Climate risk adaptation through livestock insurance: evidence from a pilot programme in Nigeria

ABSTRACT

This study investigates the option of choosing index-based livestock insurance (IBLI) to mitigate the adverse effects associated with climate change in Kwara State, Nigeria. Previous studies indicate that the failure to include farmers during the early stage of pilot programme design is a major factor in Africa’s poor participation rate in index insurance. We conducted a survey with 392 farming households across eight of the state's 16 Local Government Areas (LGAs). This survey employed a contingent valuation method (CVM) to assess the value respondents are willing to pay for such insurance and add to the growing literature on IBLI uptake. We discuss important issues ranging from the cost of the IBLI premium to the impact of social capital and access to alternative sources of income on adoption. Our findings show that farmers are willing to pay a 1.3% premium for IBLI, which is lower than the current premium charged for traditional agricultural insurance in Nigeria which typically ranges from 2% to 5%. The results also highlight the need to consider insurance uptake and access to credit as complementary measures, rather than substitute strategies for managing the risks posed by severe climate shocks and extremes.

KEYWORDS:

• Index-based livestock insurance
• agricultural risk
• adaptation strategy
• and sub-Saharan Africa

1. Introduction

Livestock production is a risky enterprise in many underdeveloped countries and climate hazards pose serious dangers to the sustainability of agricultural development in these regions (Mills et al., 2015). In Nigeria, the livestock sector’s capacity to adapt to and recover from shocks in a timely and effective manner is still almost nonexistent. Nearly 40% of the country’s GDP comes from agriculture, with animal production playing a significant role (Fakoya, 2007; Hirfrfot et al., 2014). Losses and damage from climate change have the potential to impede development by exacerbating the frequency and severity of shocks that contribute to poverty. In this regard, climate risk insurance can play a critical role in breaking the cycle of vulnerability and poverty by compensating for the losses incurred as a result of catastrophic weather events. Enhancing people’s capacity to manage and mitigate climate risk can significantly reduce their vulnerability and contribute to their long-term social and economic well-being. The implementation of climate change adaptation strategies is a crucial component of this effort (Aina et al., 2023a, 2023b). However, livestock insurance is still not widely adopted in developing countries, primarily because formal insurance markets have remained underdeveloped and largely ineffective in many impoverished rural areas (Fonta et al., 2018). Several factors have discouraged farmers from using this insurance option, including high administrative costs and long delays in receiving compensation after livestock losses (Bulte & Haagsma, 2021).

In response to these challenges, there's growing support for alternative strategies that help low-income farming households build resilience against climate risks (Churchill, 2006; de Bock & Gelade, 2012). The use of insurance is expected to protect farmers against production shocks and motivate them to modernize their farming practices. Within the past decade, index-based livestock insurance (IBLI) has become a relatively novel product adopted in developing countries. IBLI schemes have been widely introduced in countries such as Mongolia, Kenya, and Ethiopia, with similar programmes being designed in other arid regions, including South Africa, Mali, Nigeria, Senegal, Somalia, and Zimbabwe (Castell, 2017). Unlike conventional insurance, which compensates for proven losses, IBLI is based on the value of an observable index that captures fluctuations in rainfall and/or temperature (Barnett & Mahul, 2007; Bulte & Haagsma, 2021). The relatively transparent nature of the scheme presents important benefits and allows insurance companies to shift their risk to the global reinsurance markets (Bulte & Haagsma, 2021). It also allows farmers to avoid challenges associated with the indemnification of losses unique to traditional insurance and eliminates the significant transaction costs involved in monitoring behaviour and verifying actual losses (Miranda & Farrin, 2012).

This study investigates the potential of adopting IBLI as a means of mitigating the adverse impacts of climate change. Given the scarcity of research on livestock insurance, this study diverges from the standard approach to index insurance demand by examining the determinants of IBLI adoption in Nigeria. Additionally, we estimate the value that potential IBLI adopters would be willing to pay for the insurance scheme. Although IBLI is not yet established in Nigeria, this paper provides an initial framework that policymakers can use to develop appropriate strategies to reduce the risks and uncertainties associated with livestock-rearing activities in the country. Therefore, understanding the factors that drive the adoption of the insurance scheme would help policymakers design sound public policy measures for the livestock sector in the country. This study makes valuable contributions to the existing literature by shedding light on the distinctive socio-economic attributes and institutional frameworks present in the West African region. Specifically, factors such as premium costs, income gaps, religious beliefs, education levels, and even gender play a significant role in the adoption of insurance innovation in the region (Auriol et al., 2020; Olarewaju & Msomi, 2021; Stoeffler et al., 2022). Our approach emphasizes the importance of considering region-specific insights rather than relying solely on evidence from other regions where the scheme is popular when formulating policy initiatives and engaging stakeholders in the context of index insurance.

The remainder of this paper is organized as follows. Section (2) delivers an overview of the existing literature. Section 3 describes the methodology used in this study. This section also describes the study area and data used. The results are presented in Section 4, and Section 5 discusses the main results and concludes with several policy implications.

2. Previous Literature

The literature widely acknowledges that livestock farmers in low-income countries are exposed to significant environmental risks and that there is increasing awareness of IBLI (Takahashi et al., 2016). However, despite the optimism about IBLI as an effective mitigant of weather shocks, various scholars have raised doubts regarding its potential (Binswanger-Mkhize, 2012; Morduch, 2006). Furthermore, numerous studies have shown that the adoption rate of IBLI among the target population is typically low, rarely exceeding 30%. For example, Chantarat et al. (2013) concluded that IBLI might eliminate 25–40% of the total livestock mortality risk in Africa. However, the design of IBLI contracts still faces demand-side challenges and many livestock farmers remain reluctant to adopt IBLI. Hence, it is necessary to reconsider earlier optimism regarding index-based insurance, and experts are beginning to consider numerous elements that could impact the insurance demand (Leblois et al., 2014). A growing number of projects offer payoffs based on aggregate performance indicators, rather than individual-specific results, to close the gap between insurance demand and supply. For example, Takahashi et al. (2016) examined the purchase direction of IBLI in southern Ethiopia and focused on the role of accurate product understanding and price on adoption. This study found that offering randomly distributed coupons led to an immediate increase in product adoption, without dampening future demand.

