Social Funds 46307 Innovations Notes Microinsurance: Extending Pro-Poor Risk Management through the Social Fund Platform by Marc Maleika and annet. kuriakose M icroinsurance can be an effective complement to exist- to find in low-income countries. Social Funds can help bridge ing menus of social protection programs.A flexible and this gap, standing as a platform to organize and deliver micro- powerful instrument, microinsurance (MI) reduces vulnerability insurance products.This Social Funds Innovations Note intro- and mitigates the negative effects of external shocks on poor duces some of the primary design principles behind micro- households. However, microinsurance programs require well- insurance program development, highlighting cases of best developed institutional arrangements in order to run in an practice, and suggests how Social Funds can be used to deliver efficient and effective manner. Such conditions can be difficult microinsurance services more effectively to poor households. Why Microinsurance? household Formal insurance instruments can offer superior risk Vulnerabilityand risk considerations management alternatives, provided poor households can access these services.without insurance, low-income low-income households are particularly vulnerable to households forego higher-return livelihood strategies for risk and negative external shocks (e.g., natural disaster; lower-risk avenues that reduce risk. Insurance products illness/ death of main breadwinner) due to their low assume such risk thus reducing household efficiency asset bases. In the absence of functioning insurance mar- losses and protecting assets so that the poor can escape kets, poor people in developing countries have created poverty traps. Insurance instruments pool the risks of a number of formal and informal instruments to man- individuals of a similar risk class, and transfer it to a larger age such risk.these include risk-pooling schemes (e.g., and more diverse group of market participants through funeral and burial societies); income support (e.g., credit the `hedging' process.traditional forms of insurance, how- arrangements; transfers), and consumption smoothing ever, have often been beyond the reach of poor persons. arrangements (e.g., savings; grain banks) (see bhattamishra Innovations in microinsurance aim to increase outreach and barrett 2008). however, such informal and formal and coverage across lower income tiers. approaches offer limited protection, low returns for households, and are prone to breakdown during emer- What is Microinsurance? gencies. Community-based risk management schemes also rely heavily on personal relations between participants, Microinsurance is "the protection of low-income people limiting scalability and geographic spread. even formal sup- against specific perils in exchange for regular premium pay- port programs such as food-for-work may be exclusion- ments, proportionate to the likelihood and cost of the risk ary, as in the case of female-headed households often left involved" (CGaP 2003). In contrast to savings or transfers, out of such work programs as they face difficulty making microinsurance is not limited in outreach or coverage. It the required labor contribution. can be provided by a range of different providers. Products t h e wo r l d b a n k o C to b e r 2 0 0 8 Vo l U M e 5 n o. 2 and useful in mitigating small loss events that occur box 1:benefits of Microinsurance frequently and predictably. In contrast, only risks result- Microinsurance is a powerful tool for: ing in exceptional losses are considered insurable (levin 2007). risk pooling allows for broader coverage against · Protecting the poor and their assets from nega- risk than individual households can provide on their own tive external shocks (particularly for covariate risks that frequently lead to breakdown of community-based risk management when · Compensating the effects of covariate shocks all households face financial strains simultaneously). (e.g., natural disasters) · addressing gender-specific vulnerabilities need for iteratiVe Product deVeloPMent · Freeing up household capital for investment in small enterprise the success of a microinsurance (MI) program depends on the trust clients have in the insurance product and · helping households avoid poverty traps the MI institution servicing it. Product outreach and suc- · expanding informal insurance schemes and cessful uptake by clients requires simple explanations social protection of risk-pooling and insurance services more generally. Similarly, transparent, easily accessible policies and claims procedures help maintain trust between the MI institu- tion and policyholders. Social funds and similar institu- may develop from a natural extension of existing micro- tions with good community presence can help ease this finance provision or in coordination with health care transition.an iterative process of product design, testing service delivery (ibid).to serve the poor, microinsurance and roll-out is preferable. Product development has to products must be specifically tailored to the poor's prior- be client-centered, competition-focused and matched ity needs for risk protection in terms