Index Insurance For Developing Countries

Applied Economic Perspectives and Policy (2012) volume 34, number 3, pp. 391–427.
doi:10.1093/aepp/pps031
Featured Article
Index Insurance for Developing Countries
Mario J. Miranda* and Katie Farrin
Mario J. Miranda is the Andersons Professor of Finance and Risk Management

and Katie Farrin is a doctoral candidate, Department of Agricultural,
Environmental, and Development Economics, The Ohio State University.
*Correspondence may be sent to: miranda.4@osu.edu.
Commissioned April 2011; accepted July 2012.
Abstract Unlike conventional insurance, which indemnifies policyholders for veri-

fiable production losses arising from multiple perils, index insurance indemnifies pol-
icyholders based on the observed value of a specified “index” or some other closely
related variable that is highly correlated with losses. Index insurance exhibits lower
transaction costs than conventional insurance, potentially making it more affordable
to the poor in the developing world. However, it also offers less effective individual
risk protection. This article provides a review of recent theoretical and empirical
research on index insurance for developing countries and summarizes lessons learned
from index insurance projects implemented in the developing world since 2000.
JEL codes: O16, Q14, G22.
Introduction
Weather-related perils such as droughts, floods, freezes, and wind-
storms present pervasive risks for the billions of poor in the developing
world, particularly those whose livelihoods depend on agriculture. The
adverse impacts of these perils on the lives of the rural poor are aggra-
vated by the absence of formal insurance, credit, and deposit services in
rural areas. A lack of access to affordable financial services forces the poor
in developing countries to employ risk-avoidance, risk-diversification, and
informal risk-sharing practices that are either costly or offer inadequate
risk protection, particularly against widespread catastrophic weather
events (Coate and Ravallion 1993; Townsend 1994; Ligon, Thomas, and
Worrall 2002; Dubois, Jullien, and Magnac 2008).
Although the effects of catastrophic weather are felt most immediately
and most profoundly at the farm level, the effects are also propagated
through the agricultural marketing chain via contractual relationships
(Miranda and Gonzalez-Vega 2011). In particular, agricultural banks,
input suppliers, cooperatives, and processors who provide loans to or
enter into marketing contracts with farmers can experience dramatic
increases in loan delinquency and contract performance failures after a
catastrophic weather event that simultaneously affects a large number of
# The Author(s) 2012. Published by Oxford University Press, on behalf of Agricultural and Applied
Economics Association. All rights reserved. For permissions, please email:
journals.permissions@oup.com.
391
Downloaded from https://academic.oup.com/aepp/article-abstract/34/3/391/9031
by Stellenbosch University user
on 01 January 2018
Applied Economic Perspectives and Policy
clients. As such, uninsured catastrophic risk can severely limit the evolu-
tion of efficient agricultural credit markets and value chains, undermine
investment in productive farm activities, discourage the adoption of new,
more productive technologies, and generally impede the efforts to emerge
from poverty made by the rural poor of the developing world.
Developing countries have looked to developed countries for guidance
on ways to promote establishing sustainable private agricultural insurance
markets. However, virtually all insurance programs operating in the devel-
oped world have exhibited poor actuarial performance and have been sus-
tained only through significant government subsidies. These subsidies
usually take the form of government payment of the greater part of insur-
ance premiums on behalf of farmers, but can also include government
reimbursement of administrative costs borne by private insurers and
government-run reinsurance agreements that are actuarially favorable to
insurers. The indemnities paid by agricultural insurance programs in
developed countries typically substantially exceed the value of premiums
paid by agricultural producers, indicating that similar programs would
not be fiscally sustainable for governments of limited financial resources
in the developing world (Hazell 1992). For example, from 2003-2007, the
ratio of indemnities paid to producers to premiums collected from pro-
ducers for the five largest national agricultural insurance programs were:
USA, 1.70; Canada, 1.86; Spain, 2.44; Japan, 1.84; and Italy, 1.47 (Mahul
and Stutley 2010).
Most agricultural insurance programs currently offered by developed
nations are based on “conventional” insurance contracts in which an
insured party pays a premium and subsequently receives an indemnity
based on his verifiable losses. Conventional agricultural insurance,
however, possesses some well-known structural problems, including
moral hazard, adverse selection, and systemic risk. Moral hazard, also
known as the “hidden action” problem, arises when farmers, after pur-
chasing insurance, alter their production practices in a manner that
increases their chances of collecting an indemnity (Holmstrom 1979;
Shavell 1979; Chambers 1989; Horowitz and Lichtenberg 1993; Vercammen
and vanKooten 1994; Smith and Goodwin 1996). As a result of such
actions, the insurer’s expected indemnities rise, undermining the financial
soundness of the insurance operation. To combat moral hazard, the
insurer specifies “deductibles” in the insurance contract, which require the
farmer to absorb part of his losses, thus giving him an incentive to con-
tinue using risk-reducing production practices after the purchase of insur-
ance. The deductible payment provision, however, limits the coverage
provided by the insurance and reduces its value to the individual farmer.
Adverse selection, also known as the “hidden information” problem,
arises because farmers are better informed about the distribution of their
own production losses, and thus are better able to assess the actuarial fair-
ness of their premiums than the insurer (Akerlof 1970; Skees and Reed
1986; Quiggin, Karagiannis, and Stanton 1993; Just, Calvin, and Quiggin
1999; Makki and Somwaru 2001). Farmers who recognize that their
expected indemnities exceed their premiums are more likely to purchase
coverage than those whose premiums are actuarially high. As a result, the
insurer’s expected indemnity outlays exceed total premium income and,
in the long-run, the insurance operation loses money. Efforts by the
insurer to avoid losses by raising premiums only result in a smaller and
392
on 01 January 2018
more adversely selected pool of participants. Adverse selection problems

can be especially acute in developing countries, given that very few
farmers maintain adequate production records that could be used to accu-
rately rate a conventional agricultural insurance contract.
Yet another problem associated with conventional insurance when pro-
vided for agricultural production is systemic risk (Miranda and Glauber
1997). An ideal condition for the existence and efficient functioning of
insurance markets is that individual risks be independent. This condition
is violated in agricultural production, however, because the widespread
effects of droughts, freezes, floods, windstorms and other catastrophic
weather events cannot be fully diversified. An unprotected private insurer
offering agricultural insurance would ultimately assume a significant
amount of the systematic risk exhibited by farmer yield losses, putting it
in peril of bankruptcy were a widespread, catastrophic weather event to
occur. International reinsurance is generally available, but is expensive
and difficult to obtain due to the thinness of markets and variations in
commercial laws, business customs, and government regulatory environ-
ments across developing countries. The costs of reinsurance must be
passed on to the insured, thus increasing premium rates and reducing the
value of and demand for the insurance.
Given the problems associated with conventional forms of insurance in
agriculture, over the past twenty years a growing number of academic
researchers, multilateral international non-governmental organizations,
and national governments have exhibited considerable interest in the use
of a different form of insurance known as index insurance to manage the
risks faced by poor agricultural producers in the developing world
(Miranda and Vedenov 2001; Bryla and Syroka 2007).
Unlike conventional insurance, index insurance indemnifies the insured
based on the observed value of a specified “index” or some other closely
related variable. Ideally, an index is a random variable that is objectively
observable, reliably measurable, and highly correlated with the losses of
the insured, and additionally cannot be influenced by the actions of the
insured. The most widely used index in index insurance contract designs
is rainfall (Bardsley, Abey, and Davenport 1984). However, other indices
have been used or otherwise considered in index insurance designs,
including temperature, humidity, and El Niño-Southern Oscillation
indices. Indices that are not, strictly speaking, weather variables, but
which nonetheless serve as proxies for the impact of widespread weather
events on agricultural production have also been employed in index insur-
ance designs, including area yields, flood levels, river flows, satellite-
measured vegetation indices, and regional livestock mortality rates
(Halcrow 1949; Miranda 1991; Skees, Hartell, and Murphy 2007; Khalil,
Kwon, Lall, Miranda, and Skees 2007; Leiva and Skees 2008).
Index insurance contracts can be classified according to the underlying
index used to settle indemnities and the perils they are designed to cover.
Index insurance products can also be classified according to the nature of
the intended policyholder and their uses in risk management strategies.
“Micro” index insurance contracts are designed to be held by farmers to
help manage farm-level production risk. “Meso” index insurance contracts
are designed to be held by businesses that provide financial intermedia-
tion to large groups of farmers (such as banks, input suppliers, processors,
and cooperatives) who wish to protect the integrity of the loans or other
393
on 01 January 2018
financial and marketing arrangements that exist between them and their
farmer clients. Such institutions are often referred to as “risk aggregators”
because they systematically suffer from the impact of adverse weather on
their many farmer clients simultaneously. “Macro” index insurance con-
tracts are designed to be used primarily by national governments, regional
administrative agencies, and non-governmental organizations to finance
centrally managed relief efforts during widespread catastrophes.
To concretely illustrate the notion of an index insurance contract, con-
sider the following hypothetical example of a layered indemnity rainfall
micro-insurance contract designed to protect farmers against droughts: A
farmer, prior to a specified closing date of February 1, elects a liability
(maximum indemnity) of $50 and pays the insurer a $5 premium, com-
puted from the published premium rate of 10%. In return, the insurer
promises to pay the insured an indemnity that depends on the total rain-
fall measured at the specified nearby weather station during the months
of April, May, and June. For example, the contract might pay no indem-
nity if total rainfall were to exceed 550mm, but would pay $10 if total rain-
fall were between 500 and 550mm, $25 if total rainfall were between 450
and 500mm, and would pay the maximum liability of $50 if total rainfall
were less than 450mm. Other, equally simple forms of rainfall insurance
indemnity schedules, such as all-or-none indemnities or linearly propor-
tional indemnities are possible.
Index insurance is designed to avoid many of the problems that plague
conventional insurance. Because the insured cannot significantly influence
the value of the index through his own actions, and thus the indemnity
paid by the contract, index insurance is essentially free of moral hazard.
Because an index insurance contract’s premium rate is typically based on
publicly available information, not privately held information, index
insurance is largely free of adverse selection problems. Because index
insurance contracts are standardized, they do not require individually tail-
ored terms of indemnification or separate verification of individual loss
claims. As such, index insurance is less expensive to administer. And
because index insurance has simpler information requirements and exhib-
its greater uniformity and transparency of contracts, index insurance is
easier to reinsure. These features of index insurance can substantially
reduce its cost relative to conventional insurance, making it more afford-
able, particularly for poor farmers in the developing world (Skees
2008a,b).
Index insurance, however, suffers from the drawback known as “basis
risk,” which refers to the failure of index insurance to provide indemnities
that perfectly match the losses of the insured (Miranda 1991; Doherty and
Richter 2002). In particular, since the indemnity provided by index insur-
ance is based on an index rather than on verifiable losses, it is possible for
the insured to suffer a significant loss without the insurance contract pro-
viding an indemnity. The potential benefits of index insurance to the
insured ultimately depend on the statistical relationship between the
indemnities provided by the index insurance contract and the losses suf-
fered by the insured: the greater the correlation, the greater the potential
benefit.
The principal challenge in designing index insurance contracts is to
minimize basis risk while attempting to maintain contract transparency
and containing delivery, marketing, and reinsurance costs. Index
394
on 01 January 2018
insurance contracts are especially difficult to design in areas that exhibit

significant microclimatic variation, and, unless steps are taken, are suscep-
tible to intertemporal adverse selection arising from the ability of those
insured to forecast indemnities based on knowledge of observable short-
term climatic trends, such as the onset of El Niño conditions (Luo, Skees,
and Marchant 1994; Khalil, Kwon, Lall, Miranda, and Skees 2007).
Unfortunately, the results of many micro-index insurance pilot pro-
grams have been disappointing, with significant uptake of an index insur-
ance product occurring only if it is bundled with other benefits, such as
low-interest loans. Although high basis risk is often cited as one of the
causes of low demand for index insurance, few experimental and empiri-
cal assessments of basis risk and the risk-reduction benefits of and
willingness-to-pay for index insurance have been performed (Breustedt,
Bokusheva, and Heidelbach 2008). Considerable skepticism therefore
exists, and appears to be growing regarding the benefits of index insur-
ance offered directly to primary producers (Binswanger-Mkhize 2011). As
a result, researchers and donors are now beginning to look at alternative
uses of index insurance at the meso- or macro-level as a way of supporting
or promoting the development of financial institutions that in turn
provide valuable financial and risk management services to the rural poor
of the developing world (Miranda and Gonzalez-Vega 2011).
In the sections that follow, we review research and market development
activities in index insurance for the developing world. In the next section
we develop some simple models of decision-making under uncertainty
that incorporate the essential features of index insurance and thus allow
us to understand the issues associated with its provision. In the subse-
quent section, we outline five representative index insurance programs
that have been implemented in developing countries to varying degrees
during the early 2000s. Moreover, in the remaining sections, we summa-
rize lessons that have been learned from index insurance pilot programs
and academic research, concluding with general recommendations for the
future development of index insurance markets.
Simple Model of Index Insurance Demand