While most research on index-based insurance focuses on crops, very few studies have explored its use for livestock (IBLI). Additionally, existing IBLI research is concentrated in East Africa, leaving a significant knowledge gap regarding its potential in other African regions (Aina et al., 2024; Matul et al., 2013; Miranda & Farrin, 2012). Bulte and Haagsma (2021) show that the welfare effects of insurance are unclear even without transaction costs or basis risk. Their study identifies the conditions under which insurance adoption can reduce the expected income. In a separate study, McPeak et al. (2010) experimentally investigated the concept of index-based livestock insurance in northern Kenya and analysed gameplay patterns. The study aimed to identify factors that are important to households when considering IBLI adoption. Factors identified by households include household income from livestock products, susceptibility of animal production to shocks, and the long-term impacts of livestock loss on household well-being.

In light of the limited body of literature concerning livestock insurance, this study investigates the demand for index insurance by examining the factors influencing the adoption of IBLI, specifically within the context of Nigeria. We also estimated the value that potential IBLI adopters would be willing to pay for the insurance scheme. We build on McPeak et al. (2010) and Amare et al. (2019) and look at the responses associated with the adoption of IBLI in small-holder pastoral households. Unlike previous studies that have focused only on index insurance adoption constraints, we also determine the value IBLI adopters are willing to pay and highlight the welfare implications of the new insurance regime compared to traditional insurance in the study area. Overall, we provide insights that can inform the scaling-out of this insurance product to other areas within the country and West Africa.

3. Methodology

3.1. Data

This study was conducted in Kwara State, Nigeria (Figure 1). The state covers an area of approximately 32,500 square kilometres and experiences a distinct seasonal pattern with two six-month periods: wet season and dry season. The average annual maximum temperature ranges from 31°C to 35°C, whereas the minimum falls between 20°C and 24°C. Rainfall plays a crucial role in the region's livestock population dynamics as it directly affects agricultural production, which serves as a vital source of food for livestock. Climate shocks can have a negative impact on livestock productivity (Ayinde et al., 2024; Begzsuren et al., 2004). Figure 2 illustrates weather trends over the past six decades. The average annual rainfall across the state varied considerably. The driest areas received as little as 57 mm, whereas the wettest areas received up to 145 mm.

Figure 1. Study area in Nigeria.

Figure 2. Weather Trend in Kwara State, Nigeria (1946–2011).

Ethical clearance for this study was obtained prior to the commencement of fieldwork and All participants provided written informed consent before participating in the study. We employed a mixed-method approach to collect data, which involved focus group discussions, key informant interviews, and surveys of 392 sampled households. Our main survey targeted livestock farmers who had already employed traditional insurance based on information from the Nigerian Agricultural Insurance Corporation (NAIC). This approach provides a baseline to compare traditional insurance mechanisms with the proposed innovative product, given that the farmers understand the current satisfaction levels, coverage gaps, and limitations of traditional insurance, which can inform their choice of IBLI (Ankrah et al., 2021). Furthermore, these farmers have demonstrated a willingness to engage in risk management strategies; hence, their experiences and perceptions can offer valuable insights into what they value in an insurance product, which can be critical for tailoring IBLI to meet their needs (Nshakira-Rukundo et al., 2021).

Over 70% of the farmers in the study area rely on livestock for the majority of their income (Kwara State Diary, 2009). Eight Local Government Areas (LGAs) from the 16 LGAs in Kwara State were randomly chosen, and 49 livestock farmers were randomly selected from each of the eight LGAs. A focus group discussion was conducted with development agents, community leaders, and local-level coordinators of economic associations to gain a better understanding of respondents’socioeconomic features and their associated risk characteristics.

3.2. Empirical model

Two theoretical perspectives – production and consumption channels – can be used to analyse the effects of insurance on livestock farmers. The consumption channel examines these consequences after a shock, whereas the production channel investigates how insurance could reduce the price of risks before a shock occurs. To understand the demand for index-based insurance among livestock farmers in response to production shocks, we present a dynamic model of livestock farmers facing aggregate idiosyncratic production shocks, such as natural disasters (droughts, floods, and wildfires), pests and diseases, and trade disruptions. We adapted the model using techniques from Gollier (2003) and De Nicola (2015) to examine demand for index insurance. First, we develop a model without index insurance (baseline scenario) to analyse the optimal consumption, production, and investment decisions of livestock farmers without index-based livestock insurance. Second, we incorporate the option of IBLI to explore how the problem changes when an insurance mechanism is considered in the analysis.

3.2.1. Modelling farmers without IBLI

Let us consider a livestock farmer (an economic agent) who allocates her income w_t + 1 to consumption c_t and livestock-related investment. It is assumed that this related investment provides additional income in accordance with a production function with declining marginal returns and unpredictable productivity shocks ϵ _i,t, that capture weather variability, either in terms of rainfall or temperature change. Agents are assumed to be rational and to maximise the expected discounted utility. We assume a von Neumann – Morgenstern utility function that includes the risk attitudes and certainty equivalence of representative livestock farmers. Therefore, the representative farmer maximises the expected present discounted utility of consumption denoted by $E_t{\sum }_{j=0}^{\mathrm{\infty }}{\text{β}}^{j}u(c_t+j)$, where utility is a function of consumption c and is given as $u\left(c_t\right)={\displaystyle \frac{c^{1-\mathrm{\Phi }}}{1-\mathrm{\Phi }}}$. We additionally posit that agents have a constant relative risk aversion utility where $\mathrm{\Phi }$ denotes the coefficient of relative risk aversion. Consequently, the agent’s optimisation baseline problem can be expressed as V(w_t, ϵ _i,t) = max u(w_{t –} I_t) + βEV (w_t + 1), subject to c_t= w_{t –} I_t and $w_{t+1}=Q_i{ϵ}_{i,t}{I}_{\tau }^{\alpha }{a}_i^{1-\alpha }{\eta }_{t+1}$. The total weather shock that the livestock farmer cannot predict at the time investment decisions are made is represented by the weather variation, η_t + 1. Ԛ_i is the individual-specific time-variant productivity coefficient, $a_i$ represents the assets owned by the agent. However, it is important to note that the idiosyncratic terms Ԛ_i and ϵ_i,t are both log-normally distributed with mean and variance ${\sigma }_Q^2$ and ${\sigma }_{ϵ}^2$, respectively. It is assumed that livestock farmers cannot adjust the assets to which they are endowed. Assets can be the amount of land detained.

