Micro-Insurance : Extending Health Insurance to the Excluded

MICRO-INSURANCE

Extending Health Insurance to the Excluded

David M. Dror & Christian Jacquier

 Social Security Department, ILO (Planning, Development and Standards Branch)

Abstract:

            This paper proposes a way to improve health provision for populations who are usually excluded from access to health services. It starts out from a short description of who the excluded are, and what they are excluded from. The paper then looks at the major policy statements elaborated at the international level, and proceeds to propose the missing dimension. Next, based on field testing and analysis of several tens of pilot cases, the paper proposes a concept for group-based health insurance, or “micro-insurance”, and explains its rationale and its components. Last, the paper proposes a strategy to implement this concept.

Who are the excluded?

            Access to health services is still inadequate among the majority of rural populations, among excluded populations in most low- and middle-income countries (LMIC). The term “excluded populations” should not be confused for a synonym for the poor. The difference, in essence, is that whereas poverty means non-access to goods and services due to unequal material resources, social exclusion refers to inadequate or unequal participation in social life, or exclusion from a place in the consumer society, often linked to the social role of employment or work (Duffy, 1995). This distinction is important in that it modifies the focus of interest from the link between poverty and ill-health to the link between the social gradient and health. It has been proposed (Bobak, Blane and Marmot, 1998)  that the social gradient reflects degrees of exclusion, representing varying lengths and extent of exposure to depriving conditions.

            Within populations of the same country there are significant differences in health outcomes and morbidity status that are linked to socio-economic inequalities (Bobak, Blane and Marmot, 1998). The effect of unequal distribution can be relative to the standards of specific countries, and  is more visible in countries where society has incorporated market-based social organization than in countries where society is based on traditional family and community links. In countries where family and other social networks continue to be strong, there is less risk of exclusion even in the face of inadequate provision of services by the state. In countries that used to provide virtually all health needs, such as the former Soviet Union (FSU), profound transformations since the dismantling of the FSU have reduced these services, but exclusion from health services is directly related to other aspects of social exclusion (such as work, housing, education etc.). Individuals cumulating intense or multiple exposure to problems are the most vulnerable . The problems are illegal or irregular status, migration, weak relationship to family or to personal network (prevalent among the elderly, substance abusers, orphans, people with psychological problems etc.), weak relationship to the labour market, poor housing and belonging to a discriminated-against group (Duffy, 1998). In many LMIC where the state never provided more than rudimentary services, such as in sub-Saharan Africa, exclusion is linked to the inefficiencies of the market and of society to incorporate certain sub-groups who cumulate a different profile of impeding characteristics, such as low income, malnutrition, low health status (and hence low insurance status within for-profit insurances), rural habitat, low education and ethnic/tribal origin. This profile is not unique to developing countries, but is relatively prevalent in LMIC.

What are they excluded from?

            Although the problem of exclusion has ramifications on several vital domains (e.g. work, housing, education, social protection, health), this paper deals only with exclusion from health services. An indicator of the gravity of this problem is its incidence: whereas the rate of  exclusion from health services is almost nil in most OECD countries (except Mexico, Turkey and the USA), it is estimated at 20% in all developing countries, and at 51% in the least developed countries (period 1990-95). Lack of access to health services has resulted in lower health status, reflected in higher morbidity and mortality.

            Most countries have some form of public provision of health services. Less and less  countries  provide a full range of services, based on the needs of the population. In the absence of both government-managed risk pooling and a subsidy for the needy, most people seek alternative mechanisms to reduce their exposure to the financial risk of ill-health. The degree of this exposure is influenced by many factors. Let’s look first at public expenditure on health. Where public resources are spent mainly on infrastructure (e.g. hospitals and health centres) and on professional services, only those who have access to these facilities can benefit from the public subsidy, and frequently access to the publicly-financed facilities and services is available to some more than to others (depending on their occupational or insurance status, or on their place of residence). But even in the best case, public financing represents only a fraction of the overall spending on health care (not to mention “informal care” which is often disregarded or considered as a free resource). Shortage of resources is blamed for the problem of insufficient public provision of health services. However, absolute levels of expenditure on health reveal only part of the picture; another aspect is the share of health budgets that reaches lower and remoter deciles of the population. In some countries, restrictions in the quantity (through reductions in the types of benefits) and the quality (through restrictions on individual choice) of publicly financed health services oblige people to pay for services. Although in theory the poorer segments of society should continue to receive care from the public system, those services are minimal, and involve transaction and ancillary costs (such as transportation from domicile to the service location) that may render the “free” services out of reach. In other cases, user-fees and other expenses such as drugs have made access to health care even more inequitable through regressive taxation, but have also resulted in behavioural changes (e.g. delaying or avoiding care).

