Abstract

As a consequence of climate change, extreme weather events—both in terms of precipitation and temperature—are becoming more frequent, and more intense. A higher incidence of these events, along with increased climate uncertainty, may lead agricultural producers to take measures to minimize the impact of these shocks on their production and, consequently, on their well-being. One such mechanism is the acquisition of insurance against climate risks. It has been shown that, in some contexts, greater insurance coverage is associated with increased investment in agriculture, greater adoption of risky but high-yield productive projects, and lower default rates on agricultural loans (Lee et al., 2023; Karlan et al., 2014).
However, the literature is still inconclusive about whether climate shocks drive farmers to demand insurance. It has been found that farmers instead consider the difference between their income and the cost of acquiring the policy (Sun et al., 2024). Additionally, farmers may prefer other strategies to mitigate the impacts of climate change, such as adopting technological innovations (e.g., using genetically modified crops or irrigation districts) (Liv et al., 2020).
In countries in Asia and Africa, the use of insurance largely depends on whether it is subsidized by the government (J-PAL, CEGA, and ATAI, 2016). Similarly, in Colombia, more than 80% of the policies acquired in the agricultural sector are subsidized. Although insurance coverage is low—only around 5% of cultivated hectares are insured—it has shown an upward trend (Fasecolda, 2025). The percentage of subsidized policies acquired by small producers increased between 2013 and 2023, reaching nearly 100% of the policies in the latter year, when the government eliminated the subsidy for large producers.
There are significant research gaps regarding agricultural insurance. In particular, the literature has disproportionately focused on develping countries (e.g., China and the United States) (Ayodele et al., 2022), which has led to knowledge gaps in countries with high climate risk, such as Colombia, where this topic has been scarcely addressed. While there is evidence in Colombia that farmers use credit as an income-smoothing mechanism in response to climate shocks (Gáfaro et al., 2025), it has not yet been analyzed whether farmers use insurance as an adaptation strategy to climate change. This document would fill that gap in the literature and also present a detailed analysis of the use of subsidized insurance in Colombia.
This paper analyzes the effects of extreme weather events on the uptake of subsidized agricultural insurance during the period 2013–2023. We focus on floods and droughts, since freezing temperatures are not common in Colombia. Various sources of data will be used: (i) Administrative records from Finagro , corresponding to subsidized insurance policies acquired by farmers between 2013 and 2023. (ii) Precipitation data from the Climate Hazards Group InfraRed Precipitation with Station Data (CHIRPS), developed by the University of California, Santa Barbara. CHIRPS provides daily precipitation data since 1981, with coverage across the entire Colombian territory at a spatial resolution of 0.05° × 0.05° (5 km²). For each municipality, daily rainfall will be calculated as the average of all grid cells within municipal boundaries. From these data, extreme events such as rainfall deficits or excesses will be identified. (iii) Municipal-level information on agricultural production (planted and harvested areas, production, and yields), sourced from the Municipal Agricultural Evaluations (EVA). (iv) Official records of emergency declarations, managed by the National Unit for Disaster Risk Management. We use this data to construct alternatives measurements of floods and droughts, and check the robusstness of our results.
We will estimate the following baseline equation:
y_it= 〖shock〗_it β^'+X_it+ μ_t+δ_i+ε_it
Where y_it is an insurance uptake indicator: the number of policies purchased, the number of insured hectares, the insured value, and the percentage of insured hectares relative to the total planted hectares in municipality i and period t. The vector 〖shock〗_it includes temperature and rainfall shocks in municipality i during period t; X_it represents a set of municipal characteristics potentially associated with policy acquisition, such as the share of the primary sector in the municipal GDP and agricultural productivity. ; μ_tand δ_i identify time and municipality fixed effects, respectively. The parameter β measures the effect of climate shocks on access to insurance.
We expect to find that farmers will acquire a greater number of subsidized insurance policies or insure a larger portion of their cultivated hectares in municipalities with higher occurrence or severity of extreme weather events. These findings would suggest that subsidized agricultural insurance is being used by farmers as a mechanism to mitigate the risk associated with climate variability.