of coverage types, be to the capacity of the institutions (both insurer and easy to understand, and offer affordable premiums (ibid). delivery institutions). each step must be informed by an understanding of the clients' needs and financial capacity risks faced by Poor householdsand to honor long-term financial obligations; the competi- iMPlications for Microinsurance tive landscape; and MI institutional requirements. Market Product deVeloPMent research should investigate the regulatory environment; range of competing social protection instruments and the capacity of households to cope with a shock insurance products; potential delivery channels; and risk depends in part on risk source, correlation, frequency data1 necessary to develop a MI product.2 and intensity. risks can be natural (e.g., natural disasters) or human-induced (e.g., economic shocks). risks can be the case of afatVimo highlights the role of intermediar- correlated among individuals from the same locality ies and the potential for combining insurance schemes (i.e., covariate risk), as in the case of floods or droughts, that meet the specific needs of low-income house- or be uncorrelated and affect only individual households holds at affordable premiums.3 the all-India disaster (i.e., idiosyncratic risk), as with illness or accident. Further, Mitigation Institute (aIdMI), together with the Provention risks can be low frequency but with high economic Consortium, introduced the micro-insurance project `afat impact (known as catastrophic risk), or high frequency with low economic impact (non-catastrophic).the nature 1Unavailable or flawed risk data (mortality and morbidity tables) adds to the difficulty to construct actuarial fair insurance products for low- of risks insured requires different MI product design income countries. responses. Savings, credit, emergency loans and self- 2For a more detailed discussion of MI product development see Churchill (2006). insurance are more flexible instruments than insurance, 3this description of afatVimo draws on Southasiadisasternet (2003). t h e wo r l d b a n k 2 o C to b e r 2 0 0 8 Vo l U M e 5 n o. 2 high premiums are neither market-competitive nor box 2:common user Priorities for accepted by low-income households. Premiums have to Microinsurance be set in line with the financial abilities of low-income households while meeting the costs of service provision · life insurance and the capital needed to settle claims. new MI products · health insurance should only be developed with the assistance of an actu- ary or insurance expert.5 · agricultural insurance to illustrate this, we look at the case of a Microfinance · livestock insurance Institution (MFI) in kenya.6 In 2001, the Catholic organi- zation Cent based in kisumu, kenya introduced through its MFI a health insurance program called Community health Plan (CheaP). however, staff had underestimated Vimo' in 2004 as pro-poor disaster insurance.4 afatVimo the actuarial knowledge necessary to design an insurance combines micro-insurance, microcredit and micro-miti- product and linked the insurance too closely to credit and gation products for low-cost local risk transfer. It insures savings.as a result, premiums were set significantly below policyholders in the event of 19 disaster types, including potential health care outlays. CheaP quickly realized that earthquakes, cyclones and landslides. non-life damages to that a financial failure would jeopardize its institution- a policyholder's house, household assets, trade-stock, and/ building efforts in the community and harm the potential or losses of wages and livelihood are covered up to Inr for further insurance initiatives.the organization learned 75,000.the life insurance component pays out Inr 20,000 that insurance is a more complex financial product than in the case of death.Yearly premiums are set at Inr 146 savings or credit instruments.the key sustainability chal- (i.e., roughly three days of wages for agricultural laborers). lenge for microinsurance programs is that of striking a afatVimo specifically targets the poor among the disaster- balance between coverage (of large numbers of poor affected. beneficiaries are those who are: i. at risk of being persons), costs (for the insurer) and affordability (for the disaster-affected; ii. enrolled in aIdMI's livelihood relief client) (Churchill 2006). Fund; iii. earning an annual income of Inr 12,000­18,000; iv. hold average assets of Inr 9,000.the unique product Microinsurance ProducttyPes design element of afatVimo lies in the way it bundles non- life and life insurance components from different insurance Microinsurance can be designed in myriad ways. companies into a single policy.aIdMI acts as an intermedi- depending on the risk insured, activity levels, employed ary for afatVimo, facilitating interactions between commu- assets and risk exposure, different insurance types may be nities and insurance companies.aIdMI settles upfront the appropriate.the following provides an overview of com- premiums payment of the beneficiaries to ensure immedi- mon MI product types. ate coverage. Subsequently,aIdMI collects the premiums and supports the beneficiaries with claims settlement. Health Insurance Capacity-building is a key focus:aIdMI trains policyholders for emergencies, and also on legal and procedural require- health insurance directly addresses disease, reduces ments. From 2004­06,afatVimo's membership grew by mortality, and improves health. notably, the world bank's 675% to 5597. renewal rates averaged around 88%, signal- Social Protection Strategy Paper (world bank 2001) ing the strength of its unified policy design. identifies health insurance as an important complement Pricing 4rrtI works also in partnership with the hazard risk Management Unit of the world bank and the International Federation of red Cross and Setting an insurance premium rate is a tricky task. low red Crescent. 