In the existing academic literature, the demand for and potential bene-
fits of index insurance are typically analyzed by employing a static
von-Neumann-Morgenstern expected utility framework (Gollier 2003).
Numerous articles on index insurance that employ this framework have
appeared in the academic literature (Miranda 1991; Mahul 1999a; Mahul
1999b; and Bourgeon and Chambers 2003).
To better understanding index insurance, consider the following simple
model. A farmer endowed with predetermined wealth w in this period
faces an uncertain income ỹ next period, the distribution of which
depends on an insurable factor or “index” z̃ and an uninsurable factor 1̃.
Specifically:
ỹ = m + bz̃ + 1̃ (1)
where m is expected income, b ≥ 0, Ez̃ = E1̃ = 0, and z̃ and 1̃ are inde-

pendent. Further assume, without loss of generality, that Var(z̃) = 1, so
395
on 01 January 2018
that b = Cov(ỹ, z̃) measures the sensitivity of the farmer’s income to varia-
tions in the systemic factor.
Also assume that the farmer is not permitted to borrow or save, but
may purchase an unlimited, fractionally divisible number of insurance
contracts x ≥ 0 that provide coverage against adverse realizations of the
index. More specifically, a unit index insurance contract purchased today
at the premium rate p, tomorrow provides an indemnity h(z̃) ≥ 0 that
depends on the realization of the index z̃. We assume that the indemnity
function h is continuous and non-increasing with h(z) ¼ 0 for z ≥ 0; that is,
the insurance indemnifies only values of the index that reduce income.
The farmer maximizes the sum of current and discounted expected utility
next period; that is, he solves:
max u0 (w − px) + dEu1 (m + bz̃ + 1̃ + xh(z̃)). (2)

x≥0
Here, d , 1 is the farmer’s subjective discount factor and u0 and u1 are the
farmer’s current and future utility of consumption, both of which are
assumed to be twice continuously differentiable, strictly increasing, and
strictly concave.
Appealing to the Karush-Kuhn-Tucker Theorem, it is straightforward
to show that the farmer will purchase index insurance coverage if, and
only if:
p , p∗ ; El̃(z̃, 1̃)h(z̃) (3)
where
u′1 (m + bz + 1)
l̃(z, 1) ; d . (4)
u′0 (w)
In other words, the farmer purchases index insurance if, and only if, the
premium rate p is less than the “risk-adjusted” expected indemnity p*,
defined as the expectation of the indemnity h(z̃) weighted by the marginal
rate of intertemporal substitution of consumption l(z̃, 1̃). The risk-adjusted
expected indemnity p* may be greater than or less than the expected
indemnity Eh(z̃), depending on the curvature of the utility function and
the farmer’s wealth w. On the one hand, a sufficiently poor farmer (low w)
will not purchase index insurance, even if it is actuarially favorable, that
is, even if the expected indemnity exceeds the premium. On the other
hand, a sufficiently rich farmer (high w) will purchase index insurance
even if it is actuarially unfavorable, that is, even if the premium exceeds
the expected indemnity. However, the demand for index insurance at
higher levels of wealth may be due less to its risk-reduction benefits than
to the fact that investment in insurance provides the only means to store
wealth from one period to the next.
For p , p*, the farmer will purchase index insurance coverage x . 0
such that the value of consumption forgone by purchasing one unit of
coverage today equals the expected present value of consumption pro-
vided by the additional indemnity tomorrow. Using this fact, one may
totally differentiate the first-order optimality condition to show that, for
p , p*, the optimal index insurance coverage x is a continuously differen-
tiable strictly decreasing function of the premium rate p, a continuously
396
on 01 January 2018
differentiable, strictly increasing function of the income-index covariance

b, and a continuously differentiable, strictly increasing function of
wealth w.
Dynamic models of the demand for index insurance that allow for
savings and borrowing expand upon the static model to allow for self-
insurance. In such models, consumption can be smoothed over multiple
periods through saving and dissaving, as well as insurance. Access to
saving and credit reduces demand for and welfare gains from all forms of
insurance, including index insurance (Gollier 2003). In a developing
country context, where transaction costs and premium rates associated
with index insurance might be quite high, this implies that households
will prefer to self-insure. In other words, when credit and savings are
available, the demand for index insurance is less than what is predicted
by static von Neumann-Morgenstern expected utility theory.
To illustrate our theoretical understanding of the interactions between
savings and index insurance, consider the following extension to our two-
period model: An infinitely-lived farmer begins each period endowed
with predetermined wealth w, which he must then allocate among con-
sumption, savings, and purchases of insurance. The farmer faces an uncer-
tain income ỹ ≥ 0 the following period, such as that described in equation
(1). As in the two-period model, the farmer may purchase an unlimited,
fractionally divisible number of insurance contracts x ≥ 0 that provide cov-
erage against realizations of the index. More specifically, a unit index
insurance contract purchased today at the premium rate p, tomorrow pro-
vides an indemnity h(z̃) ≥ 0 that depends on the realization of the index z̃.
However, the farmer now may also save as much of his wealth s ≥ 0 as he
pleases, earning a per-period interest rate r.
The farmer chooses the savings s and coverage x that maximize the sum
of current and discounted expected future utility of consumption over an
infinite horizon. By Bellman’s Principle of Optimality, the farmer’s value
function, V(w), which denotes the maximum attainable sum of current
and discounted expected future utility of consumption given the farmer’s
current wealth w, is characterized by the functional equation:
V(w) = max {u(w − s − px) + dEV(ỹ + (1 + r)s + xh(z̃))}. (5)

s≥0,x≥0
Here, d is the farmer’s subjective discount rate and u is the farmer’s utility
of consumption, which is presumed to be twice continuously differentia-
ble, strictly increasing, and strictly concave. The farmer’s consumption
equals his predetermined wealth, less savings, less index insurance
premium payments.
Under mild regularity conditions, it is possible to show that the
farmer’s Bellman functional equation possesses a unique solution V that is
continuous, strictly increasing, and strictly concave. However, neither the
value function V nor the attendant optimal savings policy s(w) and insur-
ance coverage policy x(w) possess closed-form solutions. These functions,
however, can easily be approximated to an arbitrary precision using
numerical methods (Miranda and Fackler 2002).
Figures 1 – 3 illustrate the basic features of optimal index insurance
demand, with and without access to savings, for the infinitely lived
farmer. For the purposes of this illustration, we assume that the farmer’s
397
on 01 January 2018
Figure 1 Optimal index insurance coverage as a function of wealth
utility of consumption function exhibits constant relative risk aversion 2;

the discount factor is d ¼ 0.9; the interest rate on savings is r ¼ 5%; a sys-
temic event identified with the index z̃ occurs with probability 0.2 and
depresses income by 20%; the correlation between the farmer’s income
and the index is 0.7; and the index insurance premium equals 110% of the
expected indemnity.
Figure 1 illustrates optimal index insurance coverage as a function of
wealth, without access to savings (red curve) and with access to savings
(blue curve). As seen in figure 1, a farmer with a very low level of wealth,
with or without access to savings, will not purchase index insurance
because the value of spending one dollar on consumption today exceeds
the value of spending the indemnity generated by the index insurance on
consumption tomorrow, if the dollar is used to purchase index insurance
today. However, at levels of wealth above the critical level w1, it becomes
optimal to purchase index insurance. Above this critical level, the quantity
of index insurance purchased rises with wealth, as with the two-period
model. However, beyond a second critical level of wealth w2, an agent
with access to savings saves ever greater amounts, simultaneously reduc-
ing his demand for index insurance, eventually reaching a level of wealth
w3 beyond which he no longer demands index insurance. In short, in a
dynamic setting, the ability to save reduces the demand for index insur-
ance, and coverage will be demanded by neither the very poor, who
cannot afford the insurance, nor the very rich, who prefer to self-insure
through savings.
Figures 2 and 3 demonstrate other features of the demand for index
insurance in a dynamic world, both with and without access to savings. In
both cases, mean long-run purchases of index insurance coverage are seen
to be lower with access to savings than without. Figure 2 shows that the
mean long-run purchase of index insurance is directly related to the corre-
lation between the farmer’s income and the index. Figure 3 shows that the
mean long-run demand for index insurance in the presence of savings can
be quite sensitive to the interest rate on savings. Higher interest rates
398
on 01 January 2018
Figure 2 Mean long-run index insurance coverage as a function of the index/income

correlation
Figure 3 Mean long-run index insurance coverage as a function of the interest rate on savings
make savings more attractive, leading to higher buffer stocks of savings

and reducing the need for index insurance. Even when the interest rate is
negative, indicating loss of value from savings, as when storing a partially
perishable food commodity, the demand for index insurance is reduced
by access to savings.
Index Insurance Pilot Programs

Since the late 1990s, index insurance feasibility studies and pilot projects
have been undertaken in a wide variety of settings throughout the
developing world, including Bangladesh, Burkina Faso, China, Ethiopia,
399
on 01 January 2018
Guatemala, Honduras, India, Indonesia, Jamaica, Kenya, Malawi, Mexico,

Mongolia, Morocco, Nicaragua, Peru, Senegal, Thailand, and Vietnam.
These activities have been undertaken and/or financially supported by
organizations as diverse as the World Bank, the United States Agency for
International Development (USAID), Asian Development Bank (ADB),
Inter-American Development Bank (IADB), International Livestock
Research Institute (ILRI), the United Kingdom Department for
International Development (DfiD), and United Nations World Food
Program (UNWFP), among many others.
Activities undertaken in these feasibility studies and pilot projects have
been broad in scope, and have included: designing and actuarially rating
index insurance contracts; educating farmers, lenders, insurers, govern-
ment officials, regulators, and academics regarding the potential beneficial
uses of index insurance; acquiring, validating, and statistically analyzing
weather and agricultural production data; assessing the adequacy of
weather station network density, security, and real-time reporting capabil-
ities; evaluating the feasibility of insurance indices other than rainfall and
developing new technologies for measuring them; forging institutional
relationships among farmers, lenders, insurers, reinsurers, government
agencies, and non-governmental organizations; identifying and testing
alternative ways to incorporate index insurance into farm, firm, and gov-
ernmental risk management strategies; improving insurance marketing
and delivery mechanisms; and developing insurance training institutes
and information agencies.
Through these extensive research and market development activities,
many valuable lessons have been learned. The number of pilot programs
is too large to permit detailed summaries of all of them. We have,
however, selected five countries in which index insurance projects have
been undertaken for more detailed discussion. The projects were selected
based on several criteria. First, all projects reviewed evolved past the stage
of conceptualization, implying that a contract was designed and rated, a
marketing framework was developed, and the insurance product was
“marketed,” leading to some uptake of the product, however modest.
Second, the projects reviewed provided important general lessons learned,
whether positive or negative, regarding proper index insurance contract
design, marketing, financing, regulation, and education. Third, the projects
reviewed exhibit geographical and structural diversity; specifically, the
included projects represent Africa, Asia, and Latin America, involve index
insurance contracts based on rainfall and non-rainfall variables, and
provide examples of micro, meso, and macro-insurance products.
India
Since 2000, the Indian government has required insurers to provide
access to insurance for the poor, regardless of whether selling such poli-
cies is profitable (Mechler, Linnerooth-Bayer, and Peppiat 2006). Indian
insurers often operate at a loss in the low-income market, using profits
from policy sales in wealthier segments of society to cross-subsidize their
obligatory microinsurance programs. As a consequence, by 2007 over
500,000 Indian farmers were covered by some form of index insurance
(Mosley 2009). In addition, many Indian states make the purchase of insur-
ance compulsory when taking out an agricultural loan (Giné 2009).
400
on 01 January 2018
The BASIX weather insurance scheme, which was initially piloted in