We calculate the first-order condition for the optimal capital level by matching the marginal utility of consumption today with the anticipated discounted marginal utility of consumption tomorrow. The expression is given as $u{}^{\mathrm{\prime }}\left(c_t\right)=\text{β}{E}_t\left[u{}^{\mathrm{\prime }}\right(c_{t+1}\left)\alpha Q_i{ϵ}_{i,t}{I}_{\tau }^{\alpha }{a}_i^{1-\alpha }{\eta }_{t+1}\right]$.

3.2.2. Modelling farmers with IBLI

To better understand the role of IBLI, the baseline framework is expanded, and it is assumed that livestock farmers can buy ι_t + 1 unit(s) of insurance to protect against weather variability, which affects livestock growth and security. Each unit of purchased insurance pays (1 – η_t + 1) to offset any bad weather shocks. The optimization problem thus becomes $V\left(w_t,{ϵ}_{i,t}\right)=\underset{I_t,i_t\ge 0}{max}\left[u\right(w_t-I_t)+\text{β}{E}_{t}V(w_{t+1})$, subject to c_t= w_{t –} I_t and $w_{t+1}=Q_i{ϵ}_{i,t}{I}_{\tau }^{\alpha }{a}_i^{1-\alpha }{\eta }_{t+1}+{\iota }_t\left(1-{\eta }_{t+1}\right)\text{–}{\iota }_{\iota +1}{P}_t{ϵ}_{i,t}$. Where the term Ƥ_t is the actuarially fair price for one unit of weather insurance and is defined as $P_t={\int }_0^1(1-\eta )f\left(\eta \right)\mathrm{d\eta }$. Ƥ_t appears in the transition equation rather than in the budget constraint since it is assumed that agents have credit to pay for the insurance premium and that they can observe their productivity level before insurance purchasing. Following the Bellman principle, the optimisation problem under full insurance can then be rewritten as follows: (1) $V\left(w_t,{ϵ}_{I,t}\right)=\underset{I_t,i_t\ge 0}{max}\left[u\right(w_t-I_t)+\text{β}{E}_{t}V(w_{t+1})$(1) (2) $c_t=w_t-I_t$(2) (3) $w_{t+1}=Q_i{ϵ}_{I,t}{I}_i^{\alpha }{a}_i^{1-\alpha }({\eta }_{t+1}+\underset{\boxed{\text{{{ Weather insurance component}}}}}{\underbrace{\left(1-{\eta }_{t+1}\right)\text{–}{P}_t{ϵ}_{I,t}}})$(3)

3.3. Estimation strategy

3.3.1. Decision to adopt specification

We used a logit model framework to examine the relationship between household factors and a binary response variable that separates adopters from non-adopters. The fact that the anticipated probabilities are limited to a range between zero and one distinguishes this approach from the linear probability model. Logit model estimations show a statistical link between numerous explanatory factors and the likelihood of a farmer choosing an IBLI. Following Gujirati (1995), a binary logit model can be expressed as. (4) $P_i=E\left(Y={\displaystyle \frac{1}{X_i}}\right)={\displaystyle \frac{1}{1++e^{-\left({\text{β}}_0+{\text{β}}_1{X}_i\right)}}}$(4) which can be further reformulated as (5) $P_i={\displaystyle \frac{1}{1+e^{-Z_i}}}$(5)

Where $Z_i=\left({\text{β}}_0+{\text{β}}_1{X}_i\right)$

If $P_i$ is the likelihood of a respondent being an adopter, then the probability of non-adopter can be written as $1-P_i$.

Hence (6) ${\displaystyle \frac{P_i}{1+P_i}={\displaystyle \frac{1+e^{Z_i}}{1+e^{-Z_i}}=e^{Z_i}}}$(6) By taking the natural logarithm of the odds ratio, a linear relationship between the explanatory variables and probability of adoption can be established. This linear relationship is not only applicable to the explanatory variables, but also extends to the model parameters. Specifically, the binary logit model can be expressed as the log of the odds ratio in favour of adoption, where ${\displaystyle \frac{P_i}{1+P_i}}$ represents the ratio of the probability that the farmer will adopt IBLI to the likelihood that they will not adopt it. (7) $L_i=L_n\left({\displaystyle \frac{P_i}{1+P_i}}\right)=Z_i={\text{β}}_0+{\text{β}}_1{X}_i$(7) Where

$P_i$ is the probability of being an adopter ranging from 0 to 1

$Z_i$ is a function of $n$-explanatory variables ($X_i$) and is expressed as: (8) $Z_i={\text{β}}_0+{\text{β}}_1{X}_i+{\text{β}}_2{X}_2+{\text{β}}_3{X}_3+\cdots +{\text{β}}_n{X}_n+U_i$(8) ${\text{β}}_0$ represents the constant term or intercept while ${\text{β}}_1,{\text{β}}_2,{\text{β}}_3,\dots {\text{β}}_n$ represents the respective slopes or parameters to be estimated. $L_i$ is the logarithm of the odds ratio, $X_i$ is a vector that contains the relevant characteristics of the respondent, and $U_i$ is the error term or disturbance term of the logit model. The Maximum Likelihood (ML) method was employed to estimate the model parameters following the approach outlined by Gujirati (1995) and Maddala (1992). The dependent variable used in the binary logit model was the adoption of index-based livestock insurance, with respondents classified as either adopters or non-adopters based on their willingness to purchase insurance. Respondents who purchased the indexed insurance were assigned a value of “1”, while those who did not were assigned a value of “0”.

