurance more complex than micro-credit or saving How to establish loss from a bad harvest? How to separate bad weather shocks is from low effort (insurance fraud)? Successfully tested product: Rainfall-indexed insurance If rain stays below or above certain thresholds, insurance pays off No moral hazard or adverse selection problem Tested in several studies in Africa, India, Today ca. 10 mio holders of rain indexed insurance

74 Micro insurance Rain-insurance study in Ghana Farmers were offered an insurance that Pays $145 per acre in case of 8 successive days with rain or 10 successive days without rain Offered at randomly selected different prices repeatedly over three years Market price: $18 per acre Actuarially fair price: $12 per acre (Highly) subsidized prices: $5 or $1.30 For Free For comparison, another group of farmers got a $85 cash grant per acre

75 Micro-insurance Demand for micro insurance 12-15% take-up at market price 40% at actuarially fair price 60% at subsidized prize (5$) Very similar rates compared to another study done in India Effects of micro insurance Free insurance: Higher effect on farm investment than $85 cash grant Uninsured risk seems to be binding constraint for agricultural investment When offered free insurance, farmers find resources for additional investments Credit constraints not as binding as traditionally thought

76 Micro-insurance Significant trust aspects from buying insurance Recency bias: When farmer got a pay-out in previous period, likelihood of buying insurance increased When farmer did not get a pay-out (no anomalies in rain), likelihood decreased Likelihood also increased if another farmer in the social network got a pay-out Problem: Low-probability, high impact events, where insurance might work best, demand will be lowest -> market failures

77 Micro-insurance Chinese Study on Sow-raising Many rural Chinese households raise pigs for a sideincome Female piglets can either be spaied Grow quicker and fatter, tastier meat or not spaied grow slower, eat more, meat less tasty bear piglets which can be raised more profitable, but also more risky (sow can die from disease) Chinese agency offers sow-disease insurance Study: In random villages, insurance agent received performance incentives => Sold more insurances

78 Micro-insurance Sow insurance: Results Significant increase in sow-raising Effect not due to crowding out of raising of other lifestock Again: Evidence of lack of trust in insurance policy that have to be overcome. Natural Experiment in study One year after start, an extreme winter storm hit study area of subtropical southwestern China Many pigs died After insurance paid out in these cases, take-up of insurance increased in hard-hit areas People needed to see with own eyes that insurance agency can be trusted

79 Micro-insurance: Take-away Un-insured production risks potentially strongly binding constraint on poor self-employed However, fraud-proof insurance policies remain difficult to design Plus micro-insurance agencies face issues of trust (too many state agencies failed their customers in the past) Demand at market priced rates also remains low (10-15% of target market) People do already take part in risk-mitigating mechanisms Networks of friends and families Loans from moneylenders sometimes implicitly linked to ability to repay Cause for high interest rates

80 Summary: Microfinance Once heralded as the ultimate, sustainable development tool Today, with high quality scientific evidence, picture much more nuanced Focus has moved towards micro-saving and insurance Financial inclusion nevertheless an important aspect in its own right In the end, microfinance, an innovation stemming from developing countries themselves, should be seen as a successful adaptation of financial systems to the needs of developing countries.