1999 and became operative in 2003, is considered to be the first major agri-
cultural index insurance program in a developing country (Mosley 2009).
In the pilot phase, groundnut and castor farmers could voluntarily pur-
chase coverage at a premium of 3% of the value insured, with payouts
triggered when cumulative rainfall for the crop season fell below the
30-year average (Mechler, Linnerooth-Bayer, and Peppiat 2006). BASIX, a
microfinance institution (MFI) based in Andhra Pradesh, in southeastern
India, worked with rural bank KBS, the World Bank, insurer ICICI
Lombard, and global reinsurer Swiss Re to sell 194 policies in 17 villages
in the first year of sales; as the program expanded, so did contract sales,
with 318 contracts sold in 43 villages in 2004, 3,214 contracts sold in 422
villages in 2005, and 6,039 contracts sold in 538 villages in 2006 (Giné
2009). Marketing was targeted to small and medium farmers in villages
located near rainfall stations and containing large numbers of existing
BASIX clients. After village selection, BASIX representatives would
explain the product to a village leader, who was then to act as an
informed motivator to other villagers, before a general meeting was held
to educate potential policyholders and sell contracts (Giné, Townsend,
and Vickery 2008).
By 2010, BASIX insured almost 40,000 farm households. While the
cumulative average loss ratio for the program through 2010 was 145%,
indemnity payouts exceed premiums collected in only two of the seven
seasons for which data are available. The loss ratios were 144% and 447%
in 2006 and 2009, respectively (Joshi 2010). The index insurance product is
designed to protect the insured against catastrophic weather events; in one
analysis, contracts were found to have positive returns in only 11% of
sales periods, while the maximum rate of return was around 900% (Giné,
Townsend, and Vickery 2007).
BASIX, however, ceased index insurance sales at the close of the
2009-2010 season. Although the coverage area had been expanding, the
number of sales per village stagnated, with fewer than 10% average
uptake in villages where the product is offered. In addition, Giné (2009)
notes that purchasing households often insured only a small part of their
total agricultural income or chose to purchase insurance in only one or
two phases of the three-phase contract offering (where phases correspond
to sowing, cultivating, and harvesting periods), even when coverage in all
three phases was relatively inexpensive. This indicates that early adopters
of the product may be have been experimenting. In fact, Giné, Townsend,
and Vickery (2008) find that relatively risk averse rural households tended
to be less likely to purchase index insurance from BASIX. In this respect,
the BASIX experience highlights a significant obstacle for micro-index
insurance uptake: a potential lack of understanding of the product among
clients.
Finally, technical transformation of agricultural production practices is a
goal of index insurance researchers, who aim to design products that
promote the adoption of higher income, but perhaps riskier technologies.
The pilot phase of the BASIX program targeted cash crops that are more
profitable but more drought prone than traditional food crops, precisely
for this reason (Giné, Townsend, and Vickery (2008). The pilot from 2005
onward was modified to include area-specific contracts for all crops
(Mechler, Linnerooth-Bayer, and Peppiat 2006). However, adoption of
401
on 01 January 2018
higher-yielding crops has not been observed among purchasers of the

insurance (Mosley 2009).
Peru
An El Niño is a disruption of the ocean-atmosphere system of the
Tropical Pacific that occurs irregularly at intervals of 2-7 years, with a
typical duration of 12-18 months. Major El Niños vary in intensity, and
can have a profound impact on coastal Peru, which is located at the
eastern extreme of the Tropical Pacific. Intense El Niño events occurred in
1983 and 1997. During these events, Piura, a province located on the
northwest coast of Peru, the country’s most important agricultural region,
experienced rainfall that exceeded 40 times normal. The rainfall caused
catastrophic floods that destroyed crops and infrastructure and led to
widespread defaults on agricultural loans.
The risk of El Niño events significantly constrains access to agricultural
credit in northwest Peru. In an attempt to improve credit availability in
rural areas, a USAID project was initiated in 2004 to develop an index
insurance contract based on the El Niño-Southern Oscillation (ENSO)
index. The ENSO index measures sea surface temperatures off the coast of
Peru, where higher temperatures are associated with severe El Niño
events and subsequent flooding. Khalil, Kwon, Lall, Miranda, and Skees
(2007), in a preliminary analysis for the design of an ENSO index insur-
ance product for purchase by banks or MFIs in the Piura region of Peru,
find the ENSO 1.2 index to be well correlated with rainfall; additionally,
they find the ENSO data are more reliable, with a 150-year series of
surface temperature data available from National Oceanographic and
Administrative Administration (NOAA), whereas rainfall records from
Peru are sparse, short and intermittent, with the accompanying concern
among international reinsurers that rainfall data can be manipulated by
the local government agencies in charge of their collection.
The proposed insurance product was approved by the Peruvian
banking and insurance regulator, but due to a government subsidization
of traditional agricultural insurance in 2006, the index insurance project
went on hiatus. With additional work supported by the Gates Foundation,
partnership with local insurer La Positiva, and reinsurance secured from
PartnerRE, marketing and sales for the first season was scheduled to
begin in December 2009 and to close in March 2010 (GlobalAgRisk, Inc.
2010); however, although two MFIs that serve the poor in Piura, Caja
Piura and Caja Sullana, had indicated an interest in the product during
the development stages, no policies were sold.
Currently, La Positiva offers ENSO coverage as “Business Interruption”
insurance (Skees and Murphy 2009). The contract, for which La Positiva is
in the process of completing its first sale to a financial institution, triggers
when the surface temperature of the water in the ENSO 1.2 zone reaches
levels of 24 degrees Celsius or above; such temperatures are indicative of a
severe El Niño event, in which catastrophic flooding is more likely to
occur in the region. Caja Nuestra Gente (CNG), a financial institution
based in Trujillo that offers agricultural loans to small and medium enter-
prizes, is working with La Positiva to revise and sign a final contract with
approximately $2 million in coverage. The deal, which was delayed after
La Positiva failed to receive complete approval from Peru’s financial
402
on 01 January 2018
services regulator, is expected to open in January of 2012, with the rele-

vant index realizations occurring in November and December of 2012. A
unique feature of the contract is that indemnities are paid by the contract
before losses due to catastrophic flooding would be incurred; specifically,
it has been determined that average ENSO 1.2 measures for the months of
November and December predict extreme flooding events in February
through April with a high confidence level (GlobalAgRisk, Inc. 2010;
Lagos, Silva, Nickl, and Mosquera 2008).
Even if ENSO contracts are successfully sold to MFIs, it remains to be
seen how MFIs will use indemnity payments, if received. According to
GlobalAgRisk, Inc. (2006), possible uses include use of payments to: i)
offset the real cost of defaults from the bank’s portfolio; ii) reduce the out-
standing debts for individuals based on an assessment of their losses and
ability to repay; and iii) reduce interest rates for consumption loans fol-
lowing the disaster, particularly for farmers who had no access to credit
because they were in flood-prone areas. While the first and third options
do nothing to reduce the portfolio-at-risk rate, the second results in
increased transaction costs for the MFI, thus cancelling out some of the
cost advantages of using an index insurance product.
While the current premium of 6.5% for CNG is unsubsidized, analysts
from the Peruvian Studies Institute (IEP) confirm that this rate was
reduced from 10% at the beginning of talks. In addition, other Peruvian
financial institutions in Piura and Sullana expressed that they were unin-
terested in purchasing ENSO insurance at a rate of 10%. The IEP, whose
role in the insurance scheme is one of monitoring and evaluation, con-
ducted talks with additional MFIs in January 2012 in order to gauge
whether a lower premium rate would make the product more attractive to
potential buyers. Thus, it appears that the “market rate” on this type of
insurance remains unsettled. While CNG provides a great deal of credit to
rural borrowers, it is not clear exactly how much of its portfolio is at risk
when climactic events affect agriculture. Additionally, it is difficult for
researchers at the MFI to assess the correlation between default rates and
El Niño events.
La Positiva also provides a variety of other index insurance products.
One such product is area-yield insurance against catastrophic risk, offered
to rural banks in the eight poorest regions in Peru. The goal is to protect
subsistence farmers with fewer than 3 hectares of land against excessive
crop loss. Catastrophic area-yield insurance indemnifies farmers with pay-
ments of 400 to 750 soles per hectare when yields fall below 40% of
average. Regional governments decide upon three to five crops that will
be covered under the catastrophic insurance program. The area-yield
insurance is in its third year of operation, with $17 million covered in the
2009-2010 year and $19 million in 2010-2011. The premium for the con-
tract, which is roughly 14% of the value insured, is 100% subsidized.
Indemnities are paid either directly to the regional governments or to indi-
vidual producers through the national bank.
Several problems have arisen with the catastrophic risk index insurance
program offered by La Positiva. Developed in Mexico by consultants from
LatinRisk, a private company whose mission is the design, development,
and evaluation of agricultural insurance in Latin America, the policies
have a “blanket” feature in that insurance for all crop types faces the same
premium, regardless of how risky the crop. Regional governments have
403
on 01 January 2018
begun to take action, however, suggesting that more area- and crop-
specific contracts be developed. The current concern is that full premium
subsidies, which were intended to be temporary and then gradually taper
off as farmers realized the benefits of the program, have not been lowered
because demand remains low. Regional government authorities have
expressed that demand for catastrophic insurance at market rates might
increase if riskier crops were covered. In response, La Positiva and the
Spanish insurance company MAPFRE are using historical information to
create models of risk assessment, with the end goal of deriving crop-
specific premium rates for catastrophic insurance. Thus, it remains to be
seen whether a fully priced, farm-level catastrophic insurance product
would be viable. La Positiva has not attempted to offer area-yield or other
index insurance to low-income agricultural households at market rates.
Kenya
Similar to ENSO insurance in Peru, other non-rainfall indices have
proven to be good candidates for the design of index insurance contracts.
For example, Makaudze and Miranda (2010) compare two types of index
insurance contracts designed to protect small maize and cotton farmers in
nine districts of varying agro-ecological regions in Zimbabwe: one based
on rainfall and another based on the Normalized Difference Vegetation
Index (NDVI), which is measured using remotely sensed data from
NOAA and NASA satellites. They find that the insurance contract based
on the NDVI exhibits payouts that are more highly correlated with actual
losses, exposing farmers to less basis risk than contracts based on rainfall.
One insurance pilot being developed with the use of satellite data is
drought-related livestock mortality insurance in the arid and semi-arid
lands (ASALs) of northern Kenya. In this region, the prospect of major
livestock losses during severe droughts reduces incentives to build herds.
Thus, insurance is expected to increase investment in livestock, as well as
access to credit. Index insurance products are appealing due to high trans-
action costs of monitoring animal deaths, especially as most herders in the
region move to adapt to spatiotemporal variability in forage and water.
The index selected for the pilot is based on NDVI data, which is repre-
sentative of the vegetation available for livestock to consume (Mude,
Chantarat, Barrett, Carter, Ikegami, and McPeak 2009). Household-level
livestock mortality data collected on a monthly basis from 2000 to 2008 are
used to estimate the relationship between the index and actual losses
using a regime-switching regression model. Fortunately for project design-
ers, the ground mortality data for herders in the region, which are neces-
sary to tie the NDVI observations to quantifiable livestock deaths, are
available from the World Bank-funded Arid Lands Research Management
Project and the Pastoralist Risk Management Project (McPeak, Chantarat,
and Mude 2005). After creating an optimal index insurance contract,
in-depth surveys were conducted randomly in five villages, with the goal
of assessing demand, risk attitudes and willingness to pay for an index
insurance product. Results indicate that almost 70% of herders would
pay premiums 20% over actuarially fair prices to purchase a 30% strike
contract.
In 2010, the ILRI, supported by Financial Sector Deepening Kenya
(FSD), DfiD, USAID, and the World Bank, initiated an index-based
404
on 01 January 2018
livestock insurance (IBLI) pilot in the Marsabit district in northern Kenya

(Barrett, Carter, Chantarat, Ikegami, and McPeak 2009). Equity Insurance
Agency is the insurance agent for the policies, with UAP Insurance acting
as the local underwriter of contracts. Premiums for the NDVI insurance
product, which are either 5.5% or 3.25% of the value insured, depending
on where the client resides, are subsidized by the international donors.
However, the premiums paid by clients are 30% to 40% higher than actua-
rially fair. Market premiums without subsidies would be an additional
30% to 40% higher. Project experts are looking forward to February 2013,
when prohibitions against other commercial insurers from marketing the
product expires. As much interest in underwriting has been expressed by
other insurers, increased competition is expected to lower market
premium rates. While the contracts are targeted to individual farmers, a
second phase of the pilot is slated to focus on meso- and macro-level
clients, such as cooperatives and NGOs or governments and donor agen-
cies, respectively.
Researchers, insurers and NGOs have used varying channels to educate
potential clients about new insurance products. In the case of the index-
based livestock insurance pilot in Kenya, experimental games were
employed. As outlined in Mude, Chantarat, Barrett, Carter, Ikegami, and
McPeak (2009), a four-round game that became increasingly complex
simulated the nature of weather shocks over a ten-year period.
Participants were offered a starting “herd size” and could choose to pur-
chase a hypothetical insurance product in the beginning of each period or
go uninsured. One of five climactic conditions was randomly drawn from
a bag, and insured players were indemnified in the case of “severe
drought” or “drought.” However, while this method was effective, it was
also expensive, as two-facilitator, five-participant play takes as much as a
day to complete. At the end of games, it was determined that most
players were able to understand the importance of insurance, as well as
how the index insurance product worked.
With three sales periods completed, demand for IBLI seems to be taper-
ing off. According to Andrew Mude, economist and project leader at ILRI,
during the first season of contract sales, in January and February 2010,
1979 policies were sold; in the following season, in January and February
2011, the number of policies fell to 599; finally, in the third sales period, in
August and September 2011, only 500 policies were sold. Groundwork
suggests that pastoralists in Kenya want to be reassured that the index
insurance product works; in the most recent sales period, the index trigger
was met, and all policyholders received indemnities. While, prior to the
opening of the sales period for the January/February 2012 season, it was
hypothesized that sales would increase due to the payouts received in the
previous season, no contracts were sold due to undisclosed complications
among partners in the program. Those working on the project hope to
find new partners by the time the next sales period is scheduled to open,
in August and September of 2012.
For the NDVI livestock insurance contracts, ILRI and its partners are
collecting panel data in order to gain insights on program impact.
Currently, researchers are collecting the third round of an annual survey.
Attrition is low, at only about 5% of the initial 925 households in the
panel. After the four-year data collection period, survey data, which
consist of random observations of both insured and uninsured herding
405
on 01 January 2018
households, will be used to analyze the welfare effects of IBLI. A recent

