Index Insurance Products Research Articles

The potential of machine learning for weather index insurance

Abstract. Weather index insurance is an innovative tool in risk transfer for disasters induced by natural hazards. This paper proposes a methodology that uses machine learning algorithms for the identification of extreme flood and drought events aimed at reducing the basis risk connected to this kind of insurance mechanism. The model types selected for this study were the neural network and the support vector machine, vastly adopted for classification problems, which were built exploring thousands of possible configurations based on the combination of different model parameters. The models were developed and tested in the Dominican Republic context, based on data from multiple sources covering a time period between 2000 and 2019. Using rainfall and soil moisture data, the machine learning algorithms provided a strong improvement when compared to logistic regression models, used as a baseline for both hazards. Furthermore, increasing the amount of information provided during the training of the models proved to be beneficial to the performances, increasing their classification accuracy and confirming the ability of these algorithms to exploit big data and their potential for application within index insurance products.

Mapping weather risk – A multi-indicator analysis of satellite-based weather data for agricultural index insurance development in semi-arid and arid zones of Central Asia

Index insurance has been introduced as a solution to tackle several challenges that prevail in the agricultural insurance sector of developing countries. One of the main implementation challenges in these countries is the lack of reliable weather data for index development and implementation. The increasing availability of satellite data could ease the constraints of data access. Meanwhile, the suitability of various satellite products for yield estimation across world regions has to undergo a thorough assessment. This study contributes to the literature by systematically analyzing the accuracy of some globally available satellite data, namely the Global Satellite Mapping of Precipitation (GSMaP), Climate Hazards Group InfraRed Precipitation with Station (CHIRPS), and the Global Land Data Assimilation System (GLDAS) compared to ground-level weather information for 14 different indicators for the case of Uzbekistan. Our analysis indicates that those sources may provide the necessary data for an accessible and adequate climate service. However, a considerable risk of overestimation and underestimation depending on the source of satellite data may exist, especially for precipitation data in the conditions of Central Asia. Among the tested datasets, GSMaP showed a relatively better performance than CHIRPS in precipitation estimation for drought and flood detection. In order to reduce detection inaccuracy, the application of satellite weather products for index insurance is possible when temporal aggregation (e.g., monthly, seasonal) is considered. Globally available climate data could serve as a good source to establish index insurance products in Central Asia; however, a careful selection of source and index is required.

Parametric and non-parametric estimates of willingness to pay for weather index insurance in Malawi

As climate-related shocks on agriculture production intensify, weather index crop insurance has emerged as one potential risk management strategy for farmers. Nevertheless, sustainable uptake of this risk management strategy will largely depend on farmers’ willingness to pay for the associated premiums. This study identifies various socio-economic factors that influence farmers' purchase of index insurance for staple food in rural Malawi. It further establishes an associated willingness to pay per hectare per cropping season. We use a hypothetical rainfall index insurance program to estimate insurance demand. The study highlights several determinants of willingness to pay for an index insurance product. Results suggest that gender, the previous record of climate shocks, extension contact, and access to remittances significantly influence willingness-to-pay for weather index insurance for a staple food crop. Both parametric and non-parametric estimates of willingness-to-pay are in the range of US$5.5 to US$15 per hectare per cropping season. At the infancy stage, government subsidies for insurance premium and linkage of premium payments to Village Savings Groups will be crucial.

