Risk Management for the Poor and Vulnerable

Transcription

1 CSIS WORKING PAPER SERIES WPE 093 Risk Management for te Poor and Vulnerable Ari A. Perdana May 2005 Economics Working Paper Series ttp:// Te CSIS Working Paper Series is a means by wic members of te Centre for Strategic and International Studies (CSIS) researc community can quickly disseminate teir researc findings and encourage excanges of ideas. Te autor(s) welcome comments on te present form of tis Working Paper. Te views expressed ere are tose of te autor(s) and sould not be attributed to CSIS Jakarta Centre for Strategic and International Studies, Jakarta

2 Risk Management for te Poor and Vulnerable Ari A. Perdana CSIS Working Paper Series WPE 093 May 2005 ABSTRACT Tis paper reviews some literatures on te mecanisms available for te poor in managing risk. Lacking access to formal mecanisms of risk management, te poor rely on informal mecanisms, wic are built based on te existing social networks and trust. But wen te socks are big or affecting te entire community, tese informal mecanisms may not be adequate. Some policy interventions are ten required to elp improving te ability of poor people in managing risk. Policy intervention sould aim to provide access for te poor on saving, credit and insurance. Microfinance scemes ave been applauded as a successful best practice in providing access to saving and credit. However, microfinance institutions still ave some room for improvement by expanding teir role in providing insurance scemes. Keywords: poverty, vulnerability, risk management, microfinance. Ari A. Perdana ari_perdana@csis.or.id Ari_Perdana@ksg06.arvard.edu Department of Economics CSIS Jakarta, INDONESIA

3 Risk Management for te Poor and Vulnerable Ari A. Perdana * CSIS, Jakarta Introduction Te existence of risk significantly affects people s life. Risk creates uncertainty, wic in turn influence people in making decision. Risk also makes individuals face some probability to experience income socks. An income sock could make some people s income fall below te poverty line. In oter words, risk makes some individuals vulnerable to poverty. Tis paper presents a literature review on te significance of risk to te lowincome individuals, and te mecanisms available to manage it. Te discussion is outlined as follows. Section one reviews te basic concepts on poverty, vulnerability and risk. Section two discusses te informal mecanisms available for te poor to manage risk. Section tree presents te policy options to elp te poor in managing risk as te informal mecanisms are often inadequate. Finally, section four concludes te discussion. Concepts: poverty, vulnerability and risk Poverty is traditionally defined as te inability of an individual or a family to command sufficient resources to satisfy basic needs (Fields 1994:88). Basically, it is a condition in wic a person s income or consumption in a certain time falls below a * I would like to tank Micael Woolcock for useful comments on tis paper. 1

4 certain tresold, wic is referred to as te poverty line. However, poverty cannot be regarded as only a static concept tat deals wit one s welfare condition at a certain point in time. In reality, poverty is also a dynamic concept, since ouseolds frequently move in and out of poverty overtime. Tis raises an issue of vulnerability to poverty. Vulnerability to poverty can be defined as te probability tat a ouseold will become poor in te near future. Since te concept deals wit probability, we can say tat tere is always a cance tat a currently non-poor may end up being poor in te near future. Non-poor ouseolds may fall into poverty due to events suc as natural socks disasters, economic sock or crisis, security problems and many oters. Conversely, people wo are currently poor also ave a cance to escape from poverty. Te improvement in economic situation may bring more job opportunities, providing sources of income to te people, ence enables poor people to climb up from poverty. Tere are several approaces and metodologies ave been applied in te studies on te dynamics and vulnerability of poverty. Te simplest approac is tabulating te frequency of te event wen a ouseold falls into poverty over some fixed time frame. Alternatively, Bane and Ellwood (1983) 1 carried te metodology based on te notion of spells of poverty, using exit probabilities to examine te lengt of time tat people spend in poverty. Tey argued tat many ouseolds climb in and out of poverty over a given period. Tey also igligted te fact tat altoug many only ave sort spells in poverty, most of tem wo are poor at a given point in time would ave very long spells of poverty before tey escape. 1 See also Stevens (1995), wose study was based on tese autors. 2

