Empirical modeling of agricultural climate risk

Effective policies for adaptation to climate change require understanding how impactsare related to exposures and vulnerability, the dimensions of the climate systemthat will change most and where human impacts will be most draconian, and theinstitutions best suited to respond. Here, we propose a simple method for more crediblypairing empirical statistical damage estimates derived from recent weather and outcomeobservations with projected future climate changes and proposed responses. We firstanalyze agricultural production and loan repayment data from Brazil to understandvulnerability to historical variation in the more predictable components of temperatureand rainfall (trend and seasonality) as well as to shocks (both local and over larger spatialscales). This decomposed weather variation over the past two decades explains over 50%of the yield variation in major Brazilian crops and, critically, can be constructed in thesame way for future climate projections. Combining our estimates with bias-correcteddownscaled climate simulations for Brazil, we find increased variation in yields andrevenues (including more bad years and worse outcomes) and higher agricultural loandefault at midcentury. Results in this context point to two particularly acute dimensionsof vulnerability: Intensified seasonality and local idiosyncratic shocks both contribute toworsening outcomes, along with a reduced capacity for spatially correlated ("covariate")shocks to ameliorate these effects through prices. These findings suggest that resiliencestrategies should focus on institutions such as water storage, financial services, andreinsurance.