In yet other situations, the problem is not so much access to primary care but coping with idiosyncratic risk. In essence then, the reply to the question “what are they excluded from” point to financial arrangements, in particular health insurance, rather than to a specific type of service. And this problem is unfortunately very prevalent: a survey done in 1987 by the World Bank in 33 Sub-Saharan African countries showed that only seven countries had health insurance systems, with coverage of the total population ranging from .001% in Ethiopia  to 11.4% in Kenya (where only civil servants and a few other formal sector employees enjoyed this benefit). This leads to a conclusion that, in the absence of public provision of free-of-charge services, excluded populations can enhance access to care if they can access health insurance. Thus, the problem we deal with in this paper is precisely exclusion from health insurance.

Can the market solve the problem? 

            The choice how to provide health insurance rests with governments, and in most LMIC  the choice has been to allow “the market” to fill the gap left by non-existent social health insurance schemes. One of the arguments put forward in favour of this process has been that the market is less “bureaucratic” and more flexible to adapt to the changing needs of people. So the question is:  can the market solve the problem, by offering health insurance to the excluded?

            It is widely agreed that health care systems left to function according to market forces alone do not result in a socially optimal quantity, quality or distribution of health care (Outreville, 1998, quoting  Hsiao, 1995 and  Newbrander, 1997). For one, there is the problem of supply of health services in provincial and peripheral areas. Assuming that sparse supply causes inequity in access to services presupposes that the market is supply-based, and that those who produce health services also create the market and facilitate the transfer. On the other hand, insufficient supply of health services may well be an inefficient reaction to what is viewed as insufficient demand. The equation linking weak demand to slight ability to pay ignores two major issues: first, the inefficient economic interactions of excluded populations in a predominantly monetary economy, and secondly the rigidity of financial instruments, which are unadapted to the conditions prevailing among these population segments. Consequently, health needs of excluded populations are often not structured in terms of “solvent demand” (i.e. prioritized needs for which the needy can pay in monetary terms). These problems can be corrected, but the correction may take a very long time unless something is done to accelerate it. This  in turn could improve supply.

            Similarly, in many LMIC, the supply side (both of services and of insurance) also needs help to reach a critical mass and a geographic spread allowing it to interact with the demand. Governments can help this process, by legislation defining the domains of activity, or by providing financial guarantees to cover certain risks (e.g. the Dutch model of Exceptional Medical Expenses Act) or by facilitating the relations between insurers and providers (particularly where they own or control the provision of care).

            The market has so far not been a guarantor of sufficient supply of health care for excluded populations. Nor is health insurance per se a guarantor of better optimality. To the extent that they have been free to control their operations, private health insurers have preferred to select the demand associated with high income groups concentrated in few urban centres. In addition, many interactions between clients, insurance institutions and providers revolve around controlling the delivery of non random, low-cost & high frequency services, rather than controlling health risks. To change this structure, health insurers must refocus their activity on health risk, but is this likely to occur without political reorientation on the role of health insurers? Unfortunately, the excluded command little political support. Hence, reliance on the market alone leaves little hope that universal access to health insurance is likely to happen.

How have governments viewed their role?

            International fora have recognized the need to provide access to health care.  As early as 1944, the Declaration of Philadelphia, adopted by the International Labour Conference, “recognises the solemn obligation of the International Labour Organisation to further among the nations of the world programmes which will achieve …. the extension of …  comprehensive medical care”…. In 1952, the International Labour Conference adopted the Convention concerning Minimum Standards of Social Security (No. 102), containing a chapter on provision of medical care by law, in States ratifying that part of the Convention.  However, this Convention was adopted in a world where most development was hinged on the formal sector. The WHO Declaration concerning Health For All by the Year 2000, also known as The Alma-Ata Declaration (1978) was first to de-linking health coverage from employment status. Alma-Ata was the launching pad for a two-pronged approach: health for all (HFA), which was fundamentally a call for social justice as part of development; and primary health care (PHC), or  first-tier universal coverage organized by governments. Ten years later, it was recognized that universal coverage under PHC was not attainable in most LMIC through governmental health systems. Instead, communities were encouraged to assume this responsibility through “their ownership and control of their own endeavours and destinies” (The Ottawa Charter, 1986). It was recognized that behavioural change occurs at community and household level, and that  even poor households can, and do, pay for health care, and their community can generate income to cover the recurrent costs of basic health units, to complement or replace weak public institutions (The Bamako Initiative – BI, 1987)

. This entailed substantial decentralization of decision making from central government to lower levels, community financing of health services and community control of the management of the health system and the flow of funds.  In the last decade, all global UN conferences have included a resolution recognizing PHC as a basic human right (UN ACC, 1997). At the same time, marketization policies aimed at creating opportunities for individuals rather than social outcomes have enabled countries to ignore the difference between opportunities and rights. The gap between these two concepts has been seen most clearly among those who were unable to seize the opportunities, and who were not covered as a matter of a right. The delegation of responsibility to communities, non-governmental organizations (NGOs), the labour movement and the private sector (The Jakarta Declaration, 1997) was testimony to a  significant downgrading of government implication. Yet,  changing the focus from top-down to bottom-up action has not made significant headway, despite international recognition of its importance as the way forward. Why?