5 Anoter approac is measuring te variability of income or consumption. A person wit ig variance of income is more likely to fall into poverty. 2 Suc approac was carried by Cauduri (2000), and later applied by Pritcett et al. (2000) and Suryaadi and Sumarto (2001) using te data from Indonesia. 3 Te studies expand te definition of poverty by combining information on current consumption wit te probability tat te future consumption levels will fall below te poverty line. In teir study, Suryaadi and Sumarto sow te significance of te vulnerability to poverty. Te combination of macroeconomic sock and political instability during te Indonesian economic crisis in te late 1990s as doubled te number of poor ouseold almost doubled from 1996 (pre-crisis data) to 1999 (post-crisis). Tis was equivalent to about 27 million additional poor during te period. However, number ouseolds wo are statistically not poor but facing relatively ig probability of falling below poverty line ave increased from 13 million to 38 million; illustrating ow te crisis as put near-poor Indonesian ouseolds at risk of falling into poverty by tree times. Vulnerability to poverty is closely related to te concept of risk. Risk refers to uncertain events tat can damage well-being (World Bank 2000:139). Tere are many sources of risk. Te nature and environment are sources of risk, as well as ealt, economic and socio-political condition. Some types of risk affect just te individual or a ouseold (idiosyncratic), for example accident, sickness or crime. Oter types of risk affect a wider range of people (covariant). Small natural disaster, epidemic disease, riots or bad weater may affect te entire community witin a village. Big disaster, war, 2 By definition e or se will also be more likely to escape from poverty. Neverteless, in most cases te likeliood to escape from poverty is lower tan to fall into it. 3 See Appendix for te model. 3

6 economic crisis or regime cange will affect te entire nations or even create a contagious effect to te neigboring countries. A person wit ig risk exposure is expected to ave greater income variance; ence e or se will be more vulnerable to poverty. But tere is also a reverse effect poverty brings more exposure to risk. Most of poor people live in unealty and unsafe environments, wic expose tem to a greater risk of ealt and security. A study by Jalan and Ravallion (1996) using data from te rural Cina confirms tat consumption level of te bottom decile of population is te least protected against income fluctuations. Morduc (1994:221) furtermore caracterize tree factors tat contribute to greater vulnerability to poverty in low-income countries: 1. Since most of te poor in low-income countries are in te agricultural sector, weater and price variability are responsible for a large part of income fluctuations and, tus, poverty. 2. Poorly developed financial institutions, wic accounts for te lack of access to protection against risk suc as credit, savings or insurance. 3. Weak social insurance institutions. How te poor manage risk? Weter tey are wealty or poor, people can not avoid risk. But people can manage risk. Risk management can be classified into risk mitigation (ex-ante) and risk coping (ex-post).te main idea of risk management is to deal wit te fluctuation of 4

7 income and consumption troug income diversification, insurance or savings-borrowing scemes. 4 Te main difference between te ric and poor people in managing risk is te latter ave more limited access to formal mecanisms of risk management. Formal risk management mecanisms include tose tat are available in te market or publicly provided. Private insurance, bank credits and pension funds are example of market-based formal mecanisms. Examples of publicly provided mecanisms are public ealt care or social security systems. Te lack of access to tese formal mecanisms makes poor people rely mostly on informal mecanisms of risk management. Individual risk management Poor individuals or ouseolds mitigate te effects of income socks by diversifying teir sources of income. In rural agricultural areas, were most of te world s poor people live, farmers often diversify teir crops and use multiple seed varieties. In some cases tey also diversify teir occupations. For example, in addition to work in farms, rural agricultural people tend to work as part-time workers in te nearby towns. If a sock occurs, te poor cope wit falling income by adjusting teir expenditure. Usually, adjusting expenditure means spending less for non-basic needs, eating less or making dietary canges, like consuming less meat and oter side dises. Anoter way of coping is work for longer ours. If tere is not too many alternatives to 4 Anoter purpose of risk management is risk reduction. Building dams reduces te risk of flooding. Public immunization reduces te risk of epidemic disease. But not all risk can be significantly reduced. At least witout certain mecanisms (Jütting 2005, World Bank 2000). 5