СТИМУЛИРОВАНИЕ АГРОПРЕДПРИЯТИЙ К АГРОСТРАХОВАНИЮ С ГОСПОДДЕРЖКОЙ В УСЛОВИЯХ ЦИФРОВИЗАЦИИ СЕЛЬСКОГО ХОЗЯЙСТВА

Considering development of the agroinsurance market, the author focuses attention to its potential opportunities taking into account increased pandemic situation. Economic consequences of Covid19 and measures undertaken by the state change methods of management of agrobusiness, the accelerated rates carry out transition to digitalization. In this connection in article the variation of application of digitalization in agroinsurance as stimulation of the agroenterprises to the conclusion of agricultural contracts with the state support is considered, the international experience of eco- systems of the insurance market is presented and the prospects of development of digital agroinsurance taking into account potential opportunities of participants of the agroinsurance market are considered, the main aspects of development of model of digital agroinsurance with the state support are proposed, offers on transformation of the operating agroinsurance system concerning index insurance products for the agroenterprises are given. Processes of digitalization in agroinsurance are connected with the Digital Agriculture project which implementation the Ministry of Agriculture of the Russian Federation began in 2019 that represents the practical importance for increase in coverage agroinsurance of the domestic agroenterprises. The model of digital agroinsurance proposed in article is offered to consideration by the author on the example of the Blockchain platform with participation of the state and insurance community which will allow agroinsurers to develop products ta- king into account features and needs of regions, and for landowners to make insurance moreconvenient and attractive. The model means use of space monitoring and remote customer service at all stages: conclusion and supportof the contract of agricultural insurance, settlement of a loss.

A new framework to quantify maize production risk from chilling injury in Northeast China

Agricultural production frequently suffers from meteorological risks. Timely and accurately estimating yield losses will greatly improve disaster risk management, especially for a useful tool to transfer risks—agricultural insurance. The weather index insurance products have increasingly become promising alternatives to traditional ones. However, the current single-index insurances, mainly based on a statistical model with a lower explanation ability and at an aggregate scale (e.g. county, province), are lack of mechanistic processes and having huge basis risk. Here, we proposed a new framework to quantify maize production risk by running the well calibrated CERES-Maize driven by various chilling scenarios. We first analyzed the chilling risk across Northeast China, then developed a composite chilling index (CCI) and CCI-based vulnerability curves to derive the corresponding yield losses. Finally, the specific pure premium rates were calculated respectively at 16 sites across the studied areas. The results showed that composite-index insurance (RMSE¯ ≤ 721 kg/ha) outperformed (R2 improved by 35%) every single index (RMSE¯ ≈ 1140 kg/ha) in quantifying yield loss. We believe our proposed framework will benefit insurer companies rapidly estimating yield losses and designing more promising products, and potentially apply into other regions for other crops and hazards to transfer climate risk in the future.

Targeting weather insurance markets

AbstractThe suitability of insurance products often depends greatly on individual circumstances. This paper examines the challenges of heterogeneity in a relatively new product, weather‐indexed insurance. This index insurance product has been launched in over a dozen countries, with the goal of enabling households engaged in agricultural activity a means to manage risk. Using data from a large‐scale field experiment, we build and calibrate a model which accounts for household investment decisions, including the scope for self‐insurance via labor markets to (risky) wage work. Our results show that insurance is most valuable to households with reduced access to wage labor, or to those who face wages that are sensitive to rainfall risk. These findings have important implications for areas where index insurance is most effective.

Can Experiential Games and Improved Risk Coverage Raise Demand for Index Insurance? Evidence from Kenya

Uninsured risk impedes agricultural production, but traditional indemnity insurance is not a viable option for smallholder farmers due to market failures. Weather index insurance is often touted as an alternative risk management tool. Demand for weather index insurance, however, has been weak. One possible reason is basis risk, the mismatch between index insurance payouts and individual outcomes. We estimate the impact of two interventions on preferred coverage under a real index insurance product, as revealed through an auction: (a) a reduction in basis risk, and (b) an experiential game that teaches farmers how index insurance works, with an emphasis on basis risk. We show all farmers demonstrate strong sensitivity to basis risk. The experiential game modestly increased knowledge, and experiential game participants indicate higher levels of preferred insurance coverage. Although offering a lower basis risk insurance product and playing an experiential game both increase preferred coverage in isolation, there is no additional impact of doing both. We adapt a theoretical model of misattribution bias to demonstrate how reference‐dependent learning with loss aversion could lead to this unexpected result.

Targeting Weather Insurance Markets

The suitability of insurance products often depends greatly on individual circum- stances. This paper examines the challenges of heterogeneity in a relatively new product, weather-indexed insurance. This index insurance product has been launched in over a dozen countries, with the goal of enabling households engaged in agricultural ac- tivity a means to manage risk. Using data from a large-scale field experiment, we build and calibrate a model which accounts for household investment decisions, including the scope for self-insurance via labor markets to (risky) wage work. Our results show that insurance is most valuable to households with reduced access to wage labor, or to those who face wages that are sensitive to rainfall risk. These findings have important implications for areas where index insurance is most effective.