article on the IBLI pilot conducts a performance-based analysis at the
household-level taking into account the presence of both individual- and
location-specific basis risk (Chantarat, Mude, Barrett, and Carter 2012).
Using simulated seasonal livestock mortality data for 2000 households, the
authors find that the index insurance product most effectively protects
herders from catastrophic losses, with almost three-quarters of individual
losses covered by a contract that is triggered when mortality rates exceed
30%.
Overall, the livestock pilot program in Kenya highlights the nuances in
index insurance contract design and sales. In this case, state-of-the art sat-
ellite imaging reduces basis risk and the burden of on-the-ground data
collection, while a lack of competition among insurers drives up premium
rates for herders whose interest in the product already seems to be
waning. Finally, panel data analysis will offer impact evaluation studies
that are critical to the further development of index insurance pilots, espe-
cially where welfare effects on the poorest groups are considered.
Mongolia
For pastoralists in Mongolia, the most prevalent risk is the dzud, a dry
summer followed by a severe winter, which often causes widespread live-
stock losses. Most livestock herders in the region follow traditional pas-
toral practices, preferring to expand herd size rather than use an intensive
approach that would emphasize investment in pasture land and animal
quality (Global-AgRisk, Inc. 2010). While motivation for this behavior is
likely a combination of the lack of land tenure rights in Mongolia and a
weak export market for dairy and meat, both of which undermine invest-
ment, farmers may also be hesitant to shift importance to livestock quality
because they do not have access to insurance to mitigate losses.
Large-scale livestock deaths like those experienced in the 2000-2002 dzuds
that resulted in 9 million head lost caused many households to fall into
acute poverty.
In response to these catastrophic losses, the World Bank initiated the
Index-based Livestock Insurance (IBLI) pilot program in 2006. The
project provided for two insurance products. The Base Insurance Product
(BIP) is a commercial risk instrument sold and serviced by insurance
companies, for which herders pay a fully loaded premium rate. The BIP
pays an indemnity when a district’s (soum) mortality rate exceeds a
defined percentage between 7% and 10% depending on species and loca-
tion. The Disaster Response Product (DRP) is a social safety net product
financed by the government that begins payments at mortality rates that
exceed the exhaustion point of the BIP. Both products were initially
offered in three provinces; in 2010, the project received approval for
additional funding of more than $10 million from the World Bank in
order to scale up operations after the pilot was deemed successful
(WorldBank 2011). According to researchers at the World Bank, the
project is on target to reach all 21 provinces (aimags) in the country by
the end of the 2012.
Underwriting the two insurance contracts involves a series of
risk-layering and risk-pooling arrangements. Losses of 6% or lower are
absorbed by herders, while local insurers cover “mezzanine” losses from
406
on 01 January 2018
6% to 30% in the base insurance product (BIP). Catastrophic losses of 30%

or higher are covered by a disaster response product (DRP) funded by the
Mongolian government, which also provides stop-loss reinsurance to the
local insurers when loss ratios exceed 110%, or 125% in a single province
(Mahul and Skees 2007). Currently, the World Bank is the de facto rein-
surer of both products, as it provides the government with contingent
credit. However, the World Bank’s credit line expires in 2014, and
researchers are currently trying to phase out grant-funded reinsurance
payments.
A special syndicate pooling arrangement, the Livestock Index
Indemnity Pool (LIIP), was created to protect individual insurers from
excessive losses. As most livestock insurers in Mongolia work in localized
areas, the LIIP serves to diversify risk. In addition, each insurer in the syn-
dicate must provide a “guarantee indemnity contribution” to the pool
before each season. The contribution includes 5% of the projected
premium volume, as well as the estimated cost of reinsurance for each
individual insurer, so insurers that are more diversified throughout the
country make smaller payments as a percentage of policy sales to the LIIP
(Mahul and Skees 2007).
In its first season in 2006, about 2,200 BIP contracts were sold in 3 prov-
inces, with 200 herders opting to purchase only DRP. Those who pur-
chased BIP are also covered under the catastrophic risk plan, and thus
receive an implicit subsidy on the DRP. Funding for the BIP and DRP are
completely separate, with the government’s coverage of losses over 30%
completely subsidized rather than priced in the premium for the BIP
(GlobalAgRisk, Inc. 2010). As a further bonus to BIP purchasers, three
major lenders in Mongolia agreed to lower their interest rates by an
average of 1.2% annually for insured herders. Approximately 9% of all
herders purchased BIP. The majority of buyers chose the minimum level
of coverage available, 30%. By 2009, the program had expanded to a
fourth province, with total sales at just over 4,300 BIP contracts, a 12% par-
ticipation rate. In 2010, the pilot grew to nine provinces, and about 7,000
policies were purchased.
From 2006 through 2010, premium payments to the LIIP were just
under $735,000. In contrast, about $1.9 million has been indemnified to
8,700 herders, giving the IBLI a loss ratio of about 2.7. Thus, in its current
state, such an insurance program would not be sustainable. However, in
light of the considerable payouts, climatologists and scientists reviewed
the pricing of the product and found that there were “no alarming trends”
in the frequency of extreme weather events. Rather, the dzud event of
2010 was the worst in recorded history, implying that the loss ratio of the
IBLI program to date is not reliable for long-term rating. In addition,
researchers hope that the extension of the pilot to all provinces will reduce
the volatility of losses. Therefore, while the IBLI has not been profitable
from a commercial standpoint in the past four to five years, experts
believe that BIP is correctly priced.
To gain insight on the welfare impact of the IBLI, the World Bank
Sustainable Livelihoods Program began conducting household-level
surveys in early 2012. The goal is to create a panel from which the
program effects can be estimated. Such data will be of great value to the
development economics community, as few empirical impact analyses of
index insurance programs have been attempted.
407
on 01 January 2018
Vietnam
Early arrival of extreme rainfall and consequential flooding in the
Mekong Delta region of Vietnam interrupts the harvest of the summer-fall
rice crop, causing losses to growers and the Vietnam Bank for Agriculture
and Rural Development (VBARD), which supplies production credit to
the growers. After attempts at traditional agricultural insurance failed in
the 1990s due to large losses by insurers, VBARD acted as a de facto
insurer for its borrowers. Pooling risk by charging a flat interest rate
nationally, VBARD was recapitalized by the government after natural dis-
asters in return for forgiving debts that could not be repaid for weather-
related reasons (Skees, Hartell, and Murphy 2007). However, market
reforms known as “equitization” have changed VBARD debt forgiveness
policies, causing a careful examination of index insurance.
Amid concerns over the lack of an efficient insurance market for rice,
combined with a desire to further commercialize the country’s banking
system, Vietnam’s Ministry of Finance partnered with the Asian
Development Bank and the World Bank to develop an index insurance
product. In designing the contract, the Southern Institute of Water
Resources conducted flood plain modeling and used 1977-2004 data to
conclude that flooding is largely determined by water flows from neigh-
boring Cambodia, prompting the use of daily water flow through the Tan
Chau hydrological station in Cambodia as a flood insurance index.
Payouts are triggered when water levels reach 250 centimeters between
the dates of June 20 and July 15, with maximum payouts occurring when
levels reach or exceed 350 centimeters. This level of flooding represents a
one-in-seven year event, with the most recent early flooding occurring in
2000 (GlobalAgRisk, Inc. 2010).
Two rice-producing districts in the Dong Thap province in the Mekong
Delta region were targeted for potential pilot projects: Thanh Binh, in
which 64% of Summer-Autumn rice experienced flood damage in 2000,
and Tam Nong, in which 48% of Summer-Autumn rice was flood
damaged in the same year. Harvesting of this crop begins in June, so the
timing of the onset of flooding is of great importance for rice farmers,
who face losses when floods occur while rice is still in the field (Skees,
Hartell, and Murphy 2007; Manuamorn and Dick 2009).
Farm households experience yield loss, quality loss, higher harvest and
post-harvest handling costs, and numerous other losses, including those
due to the diversion of household labor from productive activities to loss
mitigation efforts, such as moving animals or filling sand bags to place on
dykes (GlobalAgRisk, Inc. 2010). Because sales to individuals would be
difficult due to basis risk, and due to regulations that would require the
insurance to be marketed as loan default insurance, the policy was
designed as business interruption insurance to be sold to VBARD, who
provides implicit insurance to rice producers by restructuring loans after
natural disasters.
Bao Minh first offered the fully priced index insurance product to
VBARD in 2008, with ParisRe signed on as the reinsurer, for an insured
sum of $1 million. However, due to the late timing of the attempted sale,
the likelihood of early flooding was low and VBARD chose not to
purchase a contract. In addition, although it expressed interest in the 2009
season and the contract offer was sufficiently early and preceded by two
408
on 01 January 2018
agricultural insurance education workshops, as well as the preparation of

additional materials specifically requested by VBARD, the bank again
rejected index insurance. VBARD did not provide concrete reasons for
refusing to enter into an insurance contract with Bao Minh, although at
negotiations in Hanoi VBARD representatives expressed legal concerns
related to the tax consequences of receiving indemnities. After the sales
date closed on May 15, 2009 without a purchase, VBARD did not provide
formal feedback, although sources said it was possible that VBARD had
already set aside cash reserves, or that VBARD did not think it would
flood in 2009 (GlobalAgRisk, Inc. 2009).
Although sales of the flood index product were ultimately unsuccessful,
new efforts to design meso-level insurance products in Vietnam have
arisen in reaction to the government’s recent support of subsidies for the
rice insurance market. Currently, VBARD is said to be in talks with the
Agriculture Bank Insurance Corporation (ABIC) in Vietnam and global
reinsurer SwissRe to develop a rice area-yield index policy tied to loan
defaults.
Micro, Meso, or Macro?

Some of the most important lessons learned from index insurance
research and pilot program activities pertain to the relative merits of
micro-, meso-, and macro-index insurance products.
Micro-Products
Micro-index insurance products arguably have the greatest potential to
help farmers directly, since such products ostensibly can be tailored to
their risk management needs. However, basis risk, the failure of indemni-
ties to adequately track the losses of the insured, is greatest for micro-
products due to variations across farmers in production practices and
growing conditions. In order for micro-index insurance products to be
offered on a wide scale, basis risk, which is discussed further in the next
section, would have to be addressed.
Developing a market for micro-index insurance product also requires
substantial investment in the training and education of farmers. To create
demand for a micro-index insurance product, farmers would have to be
educated about the relative costs and benefits of such products and be
properly trained to use them effectively to manage their risks. Above all,
farmers would need to be made keenly aware of the limitations of index
insurance products, most notably that they do not cover idiosyncratic pro-
duction losses from perils not directly identified with the index.
Furthermore, the marketing of micro-index insurance products to less edu-
cated farmers is likely to receive especially close scrutiny from regulators,
creating market development costs arising from efforts to properly place
the marketing of such products within the legal and regulatory system of
the country or region.
In addition to basis risk and educational requirements, micro-index
insurance products, if they are to be offered on a wide scale, require
investment in efficient delivery mechanisms, including the training of a
sales force and the development and maintenance of physical points of
409
on 01 January 2018
purchase and indemnity collection. These costs can be reduced by using

established agricultural marketing channels such as banks, input suppli-
ers, processors and other risk aggregators to serve as the vehicles of deliv-
ery, either by bundling index insurance with production loans or making
it available for purchase in conjunction with production loans.
However, the frequency of interaction and strength of farmer marketing
relationships can be quite varied, raising questions about the general
applicability of market development strategies that rely on established
marketing channels. To date, micro-index insurance pilot program suc-
cesses have been limited to controlled, relatively small-scale programs
designed around specific commodity farmer groups with strong, well-
established contractual relationships with lenders or processors who have
supported program development by providing educational and delivery
services. Whether the lessons learned from these limited experiences will
be directly applicable to larger scale index insurance programs that
attempt to reach less formally organized and poorer farmers remains a
largely open question.
Meso-Products
Meso-index insurance products avoid some of the problems associated
with micro-products, but share others and present unique problems of
their own. The basis risk problems that undermine micro-index insurance
contracts can be expected to be less pronounced for meso-products offered
to agricultural lenders and other actors in the marketing chain that
provide financial services to farmers. Lenders and other “risk aggregators”
possess loan portfolios that include loans to many farmers who are dis-
persed geographically and may further vary in production practices.
These loan portfolios, if reasonably large, should diversify much of the
idiosyncratic production risks borne by clients, but retain the systemic
risks due to widespread adverse weather effects. As such, an index that
measures systemic adverse agricultural production shocks in the lender’s
geographical scope of business should track lender cash flow shortfalls
more closely than those of any one individual farmer client.
Compared to micro-index insurance products, meso-index insurance
products are targeted to an audience with greater financial analytical
capacity. An important advantage of meso-products is that lenders and
other users of meso-products should be better able to understand the limi-
tations of index insurance products. Meso-index insurance products can
be more complex than micro-products, allowing indemnity schedules to
be designed to more precisely capture the complex relationship between
weather and policyholder losses. Reducing basis risk by designing index
insurance contracts based on compound indices and highly nonlinear
indemnity schedules becomes possible. Furthermore, obtaining regulatory
and legal authority to market meso-insurance products is likely to face
fewer logistical obstacles, given that lenders have a history of working
with regulators. In addition, meso-index insurance products require less
extensive delivery mechanisms, since in many cases they can be sold
through over-the-counter transactions. And finally, since the number of
decision-makers who need to be properly educated is much lower with
meso-products, educational costs can be substantially reduced.
410
on 01 January 2018
Meso-index insurance products, however, present very different techni-