8 work more in teir village, tey usually move to te oter places seasonally or temporary. Income socks also create te needs for te oter family members wo were previously not working to searc for a job. Tis in turn creates te increased number of cild labor or scool drop-out rate during economic socks. Even toug income fluctuates, people would try to smoot consumption level over time. People smoot teir consumption by saving some of te current income instead of consuming everyting tey earn today. During bad times, tey eat up some or all of wat tey ave saved. Or, tey may borrow from someone else and repay it back later in te future. Wealtier people ave access to te saving, credit and insurance do teir consumption smooting troug financial market. Tey save some of teir wealt into bank deposits or oter financial assets. Ten during periods of low income tey use teir savings or ask for credits. However, poor people generally lack access to formal financial system. As te result, tey can not save in modern financial assets or instruments. Instead, tey to save teir wealt in terms of land, cattle or durable goods and oter valuables assets like jewelries. 5 Tey will sell teir assets or consume teir cattle wen bad time comes. Community or group risk management Te above illustrations are examples of ow poor people manage risk individually. But in most cases, poor people rely on te oters in te group or community in managing teir risk. In te most common case, poor people borrow from teir 5 Rozensweig and Wolpin (1993) presents a study on te importance of bullocks as an important part of portfolio assets for rural farmers in India. 6

9 extended family members, distant relatives or neigbors wen facing financial ardsips. Apart from tat, tere are also social institutions tat serve as community-based risk management arrangements. 6 Table 1 lists examples of informal mecanisms of risk management wic are common in developing societies. Table 1. Informal Mecanisms for managing risk Individual and ouseolds Group-based Mitigating risk Diversification Crop and plot diversification Income source diversification Investment in pysical and uman capital Occupational associations Rotating savings and credit associations Insurance Marriage and extended family Sarecropper tenancy Buffer stocks Investment in social capital (networks, associations, rituals, reciprocal gift giving) Coping wit socks Sale of assets Loans from money-lenders Cild labor Transfers from networks or mutual support Reduced food consumption Seasonal or temporary migration Source: Te World Bank, World Development Report 2000/2001: Attacking Poverty One example of suc arrangement is agricultural contracting like sarecropping or land tenancy. In many agricultural societies, sarecropping or land tenancy is a means of saring risk between landlords and tenants. Contracting is also popular among 6 In almost all traditional societies, rituals, reciprocal gift giving and oter types of communal activities are part of social capital investment, wic purpose is to build te social institution for risk management. 7

10 agricultural communities because it is often interlinked wit credit provision landlords are also suppliers of credit to sare tenants. It is a mutual arrangement, because wen tenants receive credits, tey can be wealtier and more productive. Tis means iger income for te landlords as well. Anoter form of interlinking in agricultural contracts is marketing by landlords. Landlords are te exclusive buyers of tenants output on a given price. Usually, but not always, te price is set lower tan te market price. Te landlords would ten sell te output in te market. Tis sceme benefits bot tenants ave te certainty and landlords get te profit margin. 7 Anoter example of te institution wic serves as a way to manage risk is marriage. In many societies, marriage is a contract between two families. Te consequence of marriage is an extension of family networks. Tis implies tat a family now as larger members to wom one can seek for income transfers during bad times. A study by Rozensweig and Stark (1989) analyzed ow marriage can serve as a tool for risk edging. Tey found tat many men in rural agricultural ouseolds in India tend to marry women from outside of teir villages (patrilocal exogamy). A ouseold in a village exports daugters and imports daugters-in-law. Te reason beind tis pattern is wen te ome village experiences an income sock, for example due to bad arvest, tey still ave a family member living in a different village, wic possibly does not experience te same sock, wo can be te source of income transfer. 7 A number of literatures provide discussions on sarecropping and land tenancy. Otsuka, Cuma and Hayami (1992) presented a teoretical literature about te efficiency of agricultural contracting. An earlier study by Saban (1987) presented some empirical evidences about several types of contracting using detailed data from eigt Indian villages. Ten, Foster and Rozensweig (1994) discussed te issue of moral azard in agricultural contracting. 8