Identifying Precipitation and Reference Evapotranspiration Trends in West Africa to Support Drought Insurance

West Africa represents a wide gradient of climates, extending from tropical conditions along the Guinea Coast to the dry deserts of the south Sahara, and it has some of the lowest income, most vulnerable populations on the planet, which increases catastrophic impacts of low and high frequency climate variability. This paper investigates low and high frequency climate variability in West African monthly and seasonal precipitation and reference evapotranspiration from the early 1980s to 2016. We examine the impact of those trends and how they interact with payouts from index insurance products. Understanding low and high frequency variability in precipitation and reference evapotranspiration at these scales can provide insight into trends during periods critical to agricultural performance across the region. For index insurance, it is important to identify low-frequency variability, which can result in radical departures between designed/planned and actual insurance payouts, especially in the later part of a 30-year period, a common climate analysis period. We find that evaporative demand and precipitation are not perfect substitutes for monitoring crop deficits and that there may be space to use both for index insurance design. We also show that low yields—aligned with the need for insurance payouts—can be predicted using classification trees that include both precipitation and reference evapotranspiration.

Using a regional climate model to develop index-based drought insurance for sovereign disaster risk transfer

PurposeExamine the usability of rainfall and temperature outputs of a regional climate model (RCM) and meteorological drought indices to develop a macro-level risk transfer product to compensate the government of Central Java, Indonesia, for drought-related disaster payments to rice farmers.Design/methodology/approachBased on 0.5° gridded rainfall and temperature data (1960–2015) and projections of the WRF-RCM (2016–2040), the Standardized Precipitation Index (SPI) and the Standardized Precipitation Evapotranspiration Index (SPEI) are calculated for Central Java over different time spans. The drought indices are correlated to annual and seasonal rice production, based on which a weather index insurance structure is developed.FindingsThe six-month SPI correlates best with the wet season rice production, which generates most output in Central Java. The SPI time series reveals that drought severity increases in future years (2016–2040) and leads to higher payouts from the weather index structure compared to the historical period (1960–2015).Practical implicationsThe developed methodology in using SPI for historical and projected periods allows the development of weather index insurance in other regions which have a clear link between rainfall deficit and agricultural production volatility.Originality/valueMeteorological drought indices are a viable alternative for weather index insurance, which is usually based on rainfall amounts. RCM outputs provide valuable insights into future climate variability and drought risk and prolong the time series, which should result in more robust weather index insurance products.

Demand heterogeneity for index-based insurance: The case for flexible products

A substantial literature has analyzed the challenges around weather index insurance, yet an important design issue has been generally overlooked. Most index insurance products have so far been characterized by a one-size-fits-all payout structure, intended for a representative farmer, at the cost of ignoring considerable heterogeneity in risk profiles. This paper provides unique evidence on the ways in which heterogeneity in farmers’ risk exposure affects their demand for insurance. We analyze a set of flexible insurance products against excess rainfall and exploit the substantial variation in insurance portfolios demanded by farmers. We explore the relevance of alternative sources of heterogeneity by extending a simple expected utility decision model and relying on structural estimation to test their significance. We find important aspects of farmer heterogeneity directly affecting their demand for insurance. We quantify the benefits of a flexible scheme by comparing farmer welfare to that achieved under alternative counterfactual insurance options.

Will urban migrants formally insure their rural relatives? Family networks and rainfall index insurance in Burkina Faso

Abstract Large segments of the population in developing countries, especially in rural areas, have a high level of vulnerability to weather-related shocks but have limited means to insure themselves against them. In recent years, microfinance institutions have experimented with micro-insurance products, including rainfall index insurance, but the uptake of these products has generally been very low. We present findings from a pilot study exploring whether and how existing ties between urban migrants and rural farmers may be used to provide the latter improved access to formal insurance. Urban migrants in Ouagadougou (the capital of Burkina Faso) originating from nearby villages were offered, at the prevailing market price, a rainfall index insurance product that can potentially protect their rural relatives from adverse weather shocks. The product had an uptake of 22% during the two-week subscription window. Half the urban migrants in the study were randomly offered an insurance policy in which payouts would be made, not to the subscriber, but directly to the intended beneficiary. This feature increased uptake rates by 17–22 percentage points. We argue that rainfall index insurance can complement informal risk-sharing networks by mitigating problems of informational asymmetry and self-control issues.