3.3.2. Willingness to pay specification

The contingent valuation method (CVM) is a widely used stated-preference technique for eliciting willingness to pay (WTP) for environmental goods and services that are not easily traded in markets or if prices do not exist (Steigenberger et al., 2022). CVM is appealing because its results are relatively easy to interpret. For instance, WTP estimates can be represented in terms of mean or median values per household, or aggregate values for the relevant population of interest (Fonta et al., 2010). Since we have already determined farmers’ decisions to adopt IBLI (i.e. the binary logit model above), we run an OLS estimation technique for the subsample of adopters with positive WTP values. This is based on the assumption that, when a farmer decides to adopt, he or she will provide a positive WTP value.

Let $Y_1$ denote the farmers’ maximum WTP amount for the scheme and given by (9) $Y_{1i}=u_1^{\mathrm{\prime }}\alpha +\alpha {\mu }_1$(9) for the (log) of the WTP equation, where $\alpha$ is a scale factor, $Y_{1i}$ is a binary variable observed only when the farmer adopts IBLI and ${\mu }_1$ is the error term that is assumed to be bivariate normal with zero means, variances equal to 1 and correlation coefficient $\rho$.

The conditional expected value of $Y_{1i}$is then given as: (10) $E[\ln Y_{1i}|Y_{2i}=1]=u_1^{\mathrm{\prime }}\alpha +\mathrm{\rho \sigma \lambda }\left(v_1^{\mathrm{\prime }}\text{β}\right)$(10) where $\lambda \left(u_1^{\mathrm{\prime }}\alpha \right)=\varphi \left(v_1^{\mathrm{\prime }}\text{β}\right)\left|\mathrm{\Phi }\right(v_1^{\mathrm{\prime }}\text{β})$is the inverse of the Mills ratio and $\varphi$ and $\mathrm{\Phi }$ are the standard normal density and standard normal functions, respectively. The$Y_{1i}$ (i.e. valuation equation) is estimated within an ordinary least squares (OLS) regression framework similar to the second step of Heckman’s estimation procedure (Fonta et al., 2010).

4. Results and Discussion

4.1. Socioeconomic summary

This study examined the socioeconomic characteristics of the participating farmers (n = 392), and the results are presented in Table 1. More than half (54.1%) had educational attainment beyond secondary school, which aligns with Trang (2013), who suggests that higher education can be an advantage for adopting insurance. A significant proportion (64.5%) belonged to at least one economic association, potentially providing access to valuable resources and networks for livestock management. Interestingly, while 212 respondents reported a secondary source of income, approximately 53% were classified as low-income earners.

Table 1. Socioeconomic characteristics of respondents.

Variables	Category	Observation	Percentage
Gender	Male	311	79.3
	Female	81	20.7
Marital Status	Single	47	12.0
	Married	345	88.0
Educational Level	Primary school	88	22.5
	Secondary school	92	23.4
	Vocational school	172	43.9
	University education	40	10.2
Economic Association membership	No	139	35.5
	Yes	253	64.5
Association Type	None	139	35.5
	Cooperative	135	34.4
	All Farmers Association	118	30.1
Access to Extension Service	Yes	261	66.6
	No	131	33.4
Age	≤25	12	3.1
	26–35	88	22.5
	36–45	101	25.8
	46–55	103	26.3
	≥56	88	22.5
Household Size	≤5	213	54.3
	6–10	128	32.7
	≥11	51	13.0
Monthly Off-farm Income in Naira^a (n = 212)	Low Income (≤50,000)	113	53.3
	Middle Income (50,001–10,000)	78	36.8
	High Income (≥100,001)	21	9.9
Farming Experience	≤5	99	25.3
	6–10	180	45.9
	11–15	68	17.4
	≤16	45	11.5

Source: Authors, data collected from the field.

^a USD 1 = ₦520 (Nigerian Naira).

4.2. Differences between IBLI Adopters and Non-Adopters

We further examined the differences between IBLI adopters and non-adopters, particularly in terms of access to extension services, membership in economic associations, farming experience, educational background, and involvement in off-farm income activities. The results in Table 2 indicate that a significant proportion of IBLI adopters have access to extension services and are members of economic associations, unlike non-adopters who lack access to extension services and are not members of economic associations. Additionally, the results suggest that many farmers who have adopted IBLI have less than ten years of farming experience, have secondary school education, and engage in off-farm income activities. The chi-square test confirmed that the differences in access to extension services, participation in off-farm economic activities, and membership in an economic association were statistically significant (p < 0.001) between adopters and non-adopters of IBLI. This indicates that respondents who lack access to extension services and those without other income sources are less likely to adopt the IBLI.

Table 2. Relationship between discrete variables and the adoption of index-based livestock insurance.

Variables	Category	Adopters of IBLI	Non-adopter of IBLI	Chi-Square value	Sig.
Access to Extension Service	Yes	213	48	83.311***	0.000
	No	90	41
Household Size (≥6)	Yes	145	34	49.837***	0.000
	No	158	55
Off-farm income	Yes	196	16	101.223***	0.000
	No	107	73
Farming Experience (≥10 years)	Yes	138	25	44.892***	0.000
	No	165	64
Economic Association membership	Yes	164	64	68.037***	0.000
	No	139	25
Secondary School Education	Yes	205	24	26.331***	0.000
	No	95	85
Sex HH	Male	230	66	0.628	0.327
	Female	73	23
Marital Status	Married	245	75	0.025	0.799
	Single	58	14

Source: Authors, data collected from the field.

