«Overcoming Barriers to Microinsurance Adoption: Evidence from the Field† Shawn Cole Harvard Business School, Boston, MA 02163, U.S.A. E-mail: ...»
The Geneva Papers, 2015, (1–21)
© 2015 The International Association for the Study of Insurance Economics 1018-5895/15
Overcoming Barriers to Microinsurance Adoption:
Evidence from the Field†
Harvard Business School, Boston, MA 02163, U.S.A.
This paper provides an overview of the academic literature on microinsurance adoption
in emerging markets, with a particular emphasis on randomised control trials. I discuss what we know, what we can reasonably hope to know using the extensive work on microcredit as a comparator, and what the available evidence implies for public policy. Particular attention is paid to the case for a greater role for the government in supporting the development of microinsurance.
The Geneva Papers (2015) 0, 1–21. doi:10.1057/gpp.2015.12 Keywords: microinsurance; developing countries; government advance online publication, 3 June 2015 Introduction In the next 20 years, more than a billion people will enter the formal ﬁnancial system.
Payment services will most likely be the initial draw, but savings, credit and insurance products will surely follow. While this transition to the formal economy appears inevitable, it is less clear how quickly these additional products will be adopted and the degree to which this expansion of ﬁnancial services will improve consumers’ welfare and ﬁrms’ bottom lines.
This paper explores the role of microinsurance in the domain of household ﬁnance around the world, with a particular emphasis on the barriers to adoption and the potential for public–private partnerships (PPPs) to reduce or eliminate these barriers. Much, though by no means all, of the case for public involvement in microinsurance rests on two broad arguments.
The ﬁrst is that insurance improves welfare: just as governments may support education, food aid, or other goods and services for the poor, so too might they provide insurance. The ﬁrst part of this essay, therefore, reviews recent ﬁeld work in developing countries that seeks to document the beneﬁcial value of insurance to the insured. Though this discussion will be ﬁrmly grounded in (and cite) the most recent academic literature, this paper is not intended as an exhaustive literature review (for an excellent recent review of the literature on the demand † This paper has been granted the 2015 Shin Research Excellence Award—a partnership between The Geneva Association and the International Insurance Society—for its academic quality and relevance, by decision of a panel of judges comprising both business and academic insurance specialists.
The Geneva Papers on Risk and Insurance—Issues and Practice 2 for microinsurance, see Eling et al.1 and the papers that follow in a special issue of this journal on microinsurance.2 These ﬁeld studies are expensive and time-consuming; in addition to reviewing the evidence base they create, I will discuss their strengths and limitations.
A second important argument supporting government intervention in insurance markets is the presence of classical market failures—adverse selection and moral hazard. Early economic models3 suggest that when the decision to purchase insurance is voluntary, and clients have private information about the risks they likely face, higher-risk clients are more likely to purchase insurance, resulting in a higher-cost risk pool. This, in turn, can drive up the premium insurance companies must charge and drive low-risk clients out of the market, and, in the extreme, could lead to a complete “market unravelling”. “Moral hazard” occurs when individuals take less care because they are insured; for example, driving less carefully after purchasing auto insurance.
The balance of this paper will examine, sequentially, the various barriers that may limit efﬁcient and widespread microinsurance markets, with particular attention to how government interventions may improve the functioning and reach of these markets. I will consider perspectives of clients, governments and the ﬁrms offering insurance products.
Taken together, the discussion will generate several lessons on the scope for the public and private sectors to work together to promote microinsurance around the world. To preview the
The evidence base provides at best a weak case for direct government expenditure on life ● or health microinsurance to reduce poverty. There is, to date, almost no high-quality academic evidence that life or health microinsurance products can improve welfare (nor any evidence that they reduce it); but there are also good reasons why it may be very difﬁcult to design a credible study that demonstrates insurance reduces poverty.
In contrast, robust evidence suggests production-related microinsurance can increase ● economic output. Several studies show agricultural index insurance can improve farmers’ planting decisions, causing them to plant riskier, higher-yield crops.
Reducing transaction costs is key, but the role of transaction costs in insurance adoption is ● rarely studied by academics. Robust mobile payment infrastructure, which in many cases may require government support, would greatly facilitate the provision of many types of insurance and hence lower transaction costs.
Financial education, while promising in theory, often fails in practice. While it may be ● painful for academics who teach ﬁnance for a living to admit it, there is little evidence that traditional ﬁnancial education programmes can improve ﬁnancial decision-making in a cost-effective manner.
The private sector has uniquely useful technical capacity. I discuss two case studies that ● suggest private insurance providers may be more creative in designing products to overcome frictions or distributional challenges that the government may face.
The (uncertain) value of insurance Simple theory Insurance is arguably one of the greatest inventions of humankind. Idiosyncratic risks (death or sickness of a breadwinner, ﬁre, etc.) can be devastating to human welfare. By pooling idiosyncratic risk, a well-functioning insurance product can effectively eliminate it.
Moreover, because formal insurance contracts create strong incentives for some parties (e.g. insurers) to reduce and mitigate risks, the creation of insurance systems can, in some cases, reduce risk events. For instance, ﬁre insurance companies may require homeowners to install and maintain smoke detectors.
This simple analysis—that individuals are averse to risk, and pooling risk can eliminate it— can quickly lead to the conclusion that insurance should be welfare enhancing. International aid organisations are increasingly focused on risk and the potential of insurance: the World Bank’s ﬂagship publication, the World Development Report, focused on risk in 2014,4 and the International Labour Organization (ILO) has dedicated a facility to microinsurance development since 2008.5 Much of the case for public support turns on the possibility that the private sector alone cannot achieve widespread penetration, particularly among poor populations, on a purely commercial basis.6 But, of course, the presumed beneﬁts must be weighed against the costs of providing insurance. These cost calculations should not include payouts (which are transfers) but rather transaction costs involved with issuing and servicing policies. In addition, we must consider any deadweight loss of taxation raised to fund government support of insurance, and the possibility that government intervention crowds out private insurance providers.