5For more information on formal pricing, see Churchill et al (2003). premiums can de-capitalize insurance providers, while 6this case draws on McCord and osinde (2002) and McCord (2006­08). t h e wo r l d b a n k 3 o C to b e r 2 0 0 8 Vo l U M e 5 n o. 2 to poverty reduction efforts. Microinsurance for health and debt payments and principal. life insurance products can be designed to cover groups (e.g., MFI clients, coop- are not affected by moral hazard issues, and unlike health erative members), as well as households. [Individuals are insurance, do not require existing physical infrastructure. not insured, however, in order to control for adverse the simple structure of life insurance allows for a variety selection].7 high-risk persons such as the elderly are of marketing and distribution channels. life insurance can also typically excluded to keep premiums at an afford- be relatively easily bundled with other types of insurance able level. Further, coverage periods of less than one year to structure a product that specifically meets the needs of pose administrative difficulties, and also pose the risk of low-income families. For example, credit life insurance can adverse selection due to seasonal trends in illness inci- be sold together with loans.the premiums are collected dence (e.g., onset of malaria season) (Mc Cord 2007). with loan repayments in order to reduce administrative costs. Funeral and life insurance can be delivered through health insurance is the most difficult MI product to imple- funeral parlors or MFIs. Some insurance companies in ment, requiring significant managerial and actuarial capacity India draw on consumer retailing strategies to sell their (see Figure 1 for a rough approximation of levels of pro- products in computer kiosks or bundled with cell phone gram design difficulty among insurance products). health packages (Churchill 2006). insurers need to understand risk management techniques and the solutions for controlling adverse selection. health Index-Based Insurance: A Recent Innovation in insurance is only effective where there is an existing Risk Management Instruments infrastructure of health service providers and accessible hospitals--even if at a distance from policyholders. Poor households are particularly vulnerable to cata- strophic weather events that threaten crop failure and Life Insurance livelihood loss. Index-based insurance is an innovative instrument to overcome shortcomings of traditional death of a household's main breadwinner severely agricultural insurance, including adverse selection, moral impacts household welfare. life insurance can mitigate the hazard and administration costs.the payout and payment financial shock of the breadwinner's death, by providing structure in index-based insurance is predetermined income assistance to the family; covering funeral expenses, and triggered by an index (typically rainfall) highly cor- related to a particular crop yield or livestock mortality rate. data from weather stations is used to calculate the index. Payment starts when the index falls below a certain figure 1:Programtypes:ease of design and threshold. success Index-based insurance is also suitable for risk layering. High Ease of Design/ Higher Success Rates depending on the extent of the loss, the farmer, insurance · Credit life company, state or donor community can cover the losses. · Term life/ personal accident In Mongolia, the Government of Mongolia and the world · Savings life bank introduced an index-based mortality livestock insur- · Property insurance ance in 2006 that worked with three different risk layers. · Endowment life the insurance is linked to the mortality rate of herds of · Health insurance all types and sizes and is triggered when losses exceed the average mortality rate of 7 percent. (losses under the · Agricultural insurance Low Ease of Design/ Lower Success Rates 7It is possible to insure individuals under microinsurance programs (e.g., braC program,) but costs are double those of programs using group pricing (Churchill 2006).thus, a high participation rate amongst Source:adapted from Churchill 2006 the target group is required to keep the program financially safe (ibid). t h e wo r l d b a n k o C to b e r 2 0 0 8 4 Vo l U M e 5 n o. 2 7 percent of herd value threshold are borne by the herd- ance.the Cbo pools, manages and absorbs the risk. ers). Private insurance is triggered for losses between this model fosters strong "ownership" by