The missing dimension

            International support for decentralized activity has stopped short of recognizing three determinant operative conditions:

Civil society does not usually spontaneously organize  programs for provision of access to health care at the community level, but this can be prompted and helped.

Communities can best fulfill their bottom-up activity through voluntary affiliation. To be able to attract members voluntarily into a health insurance, the group must be free to adapt the scheme to its specific health risks and living conditions.

Regardless of who initiates group-based health insurance ( NGOs, governments or international), such activity needs the political, moral, technical and  financial support of governments. One particularly important help  governments can provide is to enable direct purchaser/provider interactions by decentralizing decision making within health facilities to the local level, thus empowering the direct parties to negotiate the conditions  of access to care.

A new concept:  MICRO-INSURANCE

            A preliminary comment is necessary to explain the name “micro-insurance”. Why not use “community-based”? The answer is that “community-based” covers a large variety of  experiences which differ from each other on several important points. Some community-based activities are recognisable in the concept labelled  micro-insurances, which attempts to structure the fundamental features around precepts of social and economic characteristics. Some examples of the differences may be useful. For one, practically all existing community-based operations depend on continuing access to some form of external subsidy (Bennett, Creese and Monasch, 1998), and many depend on in-situ presence of outside facilitators. Micro-insurance is conceived as an autonomous enterprise, independent of external operators or of permanent financial lifelines. Next,  the locus of decisions in micro-insurance is with each unit, whereas many community-based operations are in fact dependant on decisions taken far away (either at the level of governments, or NGOs etc.). At the same time, micro-insurance is more than just an activity at the level of single communities; it foresees setting up networks to link multiple small area- and occupation-based units into larger structures that can enhance both the insurance function (through a wider pooling of risks) and the support structures needed for improved governance (through training, data banks, research facilities etc.). Micro-insurance does not pretend to be THE solution, but aims to integrate all types of solutions within a concept that focuses on action at grass-roots level.

            Another query about the name goes something like: bearing in mind that insurance, by its very nature, is a macro-economic tool, is “micro-insurance” not a contradiction in terms? The answer is that “micro” refers to the level of society where the interaction is located, i.e. smaller than national schemes, and “insurance” refers to the economic instrument. A more accurate descriptor of the proposed concept might perhaps be: voluntary group self-help schemes for social health insurance. For ease of reference we suggest calling it “micro-insurance”.

            The underpinning of micro-insurance is that excluded populations have not been covered under existing health insurance schemes because of two concurrent forces. The first is that – notwithstanding important differences between social and private health insurances – insurers have done too little to include these population segments. The second factor has been that excluded people have forgone claiming access because of their dis-empowerment within society. Micro-insurance proposes to change both factors, by basing its activity on the following assumptions:

To be attractive to excluded populations, insurance must be adapted effectively  to living and working conditions of people, which are usually area- or trade-specific.

Effective adaptation can occur through a process of mediation composed of two essential functions: empowerment and increased social capital. The first happens by enabling the population to express its needs and priorities; and the second occurs through forging a receptive public opinion towards the insurance. The process requires  trust-building measures that will defuse the community’s aversion towards up-front payment for a deferred return. Observations of social interactions show that the community can be mobilized to bring about this change. Local public opinion leaders are decisive in this process, as is popular involvement in decision-making;

Even the most successful local micro-insurance unit cannot develop into a full-fledged insurance on its own, and thus needs to be bound together with others to add up to the critical mass. The necessary mechanism for this aggregation needs to be elaborated, bearing in mind that it has to favour solidarity and equity.

The bottom-up and the top-down vectors of health activity need to be intertwined.   A-priori, this can be done without umbilical links to a governmental health insurance system. The relations between disparate units and the government need not be defined at the outset; nor need the debate over the relationship between the civil society and the state hinder the development of micro insurance.

            Despite the smallness of each unit, micro-insurance justifies being characterised as  insurance, and should not be confused with a credit facility or a savings account. The main difference between the two is that savings is inter-temporal trade (of current consumption for future consumption) whereas insurance is a trade across states of nature, from good state-of-nature to a bad one (Besley, 1995). Also, savings respond only to individualistic preferences, and as such offer no opportunity for inter-personal trade, whereas insurance is an economic interaction based on mutual determination of preferences. The experience with Medical Savings Accounts (MSA) in Singapore or elsewhere throws light on the difference between saving and micro-insurance. MSA is a plan to encourage (or coerce) individuals to save, with no or negligible risk pooling or solidarity. (MSA also includes rules about spending those savings under dissuasive conditions). In contrast, micro-insurance is a mechanism to pool both risks and resources of whole groups, to provide protection to all members against the financial consequences of mutually determined health risks        “Alternatively, ‘saving for an unexpected health threat’ is not an economically efficient substitute for insurance. A person with a low-probability expected loss of $3,000 could put $60 a year under the mattress for fifty years, and finally achieve (near the end of life) the protection that would otherwise have been available every year by buying the above insurance plan for $0.60 a year (for a total lifetime premium of only $30.00). The efficiency gains due to insurance are therefore obvious. There is no alternative to pooling of risks that provide the same level of protection”.  . This difference is cardinal in terms of who bears the responsibility for the health risk: the savings plan puts that responsibility exclusively on the beneficiary, whereas micro-insurance  recognizes the  complementary roles of community and individual responsibility.