Our findings suggest that farmers with extra income from outside the farm (off-farm income) are generally better able to pay back loans (higher payment capacity). We also examined whether farmers were involved in economic associations, which are considered a sign of strong social connections (social capital). We found that strong social connections might help farmers access resources to adapt to climate change. The chi-square test also revealed that social capital was statistically significant between those who adopted the insurance programme and those who did not, with a p value of 0.000. This finding suggests that the likelihood of IBLI adoption is positively and significantly related to membership in local social groups, whereas gender and marital status are not significant. The remaining factors listed in Table 3 also exhibited comparable patterns.

Table 3. Overall ranking of weather-related risks to livestock production.

Type of Weather Risk	Yes (%)	No (%)	Mean	Rank
Flood	80 (20.4)	312 (79.6)	0.23	3
Drought	292 (74.5)	100 (25.5)	0.75	1
Excessive Rainfall	0 (0)	392 (100)	0.00	5
Fire Outbreak	20 (5.1)	372 (94.9)	0.08	4
Too High Temperature	259 (66.0)	133 (34.0)	0.68	2

Source: Authors, data collected from the field.

We ranked farmers’ perceptions of the impact of weather risks on their production. The magnitude and frequency of exposure to stress in the form of weather-related events have a significant impact on the production process and overall growth of livestock (Sivakumar & Motha, 2007). Agricultural risk events are often difficult to control because of the subsistence-based nature of agricultural production. Livestock farmers are exposed to a range of risks such as extreme weather changes, pests and diseases, random physical hazards, and technological failures. The effects of weather-related risks on livestock production were rated as extremely severe by 55% of respondents, making it the highest-ranked risk event (Figure 3).

Figure 3. Impact of weather risk on respondents. Source: Authors, data collected from the field.

To identify the specific weather risks faced by livestock farmers, we conducted a ranking exercise based on the responses of survey participants. This approach enabled us to categorise different types of weather-related risks and to design more targeted policy interventions to address the challenges encountered by farmers. The findings in Table 3 indicate that drought is the most significant weather risk affecting animal production, as reported by 74% of respondents. High temperatures were also identified as a challenge to the output by 66% of the participants. None of the respondents mentioned excessive rainfall as a potential production risk. These results inform the scope of our study, which focuses on the factors driving the adoption of IBLI as a risk management strategy in the presence of drought.

Our survey also collected supplementary information on the effectiveness of farmers’ strategies for managing climate risk, and the results were ranked around the mean. Only 80 respondents confirmed the effectiveness of the traditional insurance provided by NAIC as a risk management strategy, indicating the need for an alternative and more effective risk mitigation technology, such as IBLI. Among the various risk management strategies employed, income from other business sources was ranked first, with 42% of the respondents confirming its effectiveness. Other strategies include the use of vaccines to improve livestock health, animal breeds, and water-harvesting technologies. The challenges associated with traditional insurance were also discussed, such as most claims not being paid, leaving farmers uncompensated for their losses despite having purchased insurance policies, which is consistent with previous studies (Barnett & Mahul, 2007; NARF, 2014).

4.3. Determinants of the adoption of IBLI

The results in Table 4 reveal that, at a significance level of less than 1%, access to loans positively and significantly influences the likelihood of adopting IBLI. For those with access to loans, the likelihood of IBLI implementation increased by a factor of 6.267. This implies that access to loans improves the capacity to pay the IBLI premium. This outcome is brought about by the fact that those in poverty are less able to accumulate resources for risk transfer and risk management related to climate change (Tadesse et al., 2015). Therefore, farmers’ inadequate financial resources are a major barrier to improving the index insurance demand. Similar to the findings of Nshakira-Rukundo et al. (2021), the potential insurance market is divided into farmers with higher incomes, and those with lower incomes.

Table 4. Estimates of determinants of IBLI adoption (Participation model).

Variables	Odds ratio	Std. err	z
Access to Loans	6.267	3.486	3.29***
Distance to a weather station	3.017	0.758	−4.36***
Access to Extension Service	0.155	0.172	4.11***
Household Size	4.217	2.312	−2.43**
Off-farm income	1.046	0.085	0.29***
Years of farming experience	4.198	1.568	4.23***
Economic Association membership	10.515	6.195	4.06***
Secondary School Education	0.756	0.660	−0.32***
Sex HH	0.213	0.137	−2.45
Marital Status	1.169	0.165	2.77***
Age of HH	0.1204	0.069	−4.27***
Frequency of Premium payment	1.168	0.183	1.00
Number of observations:	392
LR chi^2 (15):	418.21
Prob > chi^2:	0.000
Pseudo R^2:	0.7634
Log-likelihood:	−64.326

Source: Authors, data collected from the field

As expected, the likelihood of adopting IBLI was significantly and negatively affected by the distance from a respondent’s residence to the weather station. This could be because farmers located far from a weather station may not have easy access to climate information, potentially hindering access to the IBLI product that may rely on ground-based verification methods alongside satellite data. Table 4 also shows that, at a significance level of less than 1%, those who belong to local economic associations are more likely to adopt insurance products. Specifically, the odds ratio in favour of IBLI adoption increased by a factor of 10.515 for such farmers. This result is similar to the findings of previous studies of technology adoption. For instance, Abugri et al. (2015) and Bogale (2014) found that social capital in the form of association membership positively and significantly influences farmers’ decisions to purchase insurance.