cal and contractual enforcement issues related to their use. Proper uses of
index insurance by lenders in holistic portfolio risk management require a
deeper understanding of the cash-flow risks faced by lenders and the
practices they currently employ to manage such risks. Although lenders
arguably are more sophisticated than farmers and thus better able to
implement complex risk management practices, it is also true that many
rural lenders in developing countries lack a culture of active risk manage-
ment practices that employ insurance, reinsurance, and derivative prod-
ucts. Operational cash-flow models and risk management practices can be
intricate and opaque, and can vary from one lender to the next. Efforts to
develop lender portfolio risk management strategies that incorporate
index insurance can encounter difficulties if lenders are reluctant to
openly discuss their trade and internal cash-flow management practices
with index insurance specialists. Unfortunately, there has been very
limited experience involving the use of meso-index insurance to holisti-
cally manage loan portfolio risk. The proper uses and potential benefits of
meso-products by lenders in such contexts require more study and invest-
ment in educational programs for lenders.
Finally, another potential problem with meso-products is that donors
and other sponsors of developmental index insurance programs are typi-
cally motivated by the desire to directly and demonstrably improve the
welfare of poor farmers. As such, they are more likely to endorse the
development of micro-index insurance products because they appear to
provide higher benefits to poor farmers, compared to meso-products,
which appear to benefit relatively more affluent lenders. Thus, for meso-
index insurance products to take root with support from donors and inter-
national agencies, there is a very practical need to demonstrate that they
can generate tangible benefits to poor farmers, either through lower inter-
est rates on loans or through significant expansion of services offered to
such farmers. Of special interest is whether relaxing debt restructuring
and forgiveness policies are effective means of transferring the benefits of
meso-index insurance to farmers, as such measures could weaken the
lending relationship by creating incentives to default.
Macro-Products
Macro-level products purchased by national governments, regional gov-
ernments, and non-governmental relief agencies promise to allow more
cost-effective, transparent, and timely usages of scarce resources, better
catastrophe planning and preparedness and, in the case of risk financing,
reduce the dependence of countries on often unreliable and untimely ad
hoc external assistance in the aftermath of major weather catastrophes
(Hess, Wiseman, and Robertson 2006). Macro-products, moreover, can be
designed to provide assistance in times of regional or less widespread
weather catastrophes that do not surpass the threshold necessary to
induce international relief efforts.
Macro-index insurance offers potential improvements to slow-onset
emergencies such as drought, as it can smooth the stream of payments
and serves as part of an effective risk-layering strategy for low-probability,
high-impact events. Examples of such programs include FONDEN in
Mexico, where the federal government insures local governments for the
411
on 01 January 2018
reconstruction of local infrastructure, the implementation of temporary

employment programs and the direct payment of farmers for income
losses after natural disasters (Skees, Varangis, Larson, and Siegel 2002)
and the World Food Program’s purchase of USD 7 million in contingency
funding from AXA Re in the case of drought in Ethiopia in 2006 (World
Food Programme 2006). In addition, Chantarat, Barrett, Mude, and Turvey
(2007) outline a model of index insurance use for famine prevention in
northern Kenya, wherein rainfall is directly linked to a human outcome of
interest, such as the prevalence of human wasting.
Macro-index insurance products are designed to be used by a single, rel-
atively well-educated decision-maker. As such, macro-index insurance
products do not require complicated delivery mechanisms or educational
efforts on par with micro- and meso-products. Macro-index insurance
products can typically be sold in straightforward over-the-counter transac-
tions with few, if any, regulatory oversight issues. Macro-weather insur-
ance products can be custom-tailored to the needs of the national
government or other organizations desiring to use these products for relief
measures, and can be more sophisticated in structure, allowing for more
complex designs that better capture the relationship between weather and
agricultural losses, and thus reduce basis risk.
Macro-index insurance products, however, present some unique prob-
lems. Macro-products designed with a single aggregate national index
may fail to capture regional catastrophes. Also, macro-products, to be
effective, would in many cases require developing detailed contingency
plans for their use in times of catastrophic weather. Potential uses of the
indemnities provided by a macro-index insurance product need to be
articulated in advance of catastrophe, and emergency preparedness and
response measures need to be carefully designed around the receipt of
those indemnities. Moreover, macro-index insurance products are more
susceptible than micro- and meso-products to fraud, political rent-seeking
and other misuses of received indemnities.
Perhaps the greatest challenge to macro-index insurance products,
however, is that the purchase of such products by a national or regional
government could reduce incentives for the global community to provide
ad hoc financial support in times of widespread catastrophes. Investment
in a macro-index insurance product, therefore, could prove cost-ineffective
from a government’s perspective if such an investment leads to a reduc-
tion in donor support. Thus, macro-index insurance schemes are more
likely to succeed if they are developed in partnership with, and perhaps
with financial support from, international donor groups and relief agen-
cies. Greater efforts should be made to work with international relief agen-
cies to explore the use of macro-index insurance as effective tools in their
humanitarian relief strategies.
Index Insurance and Credit

Strengthening the credit relationship between farmers and lenders has
been the articulated objective of most index insurance pilot projects. For a
variety of reasons, this is a very sensible starting point for testing and
developing index insurance. First, focusing on the credit relationship
412
on 01 January 2018
helps frame education, contract design, and market delivery development

efforts. Second, focusing on the credit relationship provides a test of how
the benefits of index insurance will propagate through the marketing
chain. Third, strengthening the credit relationship between lender and bor-
rower could have far-reaching consequences. Specifically, if index insur-
ance can substantially reduce systemic loan defaults suffered by lenders
during times of widespread weather catastrophes, then lenders might be
able to reduce interest rates and expand services to farmers who currently
have limited or no access to credit.
In most pilot programs targeted to the credit relationship, index insur-
ance products have been offered, not as stand-alone products made avail-
able to farmers, but rather by bundling them with credit packages offered
to farmers by lenders. Many lessons regarding the interactions between
index insurance and credit have been learned from these pilot programs,
some of them unanticipated. For example, Malawian farmers’ demand for
credit is found to decrease when loans are bundled with a rainfall insur-
ance contract, even though there is considerable risk of income loss due to
drought (Giné and Yang 2009). In this study, higher levels of education
increase take-up rates of the insured loan, while education is not signifi-
cantly correlated with the choice to take out an uninsured loan. A
randomized experiment offering indemnified loans to farmers in Ghana
finds no significant difference in loan uptake among treatment and control
groups (Karlan, Kutsoati, McMillan, and Udry 2011), although farmers in
the treatment group are found to shift production to a more perishable,
and therefore riskier, crop.
Bundling index insurance with credit contracts appears to have enjoyed
some success in reducing loan defaults in cases where the lender-borrower
marketing relationship is strong, but has proven far less effective in cases
where borrowers have ready means to default on loan repayments
without suffering major consequences. While the ability to obtain index
insurance may increase credit access, there is some concern that index
insurance may increase moral hazard in the credit market if the policy is
not carefully designed. Clarke and Dercon (2009) argue that insurance can
“crowd out” credit markets by implicitly reducing the severity of punish-
ment when households default on loans. Index insurance, by effectively
increasing the minimum welfare level a household can achieve should it
default, reduces incentives for repayment and, in turn, results in lenders
having to cut back on the amount of credit they can profitably offer to
clients. It is noteworthy that the converse may also be true: index insur-
ance could reduce moral hazard in credit markets under special circum-
stances. In Morocco, for example, the country’s public agricultural bank
forgives farm loans following a drought; if weather insurance were made
available, borrower repayment discipline might improve as drought
would be less likely to influence one’s ability to repay (Skees, Gober,
Varangis, Lester, and Kalavakonda 2001).
Ultimately, bundling index insurance with credit is not a substitute for
strong loan recovery practices, and cannot compensate for weak ones.
Index insurance can strengthen credit relationships only through
improved monitoring and enforcement of contract performance, which
could be addressed, say, by establishing functional national identification
and credit information systems or employing collateral substitutes such as
413
on 01 January 2018
joint liability groups or buyer agreements. Without such measures, bun-

dling strategies may fail when attempts are made to scale up the index
insurance market to reach farmers without strong formal marketing
relationships.
However, bundling index insurance with credit has failed to demon-
strate whether it is of any real value to the farmer, since farmers may well
be buying the insurance simply to obtain the value provided by the access
to credit. Bundling index insurance with credit, but entitling the farmer to
receive the indemnities, the standard practice in most pilot programs, con-
verts the index insurance contract into a micro-product, bringing with it
all the problems and limitations associated with micro-products, most
notably higher basis risk and higher development costs. Bundling index
insurance contracts with loans could be more effective if the lender is
named the beneficiary of the indemnity, as this would reduce the impact
of widespread loan defaults on the bank, the risk of which manifests itself
in higher interest rates charged to borrowers and in reduced scope of
banking operations.
Other Lessons Learned

Index insurance pilot programs have provided lessons that cut across
the type of contract, whether micro, meso, and macro.
Basis Risk
Basis risk is endemic to index insurance contract design and arises
regardless of the index used, the policyholder for which such contracts are
intended, and the venue for which the contract is developed. Basis risk is
the most serious obstacle to the effectiveness of index insurance as a
general agricultural risk management tool. Basis risk at any level, whether
micro, meso, or macro, can be partially addressed in one of two ways.
First, the basis risk associated with index insurance products can be
reduced by offering a wider array of index insurance products, tailored to
different risk exposures. For example, virtually all rainfall index insurance
pilot project studies conducted to date have concluded that investment
into additional automated weather stations, rainfall gauges, and/or river-
level gauges is needed if the pilot project is to succeed on a larger scale.
However, the costs of developing an array of tailored weather insurance
products and constructing, maintaining, and securing a broad network of
weather stations can be substantial, creating an impediment to the expan-
sion of index insurance programs, particularly micro-insurance programs,
beyond the pilot stage. Moreover, building new weather stations on which
to write index insurance contracts introduces the additional problem that
historical index data for those stations do not exist, creating ambiguities
for the actuarial rating of the index insurance contracts written on these
stations, likely resulting in higher reinsurance premiums.
Second, the basis risk associated with index insurance products can be
reduced by designing contract indemnity schedules so that indemnities
correlate maximally with policyholder losses. However, such efforts are
frustrated by the scarcity of farm-level yield and loss data needed to opti-
mize product design. Moreover, the design of index insurance products is
414
on 01 January 2018
constrained by the lack of sophistication of their intended users. Due to a