            Affiliation to micro-insurance is assumed to be voluntary; it is thus important to understand what motivates individuals to join. The underlying economic motivation for joining a micro-insurance unit is assumed to be a desire to seek reciprocity in sustaining risk-sharing arrangements among essentially self-interested individuals. This motivation is sometimes contrasted with altruistic feelings among  members joining a “friendly society”. The difference between the two approaches influences the consequence of withdrawing from the group, particularly once an individual has gained more than s/he has paid. In a friendly society (and in commercial insurance), withdrawing after profit-taking is tolerated (although the two differ in the reasons for this tolerance). In sustained risk-sharing arrangements, the penalty for exploiting the group means exclusion from future benefits through blocks to re-entry into insurance. The defecting/excluded individual must hereafter rely only on his/her income to pay for all contingencies. This may be a very severe threat, particularly for persons who have no personal savings. This strict trade-off can explain why even poor individuals accept to remain insured although they may dislike to pay a premium and get no benefit for it for a long time.

            A second motive to join micro-insurance is people’s desire to  improve their health status by controlling the conditions in which they live and work. People do not live and work in isolation from others; on the contrary, people have a deep-rooted need to seek voluntary and repeated interaction with others. Persons who have a formal employment relationship transfer part of their attachments from the family and the immediate wider community to the workplace in return for the rewards gained through employment. The excluded population frequently cannot do so, for lack of access to a workplace or the social protection that is available through it. The alternative source of support is the community. Members of the same community rely on each other in many ways, and refer to each other in a context of roles, values, habits and customs, to satisfy moral and material needs. These links can help to improve the conditions of life, provided that individuals adhere to collective objectives. There is a hypothesis that outside pressures enhance group cohesion. Applying this hypothesis to populations which cumulate several unfavourable conditions, we deduce that members may also join micro-insurance in response to group cohesion or pressure.

            The two motivations to join are mutually supporting, as they respond to economic and societal needs. The acid test of affiliation is payment of premiums. Economic theory predicts that the incentive to remain within the group will have its greatest effect if individuals in the scheme face large differences in their incomes, for it is then that they need to transfer large sums to equalize their marginal utilities of income (Besley, 1995, p.2167). It is proposed that while health insurance does not necessarily depend on large income disparity to thrive, micro-insurance does gain relevance if it can flatten out  income  fluctuations due to large temporal variations in the flow of income, as is typical for rural or poor populations.

            The next key characteristic of micro-insurance is the choice of risks that are covered. Drawing on the analysis of over 100 test cases (Atim, 1998, pp. 17-21 and Bennett, Creese and Monasch, 1998, pp. 9-12), it is possible to say that two approaches prevail: on the one hand, high-cost and low incidence “catastrophic” events (such as treatment for snake bites, emergency treatment of delivery complications, some cases of hospitalization); on the other hand, non-random, low cost events resembling primary care (including drugs, laboratory, supplementary care beyond that provided by government). This two-pronged approach implies that at least at the outset, there is no standard model for design of the benefit package. Granted, the package may change, depending mostly on the amount of accumulated reserves, and the community-specific perception of priority risks/benefits. In this, micro-insurance units differ greatly from private, profit motivated insurances that modify the benefit package to improve profit (and/or increase market share), as well as from social insurances that focus on a heterogenous population and are oriented towards bio-medical services.

            The process of deciding on the risks that should be covered concentrates on social needs,  in contrast to the medical approach that concentrates on health services, prevalent among most other insurances and other community-based experiments. This distinction, made by Williams  in a different context (Culyer and Maynard, 1997, p.41) reflects the dominance of members’ needs on the collective decision-making process, and vice versa. This also means that anyone who is not a member of the group would normally be excluded from the decision-making process, including medical staff. In principle, local health workers (nurse, midwife, first-aid volunteer, traditional healer etc.)  can take part in the decisions, provided they join the micro-insurance. This, too, sets micro-insurance apart from traditional health insurances (both private and social ones) in that experts from outside the group can at best have a consultative position.   Micro-insurance is structured to harness group-dynamics in order to link individual economic motives with collective social responsibility, through a process of autonomous decision-making. The process itself is described in a separate section ahead, but it is mentioned here only to point out that social implication in the definition of risks is not merely consultative, but a consensus-seeking  process. Thus, this decisional process serves both motivations mentioned earlier in that it supports acceptance by individuals of the priorities of the group, and it sustains self-interest by enabling individuals to influence the group’s choices. This process demonstrates the advantages of pooling of risks in that the group can defray expenses that none of its members can assume alone. The larger the number of members joining the scheme, the stronger the economic utility. In addition, the insurance of random events enhances redistribution of income within the protected group, particularly if, as is common, contributions are income-rated.