Drought-tolerant rice, weather index insurance, and comprehensive risk management for smallholders: evidence from a multi-year field experiment in India.

In rainfed production systems throughout India, agricultural activities are dependent upon the summer monsoon, and any aberration in monsoon rainfall patterns can have severe consequences for rice production. There is considerable policy interest in designing programs to lower small-scale farmers’ exposure to these types of risk given the regularity with which adverse monsoon events occur. This paper introduces a field experiment conducted with two risk management options in the state of Odisha: a drought-tolerant rice cultivar; and a weather index insurance product designed to complement the performance of the cultivar. Uptake rates for the cultivar itself and for the joint product are compared across two years alongside an analysis of factors that predict uptake. Results indicate high levels of demand for both the products, albeit with a significant degree of price sensitivity. But this sensitivity is agnostic to the nature of price reductions, suggesting that public investments that lower the costs of risk management may be sufficient to encourage broad uptake, without necessarily relying upon distortionary subsidies as is so often done. Sustained demand between years one and two is primarily explained where individuals were indemnified in year one and had a large number of peers also purchasing the product.

Protection against risks of losing a harvest of agricultural crops by means of applying the data obtained from the space monitoring systems of plant cover

Recently, in the domestic practice of agricultural insurance, interest in index insurance products has increased from both insurers and insurants. This is due to the high efficiency demonstrated by them as a tool to protect against the risk of yield loss, with a low level of associated moral hazard, low maintenance costs of the insurance contract. Index insurance implies the conclusion of a contract providing for the payment of insurance compensation when the base indicator reaches a specified threshold, the base indicator being an indirect sign of the occurrence of actual damage. When insuring crops, such a baseline can be high temperatures, lack of precipitation during the growing season. However, a more accurate indicator of the magnitude of the crop, and hence its shortage, is the normalized differential vegetation index (NDVI), obtained using space monitoring of vegetation. This fact is a reliable prerequisite for the use of this indicator in the index insurance of agricultural crops. Mathematical modeling of the dependence of the yield of spring barley in certain districts of the Rostov region on NDVI indicators made it possible to determine the main parameters of the corresponding index insurance products and to prove the effectiveness of their use to reduce the risks of yield loss. In particular, efficiency is expressed in stabilizing the income of agricultural producers by years, which is confirmed by a decrease in the coefficients of variation of this indicator in some cases to 27%.

Design of Temperature Insurance Index and Risk Zonation for Single-Season Rice in Response to High-Temperature and Low-Temperature Damage: A Case Study of Jiangsu Province, China.

Disaster insurance is an important tool for achieving sustainable development in modern agriculture. However, in China, the design of such insurance indexes is far from sufficient. In this paper, the single-season rice in Jiangsu Province of China is taken as an example to design the high-temperature damage index in summer and the low-temperature damage index in autumn to construct the formula calculating the weather output and single-season rice yield reduction. The daily highest, lowest and average temperatures between 1999 and 2015 are selected as main variables for the temperature disaster index to quantitatively analyze the relationship between the temperature index and the yield reduction rate of the single-season rice. The temperature disaster index can be put into the relevant model to obtain the yield reduction rate of the year and determine whether to pay the indemnity. Then, the burn analysis is used to determine the insurance premium rate for all cities in Jiangsu Province under four-level deductibles, and the insurance premium rate can be used for the risk division of the Province. The research provides some insights for the design of agricultural weather insurance products, and the empirical results provide a reference for the design of similar single-season rice temperature index insurance products.

Index insurances for grasslands – A review for Europe and North-America

Grassland based farming systems are exposed to extreme weather events causing volatile farm incomes. Grazing and lacking yield measurements make it largely impossible to insure grassland production with traditional insurance products. In contrast, index insurance products have the potential to insure grasslands, as their payoff relies on an endogenous index that is highly correlated to, but independent of, the actual grass yield. To support future development of these products, we provide the first systematic overview of 12 index insurances put into practise for grasslands in Europe and North America. Additionally, based on this overview, we present prevailing findings that are important for further research and insurance practitioners. We find that a large diversity of index insurance types is applied in practise, including insurance solutions based on regional yield levels, weather variables or satellite imagery. We reveal separated insurance markets (i.e. country-specific products), which prevent knowledge spillovers and lead to largely isolated product developments. Thus, grassland insurance schemes can be improved by knowledge exchange and combining methods that are applied elsewhere. More specifically, insurances tailored to single farm's risk exposure, the combination of satellite with other geodata (e.g. land use information) or adapting legal specifications that disadvantage some types of insurances can improve an insurance's risk reducing capacity and make grassland based farming systems more resilient to weather extremes. This paper provides an entry point for such process, ensuring the development of efficient measures for farmers to cope with climatic risks.