In our study, we found that neither the gender of the respondents nor the frequency of premium payments significantly influenced the likelihood of adopting IBLI. However, our results show a negative correlation between educational level and IBLI adoption, which aligns with the findings of Takahashi et al. (2016). This unexpected result suggests that years of formal education may not always be an effective predictor of IBLI adoption because livestock farmers may not rely solely on formal education when making practical farm-level decisions. Additionally, our analysis indicates that age negatively correlates with the likelihood of adopting IBLI, suggesting that younger respondents are more likely to rely on IBLI, whereas older respondents may be more risk-averse because of higher resource constraints from family demands. This finding is partly supported by the coefficient estimates of household size and marital status, which suggest that respondents with dependents in the household are less likely to adopt IBLI. This may be because most disposable income is shared with dependents, leaving no extra income for purchasing IBLI. This highlights the importance of considering the prevalent family structure and resource constraints in a region when designing IBLI programmes, as it is common for families in sub-Saharan Africa to have many dependents.

4.4. Willingness to pay valuations

In the context of livestock insurance, willingness to pay (WTP) refers to the amount a livestock farmer is prepared to spend on an IBLI product, given their income, risk tolerance, and other personal circumstances. This value can be determined by using the contingent valuation method (CVM). Although CVM is commonly used to evaluate non-market environmental initiatives, it is increasingly being used to evaluate private market goods and services (Steigenberger et al., 2022). In general, two methods are used to calculate WTP under the CVM. The initial format is closed-ended, often known as a referendum or the “take it or leave it.” Each question type offers only a list of pre-selected possibilities, preventing respondents from offering original or unexpected answers. The second most commonly used strategy was the open-ended technique.

To account for the actual sum, farmers would be ready to pay for index-based insurance, providing coverage for animals valued at ₦1,000,000 (US$1923.1). This study used an open-ended contingent valuation approach. This value was determined to be the average livestock valuation of the respondents through focus group discussion. Additionally, most respondents were small-scale farmers, with restrictions that prevented them from exceeding the stated value. The open-ended contingent valuation approach was used because it is free from anchoring bias and does not provide respondents with information about the value of change. Additionally, the approach is appropriate because index insurance is currently not established in Nigeria, and all respondents employed in this study currently utilise the traditional insurance programme to mitigate weather-related risks. Similar to Fonta et al. (2018), the open-ended contingent valuation technique is also important because it allows an easy evaluation of the mean amount adopters are willing to pay for an IBLI product.

First, we focus on the responses of respondents and their implications for the value they are willing to pay for the scheme. As emphasised by Fonta et al. (2018), it is important to differentiate between valid and invalid responses for analysis purposes. Table 5 shows that of the 392 completed interviews, 89 (23%) provided invalid responses to the valuation question. The reasons for these responses included protest bidders (71) and outliers (18). Protest bidders included respondents who claimed they could not afford to pay for the scheme owing to budgetary constraints (29), those who believed that traditional insurance was sufficient (16), and those who believed that it was the responsibility of the government (26). Outliers were those respondents whose maximum WTP amount exceeded their income (18).

Table 5. Reasons for not willing to pay for IBLI.

Reason for not willing to pay	Observations	%
Protest Zero Responses
Cannot afford	29	32.6
There is no problem with the traditional Insurance	16	18.0
It is the responsibility of the government	26	29.2
Outliers
WTP amounts above 5% of income	18	20.2
Total	89	100

Source: Authors, data collected from the field.

Based on the responses in Table 5, it was necessary to determine whether eliminating invalid responses from the estimation would result in a sample selection bias, which may have serious effects. For instance, if the valuation function is used to test theoretical validity, the analysis may yield inconsistent estimates similar to those reported by Maddala (1983). Additionally, the aggregated WTP values and estimated benefit measures may not be accurate. We tested for variations in the means of household factors between the two groups to determine whether there was a sample selection bias. As outlined by Fonta et al. (2018), any substantial discrepancy could be an indicator of sample selection bias. As shown in Table 6, there were no significant differences between the two groups of respondents for any factor. These results show no sample selection bias when invalid responses were not included in the econometric analysis.

Table 6. Difference in means between WTP response categories.

Dependent variable: willingness to pay	Positive WTP responses		Invalid WTP responses		t-stat
	Coef.	Std. Err.	Coef.	Std. Err.
Household Size	4.8	0.801	5.0	1.010	0.36
Years of Education	12.9	0.645	11.6	0.748	0.24
Farming Experience (Yrs)	10.6	0.988	9.7	2.692	−0.45
Farm Income (₦)	78,045.6	6,785.1	65,962.4	8,878.3	0.50
Non-farm Income (₦)	45,245.2	5,257.6	51,475.3	7,545.8	0.66
Age	46.1	0.831	50.2	2.022	−1.55
Gender (Male = 1)	0.92	0.037	0.90	0.045	0.39
Extension Visit	0.88	0.021	0.81	0.066	0.41
Access to Loan	0.45	0.087	0.66	0.071	0.61
Marital Status	0.91	0.028	0.89	0.045	0.77

Source: Authors, data collected from the field.

We then focus on the econometric analysis of the determinants of WTP for IBLI using a subsample of valid replies (where WTP > 0), as shown in Table 7. The results show that the coefficients of access to loans and off-farm income are positive and significant. This suggests that farmers willing to invest in IBLI tend to be more financially stable than their peers are. They often have access to additional financial resources, enabling them to manage the risks associated with severe climate shocks and extremes better. The results further revealed that farmers with farms situated at greater distances from weather stations were willing to pay less for IBLI than those with weather stations in close proximity to their farms. For instance, farmers whose farms are located farther from weather stations are, on average, 3.8 times less willing to pay for the scheme than those with closer farms. One possible reason for this observation could be that farmers with farms located closer to weather stations have a better understanding of local climate conditions and are more aware of the risks associated with climate-related shocks. As a result, they may place a higher value on IBLI coverage, leading to a higher willingness to pay. On the other hand, farmers with farms farther from weather stations might perceive lower climate-related risks or have limited access to weather information, potentially influencing their lower willingness to pay for IBLI.

Table 7. Determinants of willingness to pay (valuation model).