            Micro-insurance units are conceived to be group-based, without assuming a  romantic view of group life. Micro-insurance can function even in communities that do not demonstrate a high level of harmony. It is simply assumed that the joint economic activity of belonging to the same insurance may reduce antagonisms within the community in this specific context, while admitting divergent contentions on other topics among the same population.

            Because of the profile of exclusion, micro-insurance, which is designed to service excluded populations, must be sensitive  to three conditions: it needs to be simple, affordable and located close to members.

Simplicity:  Most people in the informal sector are unable to cope with procedural complexities. Many people simply cannot read or write, and even literate persons may find it very difficult to “fill the form” because they may be unable to provide such seemingly simple details as an address, date of birth or income information. Formalities and procedures must take account of this reality. Simplicity is important not just because of the technical problems. It projects a public image that micro-insurance is approachable by poor people. For the same reason, micro-insurance units have to be prepared to deal with applicants who join when they are in need, expecting immediate support, if for no other reason than that it is easiest to see the utility of insurance in moments of need. Response time on applications for affiliation should be as short as possible. However,  expeditious decisions about induction should not be confused to mean a decision to provide immediate financial help; micro-insurance units may have to impose some restrictions, particularly on expenses that are not random at the time of entry (this is developed in more detail later, under “free-riding”). Another aspect of simplicity is to forgo recourse to mechanisms of exclusion. Commercial health insurances frequently exclude prior conditions from coverage for as long a period as possible. This seems quite impractical with micro-insurance, because of the cost involved with determining such exclusions.

Affordability: What makes micro-insurance affordable?  The absolute level of premiums obviously makes a difference. Affordability is linked, at least partly, to a sense of utility. The perceived return for the premium can be as important as its absolute level. And the confidence that in case of need the insurance will pay for the member also enhances the subjective feeling of affordability. Another feature that will play a role is transparency about expenditures in general and about overhead costs in particular. People tend to view as “affordable” costs that they feel are justified. Streamlining operations to reduce costs is as important in Micro-insurance as it is in micro-finance or in any other financial dealing oriented towards the poor. Another aspect of affordability is the periodicity of payment. Regular periodicity of income flow, characteristic of wage earners in the formal sector, is rare with most people in rural or informal activity. People whose income periodicity is erratic cannot be expected to pay a regular monthly premium. Micro-insurance needs to be flexible enough to enable its affiliates to pay when they can.  Likewise, the knowledge that all members have to abide by the same equitable rules is important. Lastly, affordability may also depend on the type of transaction that is acceptable as payment. Some may have difficulty to pay in cash but may have little or no difficulty to pay in kind or by providing their labour. The community should seek ways to accept payment in kind in cases when cash payment is problematic. This is plausible if micro-insurance units operate as extensions to other communal efforts, such as sale of agricultural crops, or if the insurance recruits staff from among the local community to perform administrative tasks related to micro-insurance, at low operating costs.

Proximity: Micro-insurance units need to be near their client base, simply because the poor or the rural population have neither the means nor the latitude to travel from their place of residence to service centres.  If micro-insurance units were started in a large locality, they might be more successful if launched in several sub-sectors of a large community. Proximity is in fact necessary at all periods of operation, not just at recruitment, since the locus of decisions rests with the community.

Self management    In addition to the main characteristics of micro-insurance mentioned above, another fundamental feature is its community base.  Micro-insurance is the enterprise of the community. The democratic process of jointly defining the risks that should be covered is in itself unique to micro-insurance.  Unlike commercial insurance (where the insurer determines the offer of insurance) or social health insurance (where the government determines the benefit package) it depends on needs-based decision-making by the beneficiaries, and is capped by the resources that they can commit. This approach provides several concurrent advantages.

            First, it enables to identify the priorities and fix the qualifying conditions through participation of all.  Admittedly, democratic participation in defining health risks may present some problems, notably access to medical knowledge, which will have to come from the outside  (Although  health workers (nurses, midwives, first-aiders etc.) have much health knowledge which should not be overlooked or denigrated). However, it is assumed that the community can define random catastrophic risks that are most feared by the members, as this category of risks will likely be established not solely on medical grounds. A different logic may perhaps apply in respect of setting the priorities for health promotion, preventive and primary care. Prevention is effective when accidents occur, as they represent a tangible signal that something more serious can happen, and people needy of first aid are sensitized to prevent the risk (ILO-OSHS, 1989). This approach may open up avenues for collaboration between the micro-insurance unit and agencies that deal with prevention or health promotion, who can provide technical expertise and  funding.

            Secondly, autonomous decision-making enables a group of people to act as a cohesive social unit that can fulfill a role no-one else can do better: relate needs and priorities to their prevalent activity, location-specific conditions, the level of resources etc.