11 Te autor confirms tis by sowing te significance of te amount of intervillage in te region. Tat is wy, according to te study, marriages usually involve a man and a woman from two villages tat are relatively far away. By aving a family member tat lives in a relatively far village, te cance of tose two villages experiencing a same sock will be muc smaller. Also, two daugters from te same family will not marry partners from te same village. Limitations and policy implications Informal mecanisms are important tools for te poor in managing risk. Tey can provide protection from risk in some ways. However, informal mecanisms ave many limitations. Te may protect te poor in small income socks, but not in big or persistence socks. Since poor people by definition own only a few amount of assets, te ability to maintain te level of consumption is close to noting if te income sock is substantially large. Furtermore, if a sock persists for more tan one period, assets ave been run off in te first period so te poor are left wit no protection against income fluctuation in te latter periods. 8 Anoter limitation of informal mecanisms is tat tey often imply a trade-off between risk mitigation and efficient production. Mitigating risk troug crop diversification typically lowers te expected profits (Morduc 1995). Farmers will grow some crops wic are less risky but yield lower profits. Expected profits are also lower because poor farmers tend to be reluctant in adopting new tecnologies and taking 8 Morduc (1994) raised tis argument, quoting te case of investment in analyzed mentioned by Rozensweig and Wolpin (1993). 9

12 advantage of new economic opportunities, as sown in a study by Rozensweig (1993) in Indian rural agricultural villages during te green revolution. Community-based informal mecanisms are also limited. Tey are ineffective to protect against covariate socks suc as big natural disaster or economic crisis tat affect everybody in te community. Moreover, since informal mecanisms are based on local rules wic are sometimes unwritten, enforcement sometimes becomes te problem. 9 Tese limitations give te room for interventions. Tere is a wide range of possible policy interventions. Policies can improve risk management indirectly. Prudent monetary and fiscal policy elps controlling inflation, wic reduces te risk of falling real income. Education policy improves uman capital, wic ten brings poor people to a wider range of job options. Public ealt quality improvement elps reducing te risk of illness. But policies sould also directly improve te ability of te poor to manage risk. Tree main areas of policy intervention in elping poor people managing risk are providing mecanisms for saving, access to credit and insurance. Te roles of saving are serving as te means for accumulation of asset and precautionary purpose. Asset accumulation is important for an individual or ouseold to provide income security, for example during te old age or to finance cildren s education. Precautionary motive is driven by te needs to smoot consumption wen a sock occurs. Zeller and Sarma (2000:160) argued tat te saving as a precautionary purpose is more important wen te agent is poorer and more vulnerable, since poor, income insecure people would want easy and quick access to teir money at all times. 9 Te common problem of enforcement is members of te community end up playing strategic beavior. As Bloc, Genicot and Ray (2005) argued in teir study tat lack of commitment is an example of a source of instability in social networks. 10

13 Te importance of credit is quite obvious. Production credit relaxes te liquidity constraint tat prevents te poor to be productive. Consumption credit serves as an external source of consumption smooting wen tere is an income sock. Meanwile, te presence of insurance maintains a degree of income certainty over a period of time. Wen access to insurance exists, agents do not need to sacrifice more productive and profitable economic activities for te more certain income. 10 Since access to te markets of saving, credit and insurance are generally not available for te poor, tey rely on te informal mecanisms of risk management, wic are built on te community s social capital. Two important components of social capital are networks and trust. Te above examples of community credit association, contracting and marriage reflect te importance of social networks and trust for te poor people in providing protection against risk. Networks provide te basis of building te institutions, and trust is wat makes te institutions sustained. A good and effective intervention sould be built on te existing social capital. It sould neiter eliminate nor displace it. Microfinance is peraps te most applauded example of an intervention tat is built on existing social capital. Te basic idea of microfinance is providing small-scale loans to people wo can not borrow from te banks because tey do not ave assets to serve as collateral. 11 As te solution to te collateral constraint, instead of granting loans to a single individual, te sceme provides group lending. Referring to te Grameen Bank practice, in eac village groups of 3-5 self-selected members are formed. Loans 10 Te types of insurance needed vary among different population. Poor rural farmers will need weater insurance to guarantee a certain level of income during bad weater. Urban poor working in factories or industrial complexes will need unemployment insurance. And in general, ealt insurance is needed by everyone. 11 Pioneered by Grameen Bank in Banglades, founded by Dr. Moammad Yunus, microfinance is now a widely applied sceme of community savings and lending. 11