Variables	Coeff.	Std. err	t-value
Constant	4.567	4.785	0.95
Access to Loans	0.964	0.435	2.22**
Distance to a weather station	−0.378	0.141	−2.68**
Access to Extension Service	0.074	0.035	2.11***
Household Size	5.311	11.874	0.45
Off-farm income	0.889	0.296	3.00**
Years of farming experience	2.681	8.974	0.30
Economic Association membership	0.984	0.439	2.24**
Secondary School Education	0.614	0.233	2.64**
Sex HH	0.665	0.278	2.39***
Marital Status	0.322	0.261	1.23
Age of HH	3.341	6.587	0.507
Frequency of Premium payment	3.824	5.214	0.73
Number of observations:	303

Source: Authors, data collected from the field.

The results also show that higher WTP amounts are associated with farmers who possess at least a secondary school education. One potential reason for this could be that higher education levels may provide a better understanding of the benefits and workings of IBLI. They may also have more financial resources or be in a more secure financial position, allowing them to afford higher insurance coverage premiums. Moreover, farmers who belong to economic associations and social groups show greater demand for insurance products than those who are not members of such organisations. This could be because membership in economic associations provides farmers with access to information, resources, and support networks, making them more aware of the benefits of insurance and more likely to seek it. Additionally, group membership may enable collective decision making and resource pooling, which can facilitate the purchase of insurance as a group, potentially reducing costs and increasing affordability for individual farmers.

The mean WTP value for the IBLI pilot initiative was predicted using a subsample of valid replies, because there was no significant evidence of sample selection bias. The results are presented in Table 8. Mean WTP for the IBLI offering coverage for livestock valued at ₦1,000,000 (US$1,923.1) is estimated as ₦12,563 or US$24.2 (1.3%) monthly with a 95% confidence interval of ₦10,863 (US$20.9) and 14,917 (US$28.7), respectively. While premiums for conventional agricultural insurance in Nigeria can range from 2 to 5 per cent of the livestock’s insurance value (Babalola, 2014), our estimate is slightly less expensive. This may be attributed to the different product coverages usually offered by traditional insurance contracts. To represent consistency and fairness in pricing choice, it is crucial to note that the ongoing IBLI pilot in Nigeria lacks a standardised process for developing and charging insurance rates. The estimated WTP value from our analysis provides an important reference for the IBLI premium rate.

Table 8. Monthly mean WTP bid^a.

Model	Observation	Mean	CI-mean (95%)^B
OLS	303	12563.28 (20.9)	10863.22 (9.4) – 14917.62 (28.7)

^a Nigerian Naira ($US).

^b Normal-based 95% confidence interval generated with Bootstrap replications.

5. Conclusion and policy implications

Index-based insurance is a promising strategy for managing agricultural risks and reducing the impacts of natural disasters and climate hazards on farmers. This insurance helps limit losses from droughts, floods, rainfall fluctuations, temperature fluctuations, dry spells, and heat waves (Amare et al., 2019). It offers several benefits including minimising moral hazard and adverse selection, fast payouts during emergencies, increased investment in production inputs, and improved resilience to food insecurity (Aina et al., 2024). By providing a safety net, agricultural insurance helps farmers and households escape poverty in many regions directly affected by weather shocks and natural catastrophes (Aina et al., 2024; Fonta et al., 2018). Unfortunately, the uptake of agricultural insurance in sub-Saharan Africa remains the lowest worldwide (Nshakira-Rukundo et al., 2021). Instead, smallholder farmers continue to rely on less effective mechanisms such as asset depletion (Yilma et al., 2017) and savings, even when insurance options are available (Delavallade et al., 2015). Africa makes up only 0.5% of the global insurance market, while Europe, North America, and Asia account for 20%, 55%, and 20% of the total agricultural insurance premiums, respectively. The coverage and uptake of the few programmes being established, particularly those in West Africa, are still severely low (Fonta et al., 2018).

This study investigates the critical factors influencing the uptake of index-based livestock insurance (IBLI) in Kwara State, Nigeria. We also estimate the value IBLI adopters are willing to pay for the insurance scheme. IBLI is a financial product that requires careful implementation, and its widespread adoption requires a clear and well-articulated policy that considers the target farmers’ characteristics to encourage acceptance. The findings from this study contribute to filling the gaps related to promoting the uptake of IBLI and provide guidance to policymakers in their approach to mitigate the effects of climate change on livestock production in Nigeria. This study is one of the first to estimate the factors that drive the adoption of an IBLI scheme in Nigeria. In this respect, our study highlights the need to combine access to loans with the use of IBLI, so that households with limited cash can afford insurance. Considering insurance and credit as complementary rather than substitutes is crucial for managing the risks posed by severe climate shocks and extremes. Therefore, it is essential to rethink the relationship between these two financial tools to effectively manage risks.

Furthermore, access to extension services and economic association membership offers farmers social capital and encourages their decision to purchase IBLI. Since most farmers have no prior experience with IBLI, there is a need to explain the basic concept of insurance payouts, risk mitigation, and the implications of adoption on overall household welfare. Training and financial literacy in insurance should be provided through organised social groups to farmers to enhance their knowledge of climate risk perception and minimise any potential ambiguity that may arise from misperceptions (Getahun & Chotai, 2022). This is expected to reduce the stress on farming communities. When farming households feel secure about adopting IBLI, they can speculate on a favourable market, sell at the best price, and presumably increase their income and livestock accumulation. Ultimately, this could promote economic development and reduce poverty.

This study revealed that the mean WTP for the IBLI offering coverage for livestock valued at ₦1,000,000 (US$1923.1) is estimated as ₦12563 or US$24.2 in the study area. However, our estimate suggests a 1.3% premium for the IBLI, which is slightly lower when compared to what is currently being charged for traditional agricultural insurance in Nigeria (2–5%). This result implies that premium costs can be a strong barrier to insurance uptakes. This also indicates that the long-run welfare impact of the IBLI policy on vulnerable households could be significant as an alternative to traditional insurance. Therefore, policy decisions should focus on initiatives such as cash transfers and income support programmes to help poor farmers protect themselves against shocks and take advantage of index insurance programmes. This would include IBLI potentially acting as a safety net to prevent further impoverishment for vulnerable farmers after livestock losses due to drought, thus helping them mitigate the impact of shocks. The first step may require initial targeting of farmer clusters with similar risk profiles. This is crucial, especially in regions where expensive alternative risk-management measures and severe agricultural losses may trap households in poverty.