Cbos and 7 percent to 30 percent.the Government provides the member policyholders. Community involvement and final risk layer, indemnifying losses exceeding 30 percent peer-monitoring reduces information, enforcement costs of the estimated value of the herd (alderman 2007). (transaction costs) and the probability of adverse selec- tion and moral hazard. Studies have shown that com- Use of such a transparent index mechanism reduces munity participation achieves better targeting outcomes administration and transaction costs for identifying losses, and reduces the administrative costs of handling transfer and limits problems of moral hazard and adverse selection payments. (since the index cannot be influenced by individuals, and payouts are predetermined). Secondly, it makes agricul- however, Cbos usually lack the management and actu- tural insurance more attractive to international reinsurers arial expertise and financial backing of regulated insurers. and provides protection against correlated8 risk. It intro- Cbos often base their pricing strategies/ premiums on duces different risk layers to allow for wider coverage. Its peoples' financial abilities, and not on the required finan- design features also allow for quick response during disas- cial and managerial resources needed to provide adequate ter response efforts and help improve agency response coverage. MI insurance companies need to have enough time in settling claims. retailing of index insurance can cash reserves at hand to balance cash flow fluctuations. take different forms. Index contracts are either sold indi- Cbos usually have only limited reserves and thus can run vidually or bundled with related risk-management services the risk of not honoring their payment obligations. lack (microfinance, technical assistance, advisory services). In of reinsurance partners further constrains their ability to India, a seed company acquired rainfall insurance which properly manage deficits. Cbos often have weak manage- they sold together with their seed packages. ment controls. legislation and regulation may limit the expansion of MI (e.g., only licensed providers are allowed the major limitation of index-based insurances is the to sell insurance in India). Government regulators are probability that the insurance does not represent indi- also typically skeptical about the abilities of non-insurers vidual losses.this is referred to as basis risk. basis risk can to manage insurance programs. Finally, gender, kinship, significantly reduce the acceptability of the risk manage- geographical proximity, ethnicity, social networks, wealth ment instrument. Group coverage and risk pooling of have a strong influence on the level of inclusion in com- farmers can reduce basis risk by allowing the group or munity insurance networks.the geographical boundaries community to allocate the funds among themselves, given of Cbos effectively limit the amount of potential policy- that local users will have improved information on losses holders and the size of the risk pool, meaning Cbo-linked to individuals, including the ability to verify losses. insurance programs may fail to provide protection against covariate risks. deliVery Models for Microinsurance Full Service Model delivering the best possible benefits and affordable premi- ums to poor individuals can occur only when administra- In this model, a nGo or other organization operates tive costs are minimized. different distribution models the insurance scheme and fully absorbs risks, profit and can be considered, with various cost effects. loss. Full service models require substantially investment in human and financial resources and acquisition of actu- Community-Based Organization (CBO) In the Community-based organization (Cbo) model, 8Correlated risks arise when a weather event is affecting a large number local community organizations, MFI, nGos, or coop- of farmers in the same region. Small-scale financial institutions do not have the financial capacity to cope with such a loss and have to diversify eratives jointly develop and distribute their own insur- the risk of the portfolio by attracting international reinsurers. t h e wo r l d b a n k 5 o C to b e r 2 0 0 8 Vo l U M e 5 n o. 2 arial expertise before becoming operational.this sort of Social Funds are well placed to assume the role of the approach is not widespread.an example is SPandana intermediary in areas that lack well-established nGos in Guntur,andhra Pradesh, India which serves poor urban or MFIs.they can also help provide the start-up costs for women in coastal cities of that state. SPandana became microinsurance programs. fully operational in 1998 and now has over 115,000 cli- ents, served by 181 staff in 31 branches, covering a total agency linkages credit and insurance portfolio of over USd 12 million. Credit life insurance (i.e., loan protection) is bundled in linking MI programs to other MFI schemes and partners a compulsory manner with the nGo's loan products (at is a helpful strategy to compensate for some of the dis- 1 per cent of the loan amount), so that in the event of a advantages outlined above and to create economies of female borrower's death or death of her