14 will be made to individuals, but if any one member fails to repay, all members in te group will receive collective sanctions. Te sanctions ranged from an obligation to pay te collateral or exclusion from next lending. In addition to provide loans, many microfinance scemes ave also succeeded in inducing poor people to save because in te latter stages of loans, furter collateral requirements could be substituted by requiring members to form collective funds after te loan as been disbursed. However, microfinance institutions (MFI) in some countries ave also faces several problems, like te ineffective use of loans. Instead of using te loans for productive activities, borrowers use tem loan for oter purposes suc as to repay existing debts (if borrowers ave been in a vicious circle of ig debt) or serve as lending to a tird party. 12 Anoter kind of problem is te profitability of an existing MFI may attract competitors (Morduc 1999:1592). Competition may reduce te power of existing institution in giving sanctions to borrowers, terefore increasing te number of unpaid loans, ence treatening te sustainability of MFI. Unlike te relative success in providing access to saving and credit, MFI still overlook te potential to provide insurance to te poor. Understandably, insurance service is more difficult to provide due to te problems of asymmetric information and moral azard. Tis creates some room for MFI to improve teir role in elping te poor by providing insurance scemes to te poor. Some MFI, owever, ave innovated teir businesses by linking saving, credit and insurance, as Zeller and Sarma (2000:163-4) noted: 12 Coleman s study (1999) as provided empirical examples of tese situations in Tailand. 12

15 For example, te Association for Social Advancement in Banglades and Bank Rakyat Indonesia require borrowers to buy life insurance for safeguarding loan repayment in case of te deat of te borrower. Te Indian self-elp organization SEWA allows pregnant borrowers to rescedule teir loans, and te members of te Caisses Villagoises, for example in Mali or Te Gambia, allow te provision of consumption loans at lower interest rate if financed troug internal savings of te members. Concluding remarks Poverty alleviation strategy sould not focus only on reducing te eadcount poverty rate, or te percentage of poor ouseolds to total population. Declining eadcount poverty rate may be a sort-term objective. But in te longer-run, policy objective sould be to reduce te level of vulnerability to poverty by providing mecanisms for te poor to manage risk. Policy intervention sould aim to provide access for te poor on saving, credit and insurance. MFI scemes ave been applauded as a successful best practice of intervention. Not only MFI elp poor people in providing risk management for te poor; it as been built on not substituting te existing social networks. Wile most MFI focus on providing credit and saving, tere is still some room MFI s to improve teir role by providing insurance scemes. 13

16 References Bane, Mary Jo and David T. Ellwood (1983). Slipping Into and Out of Poverty: Te Dynamics of Spells. National Bureau of Economic Researc Working Paper No (November). Beslet, Timoty (1995). Nonmarket Institutions for Credit and Risk Saring in Low-Income Countries. Journal of Economic Perspectives, 9(3), pp Bloc, Francis, Garance Genicot and Debraj Ray (2005). Informal Insurance in Social Networks. Mimeo. Coleman, Brett E. (1999). Te Impact of Group Lending in Norteast Tailand. Journal of Development Economics, 60, pp Fields, Gary S. (1993). Poverty and Income Distribution: Data for Measuring Poverty and Inequality Canges in te Developing Countries. Journal of Development Economics, 44. Foster, Andrew and Mark Rozensweig (1994). A Test for Moral Hazard in te Labor Market: Contractual Arrangements, Effort and Healt. Review of Economics and Statistics, 762(2), pp Jalan, Jyotsna and Martin Ravallion (1999). Are te Poor Less Well Insured? Evidence on Vulnerability to Income Risk in Rural Cina. Journal of Development Economics, 58, pp Jütting, Joannes (2005). Healt Insurance for te Poor in Developing Countries. Hampsire, UK and Burlington, VT: Asgate. Morduc, Jonatan (1999). Te Microfinance Promise. Journal of Economic Literature, XXXVII, pp (1995). Income Smooting and Consumption Smooting. Journal of Economic Perspectives, 9(3), pp (1994). Poverty and Vulnerability. American Economic Review, Vol. 84(2), pp Otsuka, Keijiro, Hiroki Cuma and Yujiro Hayami (1992). Land and Labor Contracts in Agrarian Economies: Teories and Facts. Journal of Economic Literature, XXX, pp Pritcett, Lant, Asep Suryaadi and Sudarno Sumarto (2000). Quantifying Vulnerability to Poverty: A Proposed Measure, wit Application to Indonesia. SMERU Working Paper, May. 14