            Thirdly, retaining decision-making powers within the community empowers it to control the flow of its funds. Two examples can illustrate this point: the micro-insurance unit can change the benefit package rapidly, without the need to receive approval from the outside. And, the micro-insurance unit can represent its members in negotiations with interlocutors outside, notably with providers.

            Because micro-insurance services the poor, each unit must be particularly sensitive to keeping overheads low. All known alternatives to self-management, be they commercial or public, favour highly trained and expensive professional management, as is typical in the formal health sector. Replicating such a managerial profile seems  incompatible not only with the sizes and quantities that need to be managed at the level of a single community, but also with the kind of services that will be provided. Self-management is thus not only a cost saving but also a right-sizing measure.

            Self-management also enhances transparency (and hence reduces the risk of corruption) and cohesion around social objectives (and hence creates a climate discouraging abuse of the system). Reducing anonymity raises self control. Also, closeness between the members creates intangible but real links that augment acceptance of redistribution of benefits according to need rather than solely according to individual utility.

             One of the innovative features of micro-insurance is that it introduces a complex financial concept, health insurance, as an extension to familiar social interactions. This opens the door to improved contacts between excluded populations and financial markets, who frequently ignore the needs of poor and rural populations, on the assumption that they are unable to pay for health services. Also, poor and rural people do not have the habit or the possibility to initiate individual negotiations to tailor insurance products to their conditions. The community, if helped and empowered, can serve both as an organizer of local self-help as an alternative to the unavailable public help, and as a mediator between its membership and actors outside the community, such as public authorities (including providers of health services) or financial institutions.

            Setting up micro-insurance units  require a relatively small nucleus of people and very little capital or infrastructure to start operations. There is no impediment to enlarging the group over time both within the community and across communities. And the triple autonomy, whereby each micro-insurance unit defines its own insurable risk,  organises financing of the insurance, and exercises control over the flow and management of its funds provides a very potent formula.

            For as long as micro-insurance succeeds in maintaining popular participation,  it is shielded from the ambivalent pressure afflicting many voluntary health insurance schemes of on the one hand having to attract clients through financial incentives, and on the other hand dampening client satisfaction by restricting the demand side. Democratic governance also helps deal with the streamlining community perception of equity, a concept often interpreted freely. The risk of insufficient consensus on equity may lead some people to consider rationing as arbitrary. The debate thus shifts from process of distribution, to a debate over the rules governing distribution. The linkage at the grass-roots level between rule setting and distributive decisions is the community’s way of reducing the risk of alienation of its individuals from its institutions.

            The concept presented above is based on theory and on some experience. A recent synthesis of 22 of case studies (out of 50 test communities) has this to say about the viability of the concept (Atim, 1998, p. xii) :

... Even now, they make a significant contribution to health care access and extending social protection to disadvantaged sections of the population by mainly targeting people in the informal and rural sectors. This also represents a contribution to equity in health care in the areas where they are active. Another area in which MHOs make a new – and in this case, original – contribution is that of democratic governance in the health sector… [and]  in representing their communities or members before the health authorities in order to articulate the views of the consumers of health care. This gives them some weight in influencing the priorities, resource allocation decisions, and responsiveness of the health authorities to the concerns of the public on such issues as waiting times, staff behaviour, quality of services etc. This is a genuinely new contribution which reflects the role and origins of the MHOs as part of the growing and confident civic society that began to develop in Africa in the 1990s.

In short, the local level, the personal acquaintance of the membership with each other, the transparency of decisions, the autonomous low-cost management and the non-profit character of micro-insurance units are all amplifiers of social cohesion. The reliance of micro-insurance on the dynamics of social cohesion within the community is more than tactical; it is the strategy to seek acceptance of the system.

Micro-Insurance and classical insurance problems

            Health insurance has had to struggle with some classical insurance problems, such as moral hazard, free riding, adverse selection, under-insurance or information asymmetry. Having said earlier that micro-insurance is first and foremost an insurance concept, we shall now  examine its exposure to the same problems, and what responses it can give.

            Moral Hazard.  “Moral hazard” designates price sensitivity of elastic demand, and price indifference of inelastic demand. Moral hazard originates mainly from three phenomena: demand-induced over-utilisation, supply-induced over-utilisation, and system-induced mechanisms linked to insurer behaviour (including premium calculations) that encourage the first two causes.  In discussing this phenomenon, one must bear in mind two features: the risk that is covered and the information available to the insurer over the insured.  If health insurers were free to respond only to risk aversion, they might limit the benefit package to a few random,  inelastic, high-impact but low-frequency risks. However, health insurance schemes operate under imperfect conditions (due notably to difficulties to identify those risks across a large and heterogeneous population, or due to public regulation of the insurance or because of imperfect competition between different benefit packages) with the consequence that insurers do not limit their activity only to random catastrophic risks.