husband, or in scale (Churchill, 2006). national social protection pro- case of fire event, the loan is written off.a nominal life grams may also be complemented with MI elements. risk insurance policy is also included for female borrowers: layering can be undertaken by linking with reinsurers or they receive around USd 110 in case of the death of their insurance federations. Service delivery can include direct husband (M-CrIl 2005). Credit life insurance is the sim- contracting of nGos or public health programs, as well plest form of insurance and a good one for small groups as bundling with other products such as those provided entering the insurance market. by coops. Provider Model the key microinsurance challenge lies at the nexus of coverage, costs and affordability. retrospective premium Microfinance institutions and commercial banks can collection for example leads to increased risks to the directly market MI products to potential clients, as in the agency, as well as increased administrative costs. Mobile provider model.this model requires a well-established and dispersed client bases in some countries (such as pas- distribution network and is widely used in the general toralists or circular migrants) also pose significant though insurance market.the model suffers from high transaction not insurmountable challenges for product design, and costs, when applied in low-income, low-margin markets highlight the need for a socially-grounded understanding such as rural areas with dispersed populations. of the prospective client base and sufficient due diligence on household risks and carrying capacity. Partner-Agent Model and Social Funds deVeloPing Microinsurance:Policy In the partner-agent model, insurers (both commercial iMPlications for social funds and public) collaborate with an MFI/ nGo to develop a MI program.the MI programs then use as intermedi- as noted above, social funds are a particularly viable ary (such as a nGo or MFI institution, or local bank) platform for delivering microinsurance products. Social to liaise between the customer and insurance company, Funds (SF) channel grants to communities for small-scale and manage marketing and administration functions.the development projects such as road and school construc- insurer bears the risk of the insurance policy while the tion (de Silva and Sum 2007). Social funds typically assist MFI/ nGo utilizes its distribution channels to bring the groups affected by e.g., natural disasters and in some product to communities. MFIs/ nGos with strong ties to cases provide microfinance services, among other activi- communities are most successful in this model.they train ties. Institutionally, SFs enjoy a high degree of financial their clients in MFI products, are experienced in transac- and operational autonomy. SF's use of private sector-style tion processes, and raise financial services awareness management practices (e.g., use of results-based monitor- among low-income households.this model minimizes ing and evaluation; procurement guidelines; competitive distribution costs, while increasing outreach as well as recruitment) results in higher operational efficiency.at the affordability. project level, Social Funds provide a bridge to community t h e wo r l d b a n k o C to b e r 2 0 0 8 6 Vo l U M e 5 n o. 2 actors such as local government, community groups and conclusion nGos. SFs' longstanding experience in supporting local institutional development, public goods and services, and this note has discussed the role of microinsurance local arrangements for ex-ante risk management translate in mitigating external shocks on poor households. well into their potential to manage and implement micro- Microinsurance has been shown to be a powerful addi- insurance. tion to the social risk management product toolbox, and one that is flexible enough to be successful implemented Social Funds' structure and approach offer the following under a variety of institutional forms, including Social advantages for MI provision. Funds. · Institutional Capacity:MI programs can piggyback on nonetheless, careful attention and expert technical input the managerial and organizational infrastructure of is required in designing microinsurance products and pro- SFs to decrease administrative and start-up costs. grams as they are significantly more complex than savings and credits programs offered by different organizations. · Inclusion of marginalized groups:Cbos often fail to Use of risk layering, using different forms of reinsurance cover marginalized groups (e.g., women, poor) in to cover the insurer is crucial from a financial sustain- their activities. SFs can help offset costs for mem- ability standpoint, and the use of various outreach mecha- bers of marginalized groups by subsidizing premiums nisms to reach poor households is necessary from an (though this option does present some moral hazard equity point of view. Some microinsurance product types risks and should be carefully assessed and moni- are more easily designed and implemented (such as credit tored). SF funds can also be deployed to cover for life insurance, i.e., loan insurance) than others (e.g., health delayed or missing premium