17 Ravallion, Martin (1988). Expected Poverty Under Risk-Induced Welfare Variability. Economic Journal, 98(393), pp Rozensweig, Mark (1993). Wealt, Weater Risk and te Composition and Profitability of Agricultural Investments. Economic Journal, 103(416), pp and Kennet Wolpin (1993). Credit Market Constraints, Consumption Smooting, and te Accumulation of Durable Assets in Low-Income Countries: Investments in Bullocks in India. Journal of Political Economy, 101(2), pp and Oded Stark (1989). Consumption Smooting, Migration and Marriage: Evidence from Rural India. Journal of Political Economy, 97(4), pp Saban, Radwan A. (1987). Testing Between Competing Models of Sarecropping. Journal of Political Economy, 95(5), pp Stevens, Ann Huff (1995). Climbing Out of Poverty, Falling Back in: Measuring te Persistence of Poverty over Multiple Spells. National Bureau of Economic Researc Working Paper No (December). Suryaadi, Asep and Sudarno Sumarto (2001). Te Cronic Poor, Te Transient Poor, and te Vulnerable in Indonesia Before and After te Crisis. SMERU Working Paper, May. Townsend, Robert M. (1995). Consumption Insurance: An Evaluation of Risk-Bearing Systems in Low-Income Economies. Journal of Economic Perspectives, 9(3), pp World Bank (2001). World Development Report 2000/2001: Attacking Poverty. Wasington, D.C. Zeller, Manfred and Manoar Sarma (2000). Many Borrow, More Save and All Insure: Implications for Food and Micro-finance Policy. Food Policy, 25, pp

18 APPENDIX Suryaadi and Sumarto (2001) and Pritcett, Suryaadi and Sumarto (2000) proposed a metodology to measure te vulnerability to poverty using te estimate of te ouseold s variance of consumption expenditures. Consider te ouseold s consumption expenditure function as follows: (1) ln c = X β + ε were c is per capita consumption expenditure, X is a set of ouseold caracteristics, β is a vector of parameters and ε is stocastic errors. If a certain poverty line is set as c, ten a ouseold is poor if te consumption level is below te poverty line ( c < c ). A non-poor ouseold may be considered as vulnerable if it as a certain level of probability (for example 0.5) to become poor in te near future. Te probability of a ouseold to become poor is defined as: ln c X β (2) v = Pr(ln c < lnc X ) = φ ˆ σ were v denotes te vulnerability to poverty, tat is te probability tat te per capita consumption level ( c ) will be lower tan te poverty line ( c ) wit a given ouseold caracteristics ( X ). Meanwile, φ () represents te cumulative density of te standard normal distribution, and σ is te standard error of equation. Using te above calculation to define eac ouseold s status of poverty and vulnerability, we can group ouseolds into several categories: 16

19 Poor ouseolds: tose wose current consumption below te poverty line ( c < c ). Poor ouseolds can be divided into two sub-categories. Cronic poor are tose wose vulnerability is ig ( v > 0.5), and most likely to remain poor in te future. Transient poor are tose wose vulnerability are low ( v < 0.5), and ave a considerable probability to become non-poor. Non-poor ouseolds: tose wose current consumption is above te poverty line ( c > c ). Te non-poor can also be sub-categorized as ig vulnerability or non-poor wit ig probability to become poor ( v > 0.5), and low vulnerability, wo are basically non-poor wit low probability to become poor ( v < 0.5). Using te information of te ouseold categories, te next step of te researc is to evaluate te determinants of vulnerability by estimating te vulnerability function: (3) vˆ = X γ + τ were v is te estimated ouseold s level of vulnerability obtained from (2), X is a vector of ouseold caracteristics including personal (age, education level, gender), ouseold (ouseold size, marital status), geograpical (province, urban/rural), sectoral etc., γ is a vector of parameters and τ is stocastic errors. 17