Micro-insurance, for the time being, is exposed neither to a competitive market nor to regulations constraining the definition of the insurable risks. The members determine the insurable risk in light of the benefits that the community can autonomously pay for. Furthermore, the deterrent effect of a possible expulsion from the group of self-interested individuals as punishment for over-utilization shields micro-insurance somewhat from over-utilization responding to unilateral utility, and enhances  multilateral utility. The balance between individual and group utilities is imprecise and delicate, but it nevertheless influences individual behaviour for as long as  the community as a whole is a major stakeholder. Bearing in mind that small communities can establish with relative ease a consensus on what constitutes inelastic utilization, and that utilization information circulates within the community,  micro-insurance seems to be less exposed to demand-induced moral hazard, at least for high-cost events. Nor does it produce system-induced moral hazard generated mainly from its provider payment and premium determination systems. It may even have some built-in checks and balances to avert this problem which are not purely economic but social as well. As for supply-induced moral hazard, micro-insurance is not entirely shielded, and the risk may increase along with the increase in curative care.

            “Free Riding”  The issue of free-riding is inherently linked to imperfect information on user behaviour. Private health insurers have tried to tackle this by excluding from the insurance package some risks. While this remedy may reduce exposure to free-riding, such a policy has been viewed as one of the main market failures of private insurance that gave rise to government provision of social insurance. Due to its democratic governance, each micro-insurance unit  must consider members’ decisions on the insured risks, but this may not provide sufficient protection against free-riding. Free-riding can mean that self-interested individuals would wish to enjoy risk-sharing arrangements without reciprocity. If micro-insurance units could enforce on their members a commitment to reciprocate in the future for instant help, there would be no fear of free riding; however, in the absence of such a possibility, free riding can be reduced by imposing a waiting period, during which new affiliates would not be covered or only for random risks. This waiting period would facilitate constituting some modest reserves, and it would reduce the risk of defection after instant drawing of benefits. The only tool that micro-insurance can exploit would be community interaction and flow of information. This may not be a guarantee for success, but without community involvement there seems very little chance to find a solution. Also, free riding can exist so long as the free-rider can hide his or her advantage from others, and particularly from the insurer. We believe that this is more difficult in small communities.

            “Adverse selection” occurs when people with a high probability of “health loss” predominate in the membership, while those with low probability of loss do not join. It is undeniable that some types of occupations may entail exposure to higher risk of injury, or that some persons may live in conditions or regions that are more prone to specific types of diseases. The least expensive way to minimize this problem is to insure groups that are selected on a characteristic other than health status, thus randomizing the health risk. Micro-insurance units may be exposed to some risk of adverse selection particularly at start-up of operations, when the most insurance aware individuals within the community might be particularly vulnerable. However, micro-insurance cannot consider solutions which are employed by commercial insurers, such as “risk rating” or personalizing deductibles according to “experience rating”. Such solutions are based on statistical information that may be expensive to obtain, and unadapted both to the specific conditions of the clientele base and to the specific composition of the benefit package. At the same time, the group has a qualitative advantage in the form of  acquaintance with the client base, who shields micro-insurance units from “adverse selection” arising from incomplete or inaccurate disclosure of information by the applicant, thereby leading the insurer to under-assess the insured risk.

            “Under-insurance” represents the choice of a person to buy less insurance than needed. If the person bears the consequences of such a decision, there is no risk to anyone else (either other individuals or the insurance). However, if someone else is obliged to pay for uninsured services, then this situation becomes somewhat similar to “free riding”. Under-insurance can be avoided if the insurance is not obliged to pay for anyone who is uninsured. The communities where micro-insurance would be implemented have two potent tools to validate this rule: first, they are composed of poor people, who cannot be expected to subsidize others; and secondly, the democratic nature of the micro-insurance unit gives social legitimacy in implementing even seemingly tough rules. Another point that may reduce the likelihood for under-insurance is to offer only one type of policy, putting the client before an “all-or-nothing” choice. Micro-insurance is quite likely to practice this policy, because of its ease of administration and its more coherent synchronization with its collective character.

            “Information asymmetry”   Several forms of information asymmetry exist. Firstly, does  the insured have sufficient information to decided whether or not to buy insurance? Many people may simply not be educated enough to understand the notion of risk or their decisions may be based on a feeling about how exposed they are to certain risks, or that the premium is too expensive. A second phenomenon, labelled demand-side information asymmetry, refers to  client ignorance and provider knowledge of the care needed in a specific situation. Provider choices are assumed to respond to concerns about practising good medicine, retaining patient satisfaction and perpetuating high income. All three elements are usually linked to increased utilisation. Consequently, in private health insurance demand-side information asymmetry is countered by measures to limit provider-generated cost escalation. Micro-insurance has no specific shielding from this, other than in the limited coverage it assumes.

Information asymmetry may also occur on the supply side. Supply-side asymmetrical information means that the provider of care does not have sufficient information on the outcome of treatment and engages in defensive or trial-and-error treatment which increases the cost of care. It may also occur when the provider does not have some vital information about the patient which the patient prefers to withhold.

Micro-insurance units can reduce their exposure to information asymmetry mainly through the control of the benefit package, and through their role as negotiators with providers over prices of services.  