Despite the interesting findings described above, one limitation of this study is related to our exclusive focus on IBLI despite the existence of a variety of insurance programmes accessible to farmers to alleviate climate-related risks. In addition to IBLI, farmers in developing countries have access to other forms of insurance such as revenue insurance and commodity price insurance. Family and peer support also plays an important role in reducing vulnerability and increasing resilience in African communities. Thus, in Nigeria and many African countries, low participation in formal insurance programmes could also be due to the presence of alternative risk-management strategies and the influence of familial support networks (Abdul Mumin et al., 2022; Ramsawak, 2022). Finally, similar to several studies that employed similar methodologies, a limitation of the study is the potential overlap between the adoption decision of BLI and the willingness to pay for IBLI, which could lead to potential bias, as the price and conditions of the insurance may influence the pool of people willing to sign up (Amare et al., 2019; Fonta et al., 2018).

Acknowledgments

An earlier version of this paper benefitted from comments during sessions at the 30th International Conference of Agricultural Economists (ICAE) and the Annual World Bank Conference on Development Economics. The study was partially supported under the U.S. Agency for International Development (USAID)/Feed the Future Advancing Local Leadership, Innovation and Networks (ALL-IN) project grant titled “Building Resilience among West African Women Smallholders by Promoting Greater Access to Insurance, Financing and Advanced Agricultural Technologies”(2021–2024) https://basis.ucdavis.edu/project/linking-financial-and-agricultural-innovations-women-farmers-resilience-nigeria

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by Uma lele Mentorship Grant.

Notes on contributors

Ifedotun V. Aina

Dr Ifedotun V. Aina is a postdoctoral research fellow at the School of Economics, University of Cape Town, South Africa, where he conducts research on food price analysis and resource management by leveraging insights from economic and environmental modelling methods. He has an active research portfolio at the intersection of environmental economics, agricultural, economics, and behavioral economics. He is passionate about employing state-of-the-art econometric techniques and applied economic theory to address urgent questions in development policy. His doctoral dissertation focused on understanding specific themes in the water sector, namely, households' preferences for water conservation technologies, the impact of tariffs on the choice of water supply source and policy measures to optimize water allocation decisions in large water systems. He is a Lindau Nobel Laureate Meeting on Economic Sciences Fellow and a recipient of the Africa Regional International Student Exchange Intra-ACP Mobility Scholarship. He was a visiting scholar at the prestigious International Institute for Applied Systems Analysis (IIASA) in Austria, and a recipient of several academic awards.

Opeyemi E. Ayinde

Opeyemi E. Ayinde is a professor of agricultural economics and farm management at the University of Ilorin, Nigeria. She was a Fulbright Senior Research Scholar from her tenure at the Ohio State University in 2018, and previously served as a Senior Research Fellow at the South African Research Chair Initiative (SARCHI) at the Institute for Economics Research on Innovation, Tshwane University of Technology, South Africa, from 2015 to 2018. Prof. Ayinde's research spans agricultural economics, innovation, and gender issues, with a focus on climate risk management strategies. She has presented her work extensively at international conferences and workshops, contributing significantly to the International Association of Agricultural Economists and the Applied and Agricultural Economists Association, where she received the Uma Lele Award in 2016. Notably, she has also played pivotal roles in various executive capacities, including Vice President for the Africa Agricultural Economists Association and as West Africa representative on the board of African Network for Economics of Learning, Innovation, and Competence Building Systems (AFRICALICS).

Djiby R. Thiam

Djiby R. Thiam is an Associate Professor in Economics at the University of Cape Town, South Africa. His research focuses on resources and environmental economics, water resources economics, agricultural economics and development economics. His work looks at (i) the efficiency of water use in agriculture, (ii) the factors that drive the adoption of index-based drought insurance in the basin, (iii) the evolution of the hydrological cycle of the basin under various climate and socio-ecological changes, and (iv) the questions of land subsidence that affects groundwater management and water supply infrastructure planning. He has published articles in Water Economics and Policy, Agricultural Economics, Environmental and Resources Economics, Water Resources and Economics, Journal of Agriculture and Applied Economics and Comparative Economic Studies, among several other journals. He has consulted on production and environmental economics for the World Bank, the African Development Bank and several governments across the African continent. He has reviewed project proposals for funding organisations domestically and internationally. Djiby is a member of the editorial board of the Water Economics and Policy journal and 10 New Insights in Climate Science 2024, a joint initiative between Future Earth, the Earth League and the World Climate Research Programme. He has co-guest edited papers in Applied Economic Perspectives and Policy and Environmental Research Letters.

Mario J. Miranda

Mario J. Miranda is Professor of Agricultural, Environmental and Development Economics and Fellow of the Agricultural and Applied Economics Association. His research has produced one book, which has been adopted for courses offered by seven of the top ten ranked doctoral programs in Economics in the world, and over 70 peer-reviewed articles, two of which are among the three most frequently cited articles on “agricultural insurance”. He has advised twenty-eight doctoral students to completion, including four winners and honorable mention recipients of the Applied and Agricultural Economics Association Outstanding Doctoral Dissertation Award. Professor Miranda served seven years as Director of Graduate Programs in the Department of Agricultural, Environmental, and Development Economics at The Ohio State University, ranked by the National Research Council as the leading program in its field in the USA. Miranda has served as an Associate Editor of the Journal of Economic Dynamics and Control, the American Journal of Agricultural Economics and Computation.