Ex ante and ex post effects of hybrid index insurance in Bangladesh

This study assesses both the demand for and effectiveness of an index insurance product designed to help smallholder farmers in Bangladesh manage crop production risk during the monsoon season. Villages were randomized into either an insurance treatment or a comparison group, and discounts and rebates were randomly allocated across treatment villages to encourage insurance take-up and to allow for the estimation of the price-elasticity of insurance demand. Among those offered insurance, we find demand to be fairly price elastic, with discounts significantly more successful in stimulating demand than rebates. Purchasing insurance yields both ex ante risk management effects as well as ex post income effects on agricultural production practices. The risk management effects lead to an expansion of cultivated area with concomitant increases in agricultural input expenditures during the monsoon season. The income effects lead to more intensive rice production during the subsequent dry season, with more intensive use of both irrigation and fertilizers, resulting in higher yields and higher total rice production.

Does Index Insurance Crowd In or Crowd Out Informal Risk Sharing? Evidence from Rural Ethiopia

AbstractWe study how the introduction of a formal index insurance product affects informal risk‐sharing among pastoralists in southern Ethiopia. Using detailed social networks data, randomized incentives to purchase the insurance product, and hypothetical informal transfer data that mirror the existing customary arrangements, we find respondents’ own formal insurance uptake has no significant effect on their willingness to share risk through customary institutions. We also find weak evidence that a randomly matched peer's insurance uptake positively influences respondents’ willingness to make informal transfers to that match. Overall, our results imply that in this context index insurance does not crowd out informal risk‐sharing mediated by social networks.

One size does not fit all: Universal livelihood insurance in St. Lucia

In 2015 the G7 countries announced the InsuResilience project to quadruple, by 2020, the number of people in the Global South who have financial risk management tools against climate change. A principal way this will occur is through weather index insurance (WII), a type of insurance that pays out according to estimated losses rather than actual ones. In this paper, I examine one WII product that has been for sale in St. Lucia since May 2013, called the Livelihood Protection Policy, which has been discussed as a model WII product for InsuResilience. This article is based on twelve months of fieldwork conducted in St. Lucia, interviewing people involved in the design, implementation, and distribution of LPP, and actual and potential purchasers of LPP. In this article, I argue that LPP's design as a microinsurance product that can be purchased by any St. Lucian no matter their livelihood has made it largely undesirable because it does not fit anyone's needs well. This occurs for two main reasons. First, rather than primarily being used to help low-income St. Lucians adapt to climate change, LPP was used to test out how an abstract index insurance product can be sold without directly subsidizing the premiums of policyholders. Second, by not being designed for particular livelihoods, LPP suffers from not addressing hazards, such as drought, that many St. Lucians are highly exposed to, and from not compensating policyholders at meaningful wind and rainfall thresholds.

New modalities for managing drought risk in rainfed agriculture: Evidence from a discrete choice experiment in Odisha, India

Droughts have historically had large impacts on agricultural production in rainfed agricultural systems. Scientists and policymakers have proposed various strategies for managing risks, with limited success. In this study we consider two such strategies, specifically drought tolerant rice and weather index insurance. While neither drought tolerant cultivars nor weather index insurance products are perfect solutions for adequately managing drought risk in and of themselves, there is scope to exploit the benefits of each and bundle them into a complementary risk management product, specifically through proper index calibration and an optimized insurance design. In this study, we explore preferences for such a complementary risk management product using discrete choice experiments in Odisha, India. We are able to estimate the added value that farmers perceive in the bundled product above and beyond the value associated with each of the independent products. We also show that valuations are very sensitive to the basis risk implied by the insurance product, with farmers less enthusiastic about risk management products that leave significant risks uninsured.