lack of farmers’ experience with active farm-level risk management using
financial products, index insurance products, particularly micro-products,
typically must possess a simple and transparent design. This requirement
limits the development of sophisticated insurance contracts whose indem-
nity schedules precisely capture the complex relationships between
weather and farm-level production, including products based on com-
pound indices using multiple variables and contracts with highly nonlin-
ear indemnity schedules.
However, even if the diversity of products is expanded and the correla-
tion between indemnities and losses are maximized through optimal
design, natural limits exist regarding the degree to which basis risk can be
reduced. Losses at any level, but particularly at the farm-level, can be
caused by many factors that simply are unrelated to the index used in con-
tract design. Such factors might include, for example, sub-optimal expo-
sure to sunlight, wind damage, fire, pestilence, disease, and accidents
besetting farm workers. It is entirely possible, therefore, for a farmer to
live adjacent to a rainfall gauge for which a contract has been optimally
designed to reflect local growing conditions, yet the contract could still
embody substantial basis risk if rainfall can explain only a portion of the
variability of the farmer’s losses.
Unfortunately, little is known about the severity of index insurance
basis risk at the farm-level. This is due, first and foremost, to the scarcity
of farm-level production and loss data. Due to the lack of production and
loss data, empirical assessments of the risk-reduction benefits of index
insurance to individual agricultural producers have been few in number.
Various institutions, however, have undertaken efforts to overcome the
limitations of farm-level production data, particularly as it pertains to the
design of index insurance indemnity schedules. Research undertaken by
the International Research Institute on behalf of the World Bank, for
example, has demonstrated that models based on water requirement satis-
faction have a remarkable ability to predict aggregate yields and to
capture the complex relationships that exist among aggregate yields and
rainfall (Osgood, McLaurin, Carriquiry, Mishra, Fiondella, Hansen,
Peterson, and Ward 2007). However, these models have proven to lack
robustness with regard to optimal index insurance design. Perhaps more
importantly, these models have proven unable to explain yield variability
at the farm-level, were production is subject to idiosyncratic shocks that
are diversified away in aggregate yields. Yet it is precisely the relationship
between farm-level yields and rainfall that needs to be understood in
order to assess the basis risk associated with index insurance products.
Without a better understanding of basis risk, it is not possible to
assess ex ante and with any precision the extent to which investments in
a broader network of weather stations will provide benefits that exceed
the costs. Without a better understanding of basis risk at the farm-level,
the only way to discover the benefits of investing in expanding weather
station networks is to make the investment and wait to see if there is
take-up in the market for stand-alone products. This, however, is a
highly risky proposition. Complicating such assessments is the notion
that basis risk can vary substantially across markets and across the
globe.
415
on 01 January 2018
Spillover Effects
Index insurance can potentially undermine the informal communal risk-
sharing mechanisms that have been built up by the poor to cope with
income fluctuation they cannot address through financial markets. As
mutual support networks within communities tend to be somewhat frail,
their continued existence relies on the incentive compatibility of all
members; no individual can have a motivation to leave the group, as com-
mitment is not likely fully enforceable in these arrangements and any
defection undermines the risk-sharing system (Clarke and Dercon 2009).
At the same time, the poorest of the poor may gain inclusion into informal
social safety nets if index insurance were available to prevent asset losses
in the face of catastrophic risk. Santos and Barrett (2011), for example, find
a middle-class bias in informal reciprocal lending arrangements among
Ethiopian pastoralists, whereby those who are “too poor” (i.e., close to or
below a dynamic asset threshold) are less likely to be offered in-kind live-
stock loans from community members.
The design of index insurance products is crucial in order to avoid
further marginalizing the poorest of the poor due to other indirect effects,
namely those materializing as price changes that result in the aftermath of
weather events that affect yields. If landless laborers and net consumers
do not have access to index insurance, either because policies require land
ownership or because premiums are simply too high, insured farmers see
a relative increase in purchasing power after being indemnified, and thus
bid up prices for the scarce food products available following a bad
harvest (Morduch 2006). Also, in many cases, insurance products are not
offered to the most vulnerable landowners. Malawian farmers located in
the most drought-prone areas of the country, for example, were excluded
from a rainfall index insurance pilot because of feasibility issues tied to
relatively unpredictable rainfall patterns (Osgood, Suarez, Hansen,
Carriquiry, and Mishra 2008). Thus, the provision of insurance to some
may actually exacerbate the losses of the segment of society that cannot
purchase insurance. On a more positive note, Carter and Janzen (2012)
find preliminary evidence that livestock insurance payouts in Kenya
should help offset severe, post-drought price swings that characterize the
region and are a key component of dynamic poverty traps.
Data Limitations
Accurate rating and efficient design of index insurance products
requires long, clean, internally consistent historical data records. Because
less data leads to higher premium rates charged by insurers for index
insurance products, index insurance offered in areas where little data are
available are less likely to succeed. As learned in a World Bank index
insurance pilot program in Ethiopia, areas covered by weather stations
with 20% or more missing data from the last 30 years would not be viable
candidates for insurance pilots (Bryla 2009).
Unfortunately, most index insurance research and pilot program find-
ings have cited insufficient or low quality weather data as a major obstacle
to the expansion of index insurance programs beyond the pilot phase. The
availability of reliable weather data at needed temporal resolutions varies
across developing countries and in some countries the scarcity of data can
be especially severe. It is not uncommon, for example, for available rainfall
416
on 01 January 2018
data series to span fewer than twenty years or to exhibit major gaps of
missing observations, even at World Meteorological Organization-
registered weather stations. Moreover, it is not uncommon to find evi-
dence of ad hoc attempts to fill in missing data. For example, in some
rainfall data series, it has been noted that the exact same nonzero level of
rainfall is recorded for a sequence of consecutive days, a highly improb-
able occurrence.
One example of where a scarcity of data severely affects contract design
is that of area-yield index insurance. Because the basis risk associated
with area-yield index insurance is, in theory, lower than that of contracts
based on weather or irrigation water supply indices, farmers’ willingness
to pay for area-yield insurance has been estimated to be twice as high as
that for rainfall insurance (Carter, Boucher, and Trivelli 2007). However,
although area-yield insurance may be more desirable, rainfall data are
more likely to be available in developing countries, whereas reliable and
sufficiently detailed yield records may be nonexistent. Rainfall collection
systems are also less costly than developing a reliable yield estimation pro-
cedure for small geographical regions (Skees 2000).
The World Bank has invested in technical meteorological and statistical
studies aimed at creating synthetic rainfall data by correlating satellite
data and ground station rainfall measurements. These methods can
provide reasonable estimates of conditional expectations of rainfall on
highly-refined spatial grids. However, synthetic rainfall index data con-
structed in this fashion have also been found to underestimate the varia-
bility of rainfall at any location and to under-predict the duration of dry
spells. The use of synthetic data could therefore substantially bias esti-
mates of expected index insurance indemnities, thus undermining the
actuarial performance of the index insurance product and leading to sub-
optimal index insurance contract designs, as well as higher reinsurance
rates.
Similar to client education initiatives, the public good nature of a reli-
able times series of index data suggests that governments and donors
should take responsibility for collecting and maintaining such data, given
that it is unlikely that private sector companies will wish to assume the
costs. Such public investments include the upkeep of existing weather sta-
tions and the building of new ones in areas that are too distant from
current infrastructure to be considered for index insurance pilots.
Data Security
For index insurance contracts to be viable, both the insured and the
insurer must be confident that measurements of the index used to settle
contract claims are reliable and secure. Security of index measurements at
ground-level meteorological stations can be enhanced in a variety of ways.
First, meteorological stations can be made more secure by fencing them in
or staffing them with a reliable employee. Second, meteorological stations
can be upgraded to provide continuous, real-time reporting of data, as
this would allow easier detection of attempts to tamper with the equip-
ment. Third, to detect anomalies in the data that could indicate tampering,
a “buddy-check” system could be implemented in which index measure-
ments at one ground station are systematically compared to measurements
417
on 01 January 2018
at nearby ground stations, or to measurements obtained through remote-

sensing, satellite technology.
Another way to address security concerns associated with ground-level
measurements is to write the index insurance contract based exclusively
on data provided by remote-sensing technology. Drawbacks to this
approach are that satellite data series are typically shorter than those
available at established ground-level stations, they are more expensive to
obtain and analyze, and they may be confounded by the fact that
remote-sensing procedures have been changed from time to time, partic-
ularly when older weather satellites are replaced by newer ones. In addi-
tion, contracts based on satellite measurements will naturally employ
more complicated indices, such as the NDVI and its variants, which are
not as readily understood as more tangible rainfall-based indices,
which at least are partially but directly observed by the insured. Efforts
to employ remote sensing technology in writing index insurance con-
tracts appear promising, but are still in early stages of design and
development.
Capacity Building
By capacity building, we refer generally to efforts to create the condi-
tions wherein the market for index insurance can, in the future, be inde-
pendently sustained and developed to the maximum beneficial extent by
local stakeholders, without continued direct involvement by international
agencies. Local capacity is built primarily through education, outreach,
technical assistance, and other forms of knowledge transfer, but also
involves the forging of sustainable cooperative working relationships
among stakeholders. More specifically, effective capacity building
requires: establishing an institutional framework that allows stakeholders
to cooperate in finding new and appropriate ways to use index insurance
to enhance the performance of the agricultural value chain and to
improve the welfare of the rural sector; transferring essential technical
skills to insurers, reinsurers, and government agencies that would permit
new index insurance products to be designed and old ones to be refined
to meet the varied and changing risk management needs of farmers,
lenders, agri-businesses, and government authorities; transferring essen-
tial risk management skills that would allow for the proper use and
optimal integration of index insurance with the operational practices of
farmers, lenders, agri-businesses, and government agencies; and develop-
ing a legal and regulatory framework that would foster the evolution of
markets for index insurance through entrepreneurial initiative.
Pilot program experiences have repeatedly demonstrated that strong
institutional working partnerships among stakeholders are essential for
establishing an efficient, self-sustaining index insurance program with the
potential for independent growth. Building institutional networks and
working partnerships has been a high priority in all index insurance pro-
grams, and has been most successful when it has involved multiple stake-
holders, without focusing on only one. Nontheless, building institutional
networks has been found to be most successful if a leader can be identi-
fied among stakeholders, either in the form of a powerful farmer group,
processor, or lender. Strong leadership serves as an effective catalyst to
418
on 01 January 2018
the creation of strong institutional partnerships, as the leader can serve as

an example for other stakeholders and can facilitate the creation of a
network by drawing on its extensive political and business contacts.
The transfer of technical knowledge pertaining to the design of index
insurance products is also essential if the marketing of such products is to
pass to local insurers. Virtually all insurers employ resident actuaries with
sound analytical skills. However, most of these skills have evolved from
their application to conventional insurance, and are based largely on expe-
rience and loss-rating methods that assume independence among losses of
the insured. The design and rating of index insurance products, however,
presents unique problems and requires special skills that are not part of
traditional actuarial training. To properly design index insurance con-
tracts, it is important to understand the basic agronomic relationships that
exist between weather indices and agricultural production. The design of
such contracts, moreover, must rely on basic weather and production data,
rather than the experiential loss data most commonly the object of analysis
in conventional actuarial analysis. Furthermore, correlated risk and basis
risk, which are relatively minor concerns with conventional forms of
insurance, become major issues when designing index insurance contracts.
Substantial effort has been invested in educating stakeholders who
directly use index insurance products, most notably farmers, farmer
groups, lenders, and processors. From the outset, most pilot programs
demonstrate that index insurance, because it is a relatively new financial
product without an established record of success, is often received with
considerable skepticism by such stakeholders. Educating stakeholders can
be achieved only if educational programs are targeted to the unique con-
cerns of each stakeholder and are attentive to the differences in educa-
tional level and technical training across stakeholders. The use of
experimental games to educate farmers is one example of index insurance
educational strategies. Experiments are advantageous in that repeated
games can simulate multiple cropping periods, so that those who would
potentially benefit from index insurance can see the complexities of the
process without having to purchase the product and wait ( possibly for
several years) to see its effects (Carter et al. 2008). Such games have been
employed in Kenya, where almost 100% of pastoralists chose to insure,
with most selecting full herd coverage, by the end of play where premi-
ums were actuarially fair; in Peru, where two-thirds of the 450 participat-
ing cotton farmers chose to “purchase” insurance during the game; and in
Malawi, Ethiopia and India, where the goal was to increase trust among
farmers and index insurance providers (Patt, Peterson, Carter, Velez, Hess,
and Suarez 2009).
Outreach to smallholders can be especially expensive and challenging,
given that most farmers lack experience using market insurance and, in
most cases, a lack formal education. Outreach to lenders also presents
challenges, given that most lenders do not have a history of active risk
management using financial instruments such as insurance. Most pilot
program educational efforts have largely been limited to situations in
which lenders and farmers have well-established marketing relationships.
The successes of these efforts owe a great deal to the fact that farmers and
lenders clearly have a joint stake in using index insurance to support and
sustain their credit relationship. Little is known, however, as to whether
419
on 01 January 2018
the same educational methods will be transferable to the education of

farmer without such relationships.
The Way Forward

Research on index insurance and index insurance projects has answered
some questions about the cost-effectiveness of index insurance, but has
also raised new ones. From these efforts, new ideas regarding the use of
index insurance and new challenges to the development of index insur-
ance markets have emerged. As such, there is still a need for additional
work before we can establish whether index insurance can help address
the catastrophic risk management problems faced by the poor living in the
developing world. In this concluding section, we provide some directions
for future research and suggestions for pilot projects that might be under-
taken to advance index insurance in the future.
Develop Meso-Insurance
Micro-insurance products have not been proven to provide direct sub-
stantive risk-reduction benefits to farmers in excess of costs. In many
index insurance pilot projects, demand for micro-products has for the
most part been artificially generated by making their purchase compul-
sory as a condition of obtaining credit or participating in marketing con-
tracts that provide value to farmers. Evidence from many pilot projects
indicate that farmers would not be willing to pay market rates for index
insurance. This would impose severe constraints on efforts to expand the
market for index insurance by offering stand-alone micro-products to
poorer farmers without strong formal marketing relationships. It also
raises questions about the value to farmers of bundling index insurance
contracts with loans, unless such a practice leads to significant reductions
in interest rates or an expansion of credit market services as a result of
reducing the portfolio risk borne by lenders.
As an alternative to developing micro-index insurance products, greater
emphasis should be placed on developing meso-index insurance products.
Meso-products are subject to lower basis risk and lower development
costs than micro-contracts, and stand a better chance of passing the
benefit-cost test. However, uses of meso-contracts to manage lender port-
folio risk and methods to share the benefits with farmers have not been
fully tested in pilot projects. Proper uses of meso-index insurance by
lenders in holistic portfolio risk management require a deeper under-
standing of their cash-flow risks and risk management practices.
Operational models of lender cash flows that capture the effects of
adverse catastrophic weather events and can be used to assess alternative
risk management practices employing meso-index insurance need to be
further developed in close cooperation with lenders. Promising lender
portfolio risk management practices should then be tested in the field.
Another question that demands further examination pertains to how the
benefits of meso-index insurance products can be shared with borrowers.
Pass-through schemes that disburse indemnities received by lenders to
farmers will be effective only if the lender exploits informational advan-
tages from personal knowledge of the borrower. In short, pass-through
schemes will provide substantive benefits and address basis risk only if
420
on 01 January 2018
benefits are awarded only to those farmers with legitimate losses beyond
their control. It has been proposed that mutual groups of farmers should
be established to purchase meso-products and distribute benefits to
farmers. Here again, however, for such mechanisms to be efficient in risk
transfer, the benefit disbursement rules must ensure that only farmers
who have demonstrable losses receive indemnities. Simple rules that call
for equal sharing of indemnities paid by an index insurance contract to a
mutual would effectively turn the product into a micro-product and
reduce its benefits by re-introducing farm-level basis risk. Disbursements
to farmers should take the form of having the mutual repay the produc-
tion loans of its members first, so as to contribute to the strengthening the
credit relationship.
Perform Impact Analysis