To explain why insurance markets sometimes fail to develop, the academic literature has long focused on information frictions, but many practitioners have worried more about the prosaic concern of transaction costs, which loom especially large for microinsurance products. Many insurance policies that are proﬁtable when offered to wealthy or middleclass consumers cannot simply be scaled down (say, by reducing both premium and payout amounts by 90 per cent) to reach the “base of the pyramid” customers, as the transaction costs associated with acquiring these customers, writing policies, maintaining records, and processing claims might well exceed the entire premium the company could expect to collect, or even the value consumers place on reducing risk.
And the messy reality of building the evidence base Much of the recent work on microinsurance has focused on consumer demand for such new products. Eling et al.1 compare the post-2000 academic literature on microinsurance demand with existing research on traditional insurance markets. Adopting a framework from Outreville,7 the authors identify four key factors (economic factors, 4 World Bank (2014).
5 ILO (2009).
6 Olaosebikan (2013), which examines the proﬁtability of life insurance in Nigeria, and Yao (2013), which examines the sustainability of health insurance in Pakistan.
7 Outreville (2013).
The Geneva Papers on Risk and Insurance—Issues and Practice 4 social and cultural factors, structural factors and personal/demographic factors) as determinants of demand, comparing the evidence to that for traditional insurance products. In particular, they propose further research on improving contract design, promoting trust, providing ﬁnancial education and understanding the role of informal risk-sharing networks.
Over the past decade or so, the ﬁeld of development economics has experienced a shift towards ﬁeld-based research, often involving randomised control trials (RCTs). RCTs typically divide a study population (households, villages, districts, etc.) into treatment and control groups, administering an intervention to the treatment group to understand the causal impact of this intervention (or treatment).
RCTs came to prominence because they solved problems associated with causal inference.
As an illustrative example, consider an attempt to determine the value of health insurance by comparing households that purchase health insurance to households that do not. If insurance has no effect, but sicker people expect more insurance payout and are therefore more likely to buy insurance policies, we would observe worse health outcomes among the population purchasing insurance. However, it would be clearly incorrect to conclude from this relationship that insurance degrades health.
While this shift towards RCTs is not without controversy,8 it has focused academic economists’ efforts closely on policies and products that can be offered in a targeted manner, such as insurance schemes. Because much of the ﬁnancial (and effort) cost of an evaluation comes from the study (i.e. hiring ﬁeld research staff, surveying households, etc.) rather than the intervention (e.g. varying subsidies for insurance premiums), academics have often taken the lead in experimentally varying product offerings or programme features to test speciﬁc theories.
The past 10 years have seen a spate of ﬁeld experiments designed to measure the value of microinsurance in the developing world. The Poverty Action Lab, a confederation of 100 academics around the world, currently lists more than 40 ongoing or completed studies related to a wide variety of insurance products: life and health insurance, to be sure, but also more exotic products such as weather index insurance9 and GDP insurance.10 The ILO, Microinsurance Network, International Cooperative and Mutual Insurance Federation and the Gesellschaft für Internationale Zusammenarbeit have been driving forces behind many of these (and other) studies.
I will discuss this emerging evidence base by product line.
Life insurance Life insurance is by far the most commercially successful microinsurance product, with an estimated 48 million people receiving coverage around the world in 2012.11 The majority of individuals covered receive a credit-life product, which reimburses a lending institution (typically a microﬁnance organisation) for the principal owed by a borrower who loses her life, and often pays to the borrower’s family an additional amount roughly equal to the amount of the loan. Because life insurance is often a compulsory add-on to a microﬁnance
loan, revealed preference arguments12 in favour of insurance to beneﬁt the consumer may be less compelling, though the fact that the lender requires them suggests that they do have value for the institutions.
Despite the commercial success, I am not aware of any academic studies that seek to measure the beneﬁts of life insurance to poor households. This is perhaps not too surprising, for a number of reasons: academics like to study new things, and arrangements for payment on death (e.g. life insurance, burial societies) are perhaps as old as the written record.
Moreover, most ﬁeld projects last between one and ﬁve years, while many studies seeking to measure, for example, the effects of payouts on consumption smoothing, would require quite large sample sizes and long study periods (and hence costly data collection efforts).
Health insurance There have been many attempts to offer micro health insurance, often through NGOs, and at least half a dozen ongoing or completed attempts study the impact of health insurance on the behaviour of the insured. An important challenge these studies face is the relatively immature nature of micro health insurance products, which may not have effective systems in place or face strong effective demand.
The title of one of the ﬁrst published studies on micro health insurance is revealing:
“Bundling health insurance and microﬁnance in India: There cannot be adverse selection if there is no demand”.13 Limited demand for the product, as well as administrative difﬁculties faced by clients attempting to make use of insurance, precluded meaningful impact evaluation.
In another RCT, Thornton et al.14 subsidised insurance premiums for the treatment group and found initially promising take-up rates (about 20 per cent). After one year, treated households had substituted towards health service providers covered by the insurance and reported slightly lower out-of-pocket health expenditures. However, total health costs plus premiums were higher in the treatment group, and health outcomes were no different. Most troubling, fewer than 10 per cent of those who subscribed in the ﬁrst year continued to purchase the subsidised policies.