payments.9 insurance.as the microinsurance practitioner community develops further, it will be important to develop perfor- · Risk Management and Risk Pooling: SFs' size and mance benchmarks, refine and codify delivery models, and resources can augment Cbos with a larger risk pool engage in information exchange and shared learning pro- and well-trained staff for risk and cash management. cesses, especially South-South dialogue. Microinsurance SF's greater geographical reach and ability to medi- offers the potential for significant innovation in public-pri- ate between public and private insurers allow for vate partnership arrangements, cooperation across volun- the inclusion of additional risk layers (including rein- tary and private sectors, rural and urban services sector surers) to externalize and diversify risks to a wider development, and the extension of social protection to spectrum of market participants. underserved populations, for years to come. · Product Development and Support:SF participatory Works cited assessment and development techniques can be employed to develop demand-driven MI products aGroaSeMeX. 2006."the experience of Mexico in the and enhance "ownership".the known "brands" of SFs development and operation of Parametric Insurances can also help signal the trustworthiness of MI prod- applied in agriculture." Processed. ucts. (In as related vein, SFs must also conduct due diligence on MI products potentially on offer so as to alderman, harold, and trina haque. 2007."Insurance reduce their own reputational risk). SFs can also help against Covariate risks:the role of Index-based create demand for MI products and gather essential risk data, resulting in significant lower monitoring, transaction and enforcement costs for MI programs. 9Careful design can help bridge moral hazard problems of SFs as guar- SFs' cost advantages translate in lower insurance pre- antors as in the afatVimo case above where the intermediary is used only to pay premiums upfront in a time-bound fashion, rather than as miums therein increasing product demand. performing a full subsidy role. t h e wo r l d b a n k 7 o C to b e r 2 0 0 8 Vo l U M e 5 n o. 2 Insurance in Social Protection in low-Income Countries levin,thomas. 2007. Microinsurance aspects in of africa." World BankWorking Paper 95. agriculture. Munich re Foundation. ------ 2006."Countercyclical Safety nets for the Poor M-CrIl. 2005. Microfinance risk assessment: andVulnerable." Food Policy 34(4). SPandana. hyderabad: M-CrIl. bhattamishra, ruchira and Christopher b. barrett. 2008. McCord, Michael. 2006­2008: Microinsurance note 1­9. "Community-based risk Management arrangements: washington dC: USaId. an overview and Implications for Social Fund Program design".working draft. McCord, Michael and Sylvia osinde. 2002. Community health Program: CheaP ­ kisumu, kenya, notes from a Churchill, Craig ed. 2008. Protecting the Poor:A Visit July 2002. Microinsurance Center of Microsave africa. Microinsurance Compendium. Geneva: International labour organization (Ilo). narayan, deepa and katrinka ebbe. 1997. Design of Social Funds.world bank discussion Paper 375.washington dC: Churchill, Craig. 2007.`what is Microinsurance? access to world bank. Insurance for the Poor'. May 2007. rio de Janeiro, brazil. Powerpoint presentation. Sebstad, Jennifer. 2006."Guidelines for Market research on the demand for Microinsurance". MicroReport 69. ------ 2003. Making InsuranceWork for Microfinance USaId. Institutions:ATechnical Guide to Developing and Delivering Microinsurance. Geneva: International labour organization Sen,amartya. 1999. Development as Freedom. newYork: (Ilo). anchor books. davis, Ian ed. 2006."Community risk transfer through Qureshi, Zahid ed. 2007.Microinsurance Conference Microinsurance:an opportunity for South asia. 2007: Making Insurance work for the poor. Munich re Southasiadisasternet 13. Foundation. de Silva, Samantha and June-wei Sum. 2006." Social Funds world bank. 2003."Social risk Management:the world lending trends FY 2000­2006:an evolving role in Social bank' approach to Social Protection in a Globalizing Protection".washington dC:world bank. world".washington dC:world bank. Gross,alexandra and Samantha de Silva. 2002. ------ 2001. Social Protection Sector Strategy: From Safety "Social Fund Support of Microfinance:a review of Net to Springboard.washington dC:world bank. Implementation experience". SP Discussion Paper 215. washington dC:world bank. Social Funds Innovations notes are published informally by the Social Funds thematic group of the human development network ­ Social Protection.the findings, interpretations, and conclusions expressed in this note are entirely those of the author(s) and should not be attributed in any manner to the world bank, to its affiliated organizations or to members of its board of executive directors or the countries they represent. For additional copies, contact the Social Protection advisory Service,the world bank, 1818 h Street, nw,washington, dC 20433, USa, e-mail: socialprotection@worldbank.org. Copies are also available on-line at http://www.worldbank.org/socialfunds. t h e wo r l d b a n k o C to b e r 2 0 0 8 8 Vo l U M e 5 n o. 2