Given the range of pilot programs undertaken in recent years, greater
effort should be devoted to assessing the ex-post impact of index insur-
ance on farmers, lenders, and governments in a more comprehensive way.
Market take up is certainly one measure of success. However, such a
measure is flawed where the purchase of micro-products has been made
compulsory as a condition of entering into credit and other marketing
contracts.
Formal follow-up impact assessment studies based on standard benefit-
cost principles, conducted using appropriately designed survey instru-
ments or controlled randomized field experiments could provide answers
to a myriad of pressing questions pertaining to the large-scale viability
and the value of index insurance, including: Have farmers in pilot pro-
grams benefited from micro-products? Have indemnities provided by
index insurance contracts allowed farmers to alleviate what would other-
wise have been serious personal hardships? Have indemnities failed to
materialize at times of need? Has index insurance allowed farmers to
increase their profitability? Have changes in farmers production practices
increased farm-level income risk? Would farmers be willing to purchase
micro-index insurance at market rates, if their purchase were not made a
condition of acquiring a loan? And, if not, how much would farmers be
willing to pay for index insurance contracts and how does the willingness
to pay vary with respect to production practices, education, geographical
location, proximity to weather stations, access to financial markets, and
farmers’ asset base?
The impact of index insurance on lender practices and profitability and
the loan portfolio risk-reduction benefits it has provided also needs to be
assessed: Do lenders believe that their loan portfolios are more resilient to
systemic shocks as a result of incorporating index insurance in their practi-
ces? Have lenders seen a noticeable decline in loan defaults in times of
widespread adverse weather events? Has index insurance encouraged
lenders to reduce interest rates charged to farmers or alter loan recovery
and principal repayment scheduling practices so as to benefit farmers?
Has index insurance encouraged lenders, input suppliers, and processors
to expand their businesses? Have such expansions taken place and, if so,
has this expansion benefited farmers on or outside the boundaries of the
marketing chain? Has index insurance strengthened the lender-borrowing
relationship in tangible ways? Have lenders been able to incorporate index
421
on 01 January 2018
insurance into holistic business risk management practices, or have they

simply chosen to bundle the insurance with other services?
Measure Basis Risk

Absent individual farm-level yield and production loss data, it is not
possible to assess the risk-reduction benefits or potential demand for
stand-alone micro-index insurance contracts. The absence of such data
also frustrates efforts to develop optimal index insurance product designs.
Given the central role played by basis risk in determining benefits of and
demand for index insurance, at least some modest efforts should be made
to assess its magnitude. Efforts to measure basis risk would require the
collection of farm-level yield and loss data, which should be coupled to
the collection of other information relevant to index insurance design,
such as production practices, location, and growing conditions. Although
longitudinal farm-level yield data spanning several years and encompass-
ing inter-annual variation in weather would be ideal for assessing basis
risk, cross-sectional data would at least provide us with some indication
of how individual yields vary around aggregate yields. It is possible that
individual yield data exists if the respective agricultural ministries
perform aggregate yield estimates using cross-sectional production
surveys.
Promote Portfolio Risk Management

Most index insurance pilot programs have been developed around the
predominant cash crop grown by farmers or farmer groups with strong
marketing relationship with lenders or processors. This has been a sensi-
ble strategy, given that the focus on one crop and the existence of an estab-
lished marketing relationship could be used to streamline pilot program
delivery and development costs. However, it is axiomatic that farmers are
primarily concerned about potential threats to their income, which in
many cases is generated by more than one crop and further depends on
prevailing crop prices. This suggests that index insurance products should
be more valuable to producers if they are developed for portfolios of farm
crops and if indemnity payments are permitted to reflect prevailing
market prices. Coupling index insurance with forward and futures con-
tracts, assuming the latter are available, could substantially enhance the
value of such contracts. However, efforts to turn index insurance products
into more general income insurance instruments that depend on prices
depend on the existence of a competitive market that permits transparent
price discovery.
Establish Index Insurance Data Center

A broad range of index insurance pilot programs have been undertaken,
covering many crops, contractual designs, and regions of the world. These
pilot projects vary in their current stage of development, and the experien-
ces have varied markedly from one instance to the other. Actuarial data
summarizing the experiences in each of these projects should be published
in a systematically organized and publicly accessible fashion. Such
data would be invaluable to insurers, reinsurers, governments, non-
governmental agencies, and academics interested in index insurance
422
on 01 January 2018
market development in new venues throughout the world. Annual actua-

rial data to be published should include the total number of policies sold,
total liability, total premiums collected from policyholders, total premium
subsidies provided by governments and donor groups, and total indemni-
ties paid, and should further include detailed summaries of contract struc-
ture, who purchased the contract and where, who insured and reinsured
these products, how were risk layers apportioned among insurers, rein-
surers, donors, and the government, and, if possible, the reinsurance premi-
ums that were paid. Some of this information can be gleaned from available
publications. However, the data are not always recorded in a consistent
fashion, are often difficult to find. To the extent possible, international rein-
surers, donor groups and multilateral agencies that have been involved in
index insurance development efforts should cooperate by contributing
essential actuarial data to what ideally would be a comprehensive, publicly
available database.
Attempt Market Expansion