            “Cost escalation” occurs when either the demand side or the supply side seek to enlarge the quantity or technological sophistication of treatment more than required. This is unlikely to occur at the initial stages of incorporation of micro-insurance units but can not be discounted completely. Clients who have little to lose from cost escalation are not likely to protest. In some developing countries, insurances add “loading costs” to the regular cost of care, as a way to shift back to the customers the added charges for collection of premiums, high bank charges, high costs of credit and perhaps also the difficulty to obtain reinsurance. Micro-insurance can reduce this risk first by reducing overhead costs to the minimum through self management, by tariff negotiations with providers, and by structuring the benefit package around items that are less sensitive to individual choice, and if necessary, reduce insurable risk.

            “Risk pooling” describes aggregation by an insurer of similar but unrelated risks, which serve the function of statistical spreading over a larger number of cases. While it is frequent to speak of risk pooling in private insurance, it would be more accurate to refer to clustering of classes of risk. The difference between risk pooling and clustering of risks is that whereas the former refers to pooling of individuals at random, the latter refers to a selective composition of the pool. In many countries, for-profit insurers can  refuse to offer coverage at will. Hence, they can categorize individuals according to the

risk that they represent, and refuse “bad risks”; this is called “risk skimming”. While the reservoir of “good risks” is ample, restricting access to “bad risks” does not hurt the insurer. But if there is scarcity of good risks, insurer and client alike are hurt by this practice. The size of the group is not particularly important in clustered risks, but may be very significant in random risk-pooling, since, it is thought, the larger the group of individuals, the better the spread of the idiosyncratic risk.  Classical insurance sorts people across multiple heterogeneous communities who represent a similar though unrelated risk and who are otherwise not bound to each other by intrinsic characteristics. Micro-insurance clusters homogeneous communities, joined by their shared willingness to bear some health contingencies together. A micro-insurance unit may be exposed to very serious financial risk if many in the group are struck by an epidemic-like occurrence. Such an occurrence is statistically more probable among small and homogeneous groups. To avert or reduce this risk, the micro-insurance unit might have to ration the benefits. However, too strict  rationing may dissuade from  joining micro-insurance. In the experimental phase, some ad-hoc help may be needed to deal with particular public-health issues. Longer-term solutions might be to create federative or similar structures of micro-insurance units, and to establish some form of re-insurance. This has not yet been tried. In theory, there is no reason to prevent pooling of  risks and resources between several micro-insurance organs. However, it is usually assumed that as the size of the risk sharing group grows, the information about new members tends to decrease. This might require more monitoring against moral hazard. Considering that moral hazard, much more than idiosyncratic risk, has been targeted for repression in developed countries, the small size of the target group cannot be considered as critically biassing the risk for micro-insurance.

Implementing micro-insurance

            This section is based on lessons learnt from sporadic experiences with mutual health societies or similar self-help initiatives launched by several bodies and in several countries.  The objective of all pilot units was to encourage the organisation of communal self-help in health services. However, since the various tests were initiated separately, they did not follow a uniform conceptual framework. The initiative to move from sporadic pilot projects to promoting the concept of micro-insurance and demonstrating its potential was instigated by the ILO, in alliance with other stakeholders and has been focussed on Central and West Africa, starting in 1996.

Sporadic activity has brought some relief to needy individuals, and as such is welcome. The problem with such activity is however the difficulty to acquire sufficient information that could lead to systematic replication and better implementation. These questions relate to the characteristics of the excluded and the reasons for their exclusion; the different concepts of risk; how to enhance the social capital in order to foster willingness to pool risks and resources; how to feed-back information on success stories in terms of improved protection, member satisfaction, improved health status etc.

            ILO’s implementation strategy is designed to address these questions. It captures the need  to back-stop each micro-insurance unit by way of providing a channel for the transfer of information between units, research and monitoring infrastructure, training facilities, and , not least, a platform for public and political advocacy for this instrument. The strategy is described in the graph below:

            The first experience with this strategy comes from ILO’s activity in Central and West Africa. In June 1998, a high-level workshop was convened in Abidjan (Côte d’Ivoire) to present to many participating schemes the implementation platform and the support structures that are being set up. The meeting has adopted a document summarizing the essential elements, called  The Abidjan Platform. The Abidjan Platform contains the following salient points:

The democratic, voluntary, autonomous, participatory, community-based and non-profit nature of micro-insurance was recognized.

The bottom-up approach and the need to develop local expertise for sustainable development of micro-insurance institutions was affirmed.

The importance of “learning by doing”, field experimentation and evaluations and impact assessments as the basis to improve know-how, action and research were stressed.

It emphasised the importance of empowering the membership:

            1.    to develop more balanced relations with the authorities;

            2.  to develop contractual relations between micro-insurance institutions and health providers.

Recommended the dissemination of best practices, cross-fertilisation and coordination between stakeholders and the on-going improvement of appropriate tools.

It advocated faster coverage of population groups under micro-insurance by establishing links with  social movements.