Most pilot projects have been developed around a specific crop or
farmer group and typically with the involvement of one or a very small
number of lenders or processors. Although many lessons have been
learned from these efforts, questions remain regarding how useful these
lessons will be when attempting to develop index insurance projects on
grander scales. It has been found, for example, that bundling index insur-
ance contracts with credit can help strengthen farmer-lender contractual
relationships where such relationships already exist; however, efforts to
bundle index insurance with credit contracts has had disappointing
results where contractual relationships were weak or nonexistent. How,
then, can the reach of index insurance be extended to benefit larger
groups of farmers on the boundaries or outside the formal agricultural
marketing chain?
Major lessons pertaining to the scalability of pilot programs remain to
be learned. These lessons, however, will only be learned if efforts are
taken to expand the reach of index insurance products beyond the con-
fines of the pilot programs. Future pilot programs should therefore set
ambitious goals to substantially expand the reach of index insurance to
include poorer farmers without strong ties to the agricultural marketing
chain, to span a greater variety of crops and production practices, and to
cover wider geographical areas. Efforts should also be made to develop an
institutional framework that includes a wider variety of agricultural
market participants, including a combination of banks, input suppliers,
processors, and exporters. Efforts should also be made to integrate risk
management work into larger agricultural investment programs and
partnerships.
Efforts to expand beyond the pilot programs are sure to reveal new and
unanticipated problems in index insurance market development for which
solutions will have to be found to substantially increase the number of
beneficiaries. Index insurance market expansion is also likely to draw
increased interest from private insurers and reinsurers who are cautiously
waiting to see whether the market for index insurance can reach sufficient
volume to make it a profitable venture worth investing in on a larger
scale.
423
on 01 January 2018
References
Akerlof, G.A. 1970. The Market for Lemons: Quality Uncertainty and the Market
Mechanism. The Quarterly Journal of Economics 84(3):488–500.
Bardsley, P., A. Abey, and S. Davenport. 1984. The Economics of Insuring Crops
against Drought. The Australian Journal of Agricultural Economics 28(1):1–14.
Barrett, C.B., M.R. Carter, S. Chantarat, M. Ikegami, and J. McPeak. 2009. Index
Based Livestock Insurance for Northern Kenya’s Arid and Semi-Arid Lands The
Marsabit Pilot. Project summary, International Livestock Research Institute,
Nairobi, Kenya.
Binswanger-Mkhize, H.P. 2011. Is There Too Much Hype About Index-Based
Agricultural Insurance? Unpublished manuscript, Institute for Economic
Research on Innovation, Tshwane University of Technology, Pretoria, South
Africa. Presented at the Workshop on Weather-Index Insurance for Smallholder
Farmers in Africa, Cape Town, South Africa, September, 2010.
Bourgeon, J.-M., and R.G. Chambers. 2003. Optimal Area Yield Insurance
Reconsidered. American Journal of Agricultural Economics 85(3):590– 604.
Breustedt, G., R. Bokusheva, and O. Heidelbach. 2008. Evaluating the Potential of
Index Insurance Schemes to Reduce Crop Yield Risk in an Arid Region. Journal
of Agricultural Economics 59(2):312–328.
Bryla, E. 2009. Weather Risk Management: An Ethiopian Pilot. Experiential brief-
ing note, World Bank, Agricultural and Rural Development Department,
Commodity Risk Management Group, Washington, DC.
Bryla, E., and J. Syroka. 2007. Developing Index-Based Insurance for Agriculture in
Developing Countries. Innovation brief, United Nations, Department of
Economic and Social Affairs, New York.
Carter, M.R., C.B. Barrett, S. Boucher, S. Chantarat, F. Galarza, J. McPeak, A. Mude,
and C. Trivelli. 2008. Insuring the Never Before Insured: Explaining Index
Insurance through Financial Education Games. BASIS Brief 2008-7, BASIS
Research Program, University of Wisconsin-Madison, Madison, WI.
Carter, M.R., S.R. Boucher, and C. Trivelli. 2007. Area-Based Yield Insurance Pilot
Project For Peruvian Coastal Agriculture. Concept note, BASIS Research
Program, University of Wisconsin-Madison, Madison, WI.
Carter, M.R., and S.A. Janzen. 2012. Coping with Drought: Assessing the Impacts
of Livestock Insurance in Kenya. Brief No.2012-1, Index Insurance Innovation
Initiative, Davis, CA.
Chambers, R.G. 1989. Insurability and Moral Hazard in Agricultural Insurance
Markets. American Journal of Agricultural Economics 71(3):604–616.
Chantarat, S., C.B. Barrett, A.G. Mude, and C.G. Turvey. 2007. Using Weather
Index Insurance to Improve Drought Response for Famine Prevention. American
Journal of Agricultural Economics 89(5):1262 –1268.
Chantarat, S.M., A.G. Mude, C.B. Barrett, and M.R. Carter. 2012. Designing
Index-Based Livestock Insurance for Managing Asset Risk in Northern Kenya.
Journal of Risk and Insurance doi:10.1111/j.1539-6975.2012.01463.x.
Clarke, D., and S. Dercon. 2009. Insurance, Credit and Safety Nets for the Poor in a
World of Risk. DESA Working Paper 81, University of Oxford, Department of
Economics, Oxford, UK.
Coate, S., and M. Ravallion. 1993. Reciprocity without Commitment -
Characterization and Performance of Informal Insurance Arrangements. Journal
of Development Economics 40(1):1– 24.
Doherty, N.A., and A. Richter. 2002. Moral Hazard, Basis Risk, and Gap Insurance.
Journal of Risk and Insurance 69(1):9 –24.
Dubois, P., B. Jullien, and T. Magnac. 2008. Formal and Informal Risk Sharing in
LDCs: Theory and Empirical Evidence. Econometrica 76(4):679–725.
Giné, X. 2009. Innovations in Insuring the Poor: Experience with Weather
Index-Based Insurance in India and Malawi. Brief 7, International Research
Institute for Climate and Society, Palisades, NY.
424
on 01 January 2018
Giné, X., R.M. Townsend, and J. Vickery. 2007. Statistical Analysis of Rainfall
Insurance Payouts in Southern India. American Journal of Agricultural Economics
89(5):1248– 1254.
———. 2008. Patterns of Rainfall Insurance Participation in Rural India. World Bank
Economic Review 22(3):539–566.
Giné, X., and D. Yang. 2009. Insurance, Credit, and Technology Adoption Field
Experimental Evidence from Malawi. Journal of Development Economics
89(1):1 –11.
GlobalAgRisk, Inc. 2006. Hedging Weather Risk for Microfinance Institutions in
Peru: Comprehensive Report. Report to the United States Agency for
International Development, GlobalAgRisk, Inc., Lexington, KY.
———. 2009. Designing Agricultural Index Insurance in Developing Countries: A
GlobalAgRisk Market Development Model Handbook for Policy and Decision
Makers. Technical report, GlobalAgRisk, Inc., Lexington, KY.
———. 2010. GlobalAgRisk Projects in Vietnam, Peru, and Mongolia: Four Case
Studies. Technical report, GlobalAgRisk, Inc., Lexington, KY.
Gollier, C. 2003. To Insure or Not to Insure? An Insurance Puzzle. The Geneva
Papers on Risk and Insurance - Theory 28(1):5–24.
Halcrow, H.G. 1949. Actuarial Structures for Crop Insurance. Journal of Farm
Economics 31(3):418– 443.
Hazell, P. 1992. The Appropriate Role of Agricultural Insurance in Developing
Countries. Journal of International Development 4(6):567–581.
Hess, U., W. Wiseman, and T. Robertson. 2006. Ethiopia: Integrated Risk Financing
to Protect Livelihoods and Foster Development. Discussion paper, United
Nations, World Food Programme, Rome, Italy.
Holmstrom, B. 1979. Moral Hazard and Observability. Bell Journal of Economics
10(1):74–91.
Horowitz, J.K., and E. Lichtenberg. 1993. Insurance, Moral Hazard, and Chemical
Use in Agriculture. American Journal of Agricultural Economics 75(4):926–935.
Joshi, V. 2010. Index-Based Insurance: BASIX Experience. Unpublished presenta-
tion, International Livestock Research Institute, Addis Adaba, Ethiopia.
Presented at the Workshop on Developing Index-Based Livestock Insurance to
Reduce Vulnerability due to Drought-related Livestock Deaths, International
Livestock Research Institutue, Addis Ababa, Ethiopia, 12 July 2010.
Just, R.E., L. Calvin, and J. Quiggin. 1999. Adverse Selection in Crop Insurance:
Actuarial and Asymmetric Information Incentives. American Journal of
Agricultural Economics 81(4):834–849.
Karlan, D., E. Kutsoati, M. McMillan, and C. Udry. 2011. Crop Price Indemnified
Loans for Farmers: A Pilot Experiment in Rural Ghana. Journal of Risk and
Insurance 78(1):37–55.
Khalil, A.F., H.H. Kwon, U. Lall, M.J. Miranda, and J.R. Skees. 2007. El
Niño-Southern Oscillation-Based Index Insurance for Floods: Statistical Risk
Analyses and Application to Peru. Water Resources Research 43(10):W10416.
Lagos, P., Y. Silva, E. Nickl, and K. Mosquera. 2008. El Niño -Related Precipitation
Variability in Peru. Advances in Geosciences 14(1):231–237.
Leiva, A.J., and J.R. Skees. 2008. Using Irrigation Insurance to Improve Water
Usage of the Rio Mayo Irrigation System in Northwestern Mexico. World
Development 36(12):2663–2678.
Ligon, E., J.P. Thomas, and T. Worrall. 2002. Informal Insurance Arrangements
with Limited Commitment: Theory and Evidence from Village Economies. The
Review of Economic Studies 69(1):209– 244.
Luo, H., J.R. Skees, and M.A. Marchant. 1994. Weather Information and the
Potential for Intertemporal Adverse Selection in Crop Insurance. Review of
Mahul, O. 1999a. The Design of an Optimal Area Yield Crop Insurance Contract.
The Geneva Papers on Risk and Insurance - Theory 24(2):159–171.
425
on 01 January 2018
———. 1999b. Optimum Area Yield Crop Insurance. American Journal of Agricultural
Economics 81(1):75–82.
Mahul, O., and J.R. Skees. 2007. Managing Agricultural Risk at the Country Level:
The Case of Index-Based Livestock Insurance in Mongolia. Policy Research
Working Paper 4325, World Bank, Washington, DC.
Mahul, O., and C.J. Stutley. 2010. Government Support to Agricultural Insurance:
Challenges and Options for Developing Countries. Technical report, World
Bank, Private Sector Development Vice Presidency, Global Capital Markets
Development Department, Washington, DC.
Makaudze, E.M., and M.J. Miranda. 2010. Catastrophic Drought Insurance Based
on Remotely Sensed Normalized Difference Vegetation Index for Smallholder
Farmers in Zimbabwe. Agrekon 49(1):418–432.
Makki, S.S., and A. Somwaru. 2001. Evidence of Adverse Selection in Crop
Insurance Markets. Journal of Risk and Insurance 68(4):685–708.
Manuamorn, O. Pomme, and W. Dick. 2009. Flood Risk Management: Feasibility
Research on Index-Based Flood Products in Thailand and Vietnam. Experiential
briefing note, World Bank, Agricultural and Rural Development Department,
Commodity Risk Management Group, Washington, DC.
McPeak, J.G., S. Chantarat, and A. Mude. 2005. Explaining Index Based Livestock
Insurance to Pastoralists. Agricultural Finance Review 70(3):333–352.
Mechler, R., J. Linnerooth-Bayer, and D. Peppiat. 2006. Microinsurance for Natural
Disaster Risks in Developing Countries: Benefits, Limitations and Viability.
Technical report, ProVention Consortium/IIASA, Geneva, Switzerland.
Miranda, M.J. 1991. Area-Yield Crop Insurance Reconsidered. American Journal of
Miranda, M.J., and P.L. Fackler. 2002. Applied Computational Economics and Finance.
Boston, MA: MIT Press.
Miranda, M.J., and J.W. Glauber. 1997. Systemic Risk, Reinsurance, and the Failure
of Crop Insurance Markets. American Journal of Agricultural Economics
79(1):206–215.
Miranda, M.J., and C. Gonzalez-Vega. 2011. Systemic Risk, Index Insurance, and
Optimal Management of Agricultural Loan Portfolios in Developing Countries.
American Journal of Agricultural Economics 93(2):399–406.
Miranda, M.J., and D.V. Vedenov. 2001. Innovations in Agricultural and Natural
Disaster Insurance. American Journal of Agricultural Economics 83(3):650–655.
Morduch, J. 2006. Micro-Insurance: The Next Revolution? Oxford University Press,
337 –356. Chapter 22. Oxford, U.K.
Mosley, P. 2009. Assessing the Success of Microinsurance Programmes in Meeting
the Insurance Needs of the Poor. DESA Working Paper No. 84, United Nations,
Department of Economic and Social Affairs, New York.
Mude, A., S. Chantarat, C. Barrett, M.R. Carter, M. Ikegami, and J. McPeak. 2009.
Insuring Against Drought-Related Livestock Mortality: Piloting Index Based
Livestock Insurance in Northern Kenya. Unpublished manuscript, International
Livestock Research Institute, Nairobi, Kenya.
Osgood, D.E., et al. 2007. Designing Weather Index Insurance for Farmers in
Malawi, Tanzania, and Kenya. Technical report, International Research Institute
for Climate and Society, Palisades, NY. Final report to the Commodity Risk
Management Group, Agricultural and Rural Development, World Bank,
Washington, DC.
Osgood, D.E., P. Suarez, J. Hansen, M. Carriquiry, and A. Mishra. 2008. Integrating
Seasonal Forecasts and Insurance for Adaptation Among Subsistence Farmers:
The Case of Malawi. Policy Research Working Paper 4651, World Bank,
Washington, DC.
Patt, A., N. Peterson, M. Carter, M. Velez, U. Hess, and P. Suarez. 2009. Making
Index Insurance Attractive to Farmers. Mitigation and Adaptation Strategies for
Global Change 14(8):737 –753.
426
on 01 January 2018
Quiggin, J., G. Karagiannis, and J. Stanton. 1993. Crop Insurance and Crop
Production: An Empirical Study of Moral Hazard and Adverse Selection.
Australian Journal of Agricultural Economics 37(2):95–113.
Santos, P., and C.B. Barrett. 2011. Persistent Poverty and Informal Credit. Journal of
Development 96(2):337– 347.
Shavell, S. 1979. On Moral Hazard and Insurance. The Quarterly Journal of
Economics 93(4):541– 562.
Skees, J.R. 2000. A Role for Capital Markets in Natural Disasters: A Piece of the
Food Security Puzzle. Food Policy 25(3):365–378.
———. 2008a. Challenges for Use of Index-Based Weather Insurance in Lower
Income Countries. Agricultural Finance Review 68(1):197–217.
———. 2008b. Innovations in Index Insurance for the Poor in Lower Income
Countries. Agricultural and Resource Economics Review 37(1):1– 15.
Skees, J.R., S. Gober, P. Varangis, R. Lester, and V. Kalavakonda. 2001. Developing
Rainfall-Based Index Insurance in Morocco. Policy Research Working Paper
2577, World Bank, Washington, DC.
Skees, J.R., J. Hartell, and A.G. Murphy. 2007. Using Index-Based Risk Transfer
Products to Facilitate Microlending in Peru and Vietnam. Report to the United
States Agency for International Development. GlobalAgRisk, Inc., Lexington,
KY.
Skees, J.R., and A.G. Murphy. 2009. ENSO Business Interruption Index Insurance
for Catastrophic Flooding in Piura, Peru. Technical report, GlobalAgRisk, Inc.,
Lexington, KY.
Skees, J.R., and M.R. Reed. 1986. Rate Making for Farm-Level Crop Insurance:
Implication for Adverse Selection. American Journal of Agricultural Economics
68(3):653–659.
Skees, J.R., P. Varangis, D. Larson, and P. Siegel. 2002. Can Financial Markets be
Tapped to Help Poor People Cope with Weather Risks? Policy Research
Working Paper 2812, World Bank, Washington, DC.
Smith, V.H., and B.K. Goodwin. 1996. Crop Insurance, Moral Hazard, and
Agricultural Chemical Use. American Journal of Agricultural Economics
78(2):428–438.
Townsend, R.M. 1994. Risk and Insurance in Village India. Econometrica
62(3):539–591.
Vercammen, J., and G.C. van Kooten. 1994. Moral Hazard Cycles in
Individual-Coverage Crop Insurance. American Journal of Agricultural Economics
76(2):250–261.
World Bank. 2011. Project Paper on a Proposed Additional Credit to Mongolia for
the Second Sustainable Livelihoods Project. Report No. 61628-MN, World Bank,
Washington, DC.
World Food Programme. 2006. World’s First Humanitarian Insurance Policy
Issued. Available at http://www.wfp.org. Accessed on July 29, 2012.
427
on 01 January 2018

Index Insurance For Developing Countries

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Index Insurance For Developing Countries

Uploaded by

Copyright:

Available Formats

Applied Economic Perspectives and Policy (2012) volume 34, number 3, pp. 391–427.

Mario J. Miranda is the Andersons Professor of Finance and Risk Management

Commissioned April 2011; accepted July 2012.

Abstract Unlike conventional insurance, which indemnifies policyholders for veri-

more adversely selected pool of participants. Adverse selection problems

insurance contracts are especially difficult to design in areas that exhibit

Simple Model of Index Insurance Demand

where m is expected income, b ≥ 0, Ez̃ = E1̃ = 0, and z̃ and 1̃ are inde-

max u0 (w − px) + dEu1 (m + bz̃ + 1̃ + xh(z̃)). (2)

p , p∗ ; El̃(z̃, 1̃)h(z̃) (3)

differentiable, strictly increasing function of the income-index covariance

V(w) = max {u(w − s − px) + dEV(ỹ + (1 + r)s + xh(z̃))}. (5)

Figure 1 Optimal index insurance coverage as a function of wealth

utility of consumption function exhibits constant relative risk aversion 2;

Figure 2 Mean long-run index insurance coverage as a function of the index/income

make savings more attractive, leading to higher buffer stocks of savings

Index Insurance Pilot Programs

Guatemala, Honduras, India, Indonesia, Jamaica, Kenya, Malawi, Mexico,

The BASIX weather insurance scheme, which was initially piloted in

higher-yielding crops has not been observed among purchasers of the

services regulator, is expected to open in January of 2012, with the rele-

livestock insurance (IBLI) pilot in the Marsabit district in northern Kenya

households, will be used to analyze the welfare effects of IBLI. A recent

6% to 30% in the base insurance product (BIP). Catastrophic losses of 30%

agricultural insurance education workshops, as well as the preparation of

Micro, Meso, or Macro?

purchase and indemnity collection. These costs can be reduced by using

Meso-index insurance products, however, present very different techni-

reconstruction of local infrastructure, the implementation of temporary

Index Insurance and Credit

helps frame education, contract design, and market delivery development

joint liability groups or buyer agreements. Without such measures, bun-

Other Lessons Learned

constrained by the lack of sophistication of their intended users. Due to a

at nearby ground stations, or to measurements obtained through remote-

the creation of strong institutional partnerships, as the leader can serve as

the same educational methods will be transferable to the education of

The Way Forward

Perform Impact Analysis

insurance into holistic business risk management practices, or have they

Measure Basis Risk

Promote Portfolio Risk Management

Establish Index Insurance Data Center

market development in new venues throughout the world. Annual actua-

Attempt Market Expansion

You might also like