In Part 1 of this series, we discussed the basics of parametric insurance. As a summary, parametric insurance relies on the use of data to pay damages rather than subjective assessment of the damage by an adjuster. A natural area of application of parametric insurance is weather insurance given the availability of vast troves of weather data over long periods of time. The use of data to pay damages instead of the painstaking and subjective process of using loss adjusters on the ground holds the promise of greatly increasing affordability and adoption of insurance for adverse weather events which affects a wide set of industries from farming to energy. Here, we specifically discuss why now the time is right for parametric insurance to take off especially for agriculture.
The concept of parametric insurance has been around since the late 1980s with numerous pilots conducted across farmers in various countries. For decades, the concept remained in these pilot stages with difficulties in deploying parametric insurance widely. There were some major obstacles in the past that prevented parametric insurance from widespread usage:
1. Basis Risk
2. Computing resources
3. Farmer adoption
Basis risk remains a major issue to address for parametric insurance products. Basis risk refers to the difference in the on-the-ground experience for the end user versus what the data shows. Thus basis risk represents the “error” between what the data shows and the true damage at the farm or business level. Basis risk can be further broken down into two types of errors: first, the data itself may not match what happens on the ground and second, the crop production may not react to the realized weather condition in an anticipated manner.
The first type of basis risk is reduced by strong improvement in data availability and quality. Data availability and quality is a key foundation on which parametric insurance relies on. Without high resolution weather data the basis risk of the policy increases significantly. For example, if we can measure the weather only every 100km2, then there may be a drought or torrential rain in a 20km2 area without the data showing as such. This causes a policy to deliver sub-par results since the farmer purchasing the insurance faces adverse weather without the data triggering to make payoffs. Fortunately, weather data resolution has increased considerably over the last few decades and now global datasets such as CHIRPS by NASA produce rainfall estimates at a 0.05 degree (approx. 5km x 5km box) resolution. Part of the increase in resolution is due to higher density of weather stations while the availability of weather satellites adds to the ability to interpolate weather in areas where station density is low. This is especially useful in developing country regions such as central Africa and Southeast Asia where stations sparse. Finally, beyond government datasets, the increase of IoT sensors and personal devices such as cellphones holds the promise of vastly increasing the resolution of weather data going forward.
For parametric insurance in agriculture, the second source of basis risk stems from the uncertain link between the data being used for the insurance product and actual crop production. For example, the link between weather data and crop production can vary substantially by crop, region, type of weather data, and non-weather factors such as pests or crop diseases. Weather insurance for a farm is only relevant if the weather’s effect on production of the crop on the farm is understood clearly. In many areas, crop yield data itself is sparse making it difficult to formulate effective weather insurance policies. This issue is still to be resolved in many parts of the world for different crops where historical yield data is less available but overall we have better crop yield models for many crops and regions. For many crops such as wheat and corn, the effect of rainfall or temperatures can be understood in great detail allowing for more effective policies to be chosen. Another remedy is to use satellite data to directly infer crop health. Satellite vegetation indices have been tried in some parametric pilots but they bring the downside of reduced intuitive clarity for the end user.
Computing resources are required to collect data, process it, and determine payouts. Such resources have become exponentially cheaper and widely available in the last few years with the spread of cloud computing and data storage. At Arbol we believe blockchain and smart contracts are well suited to create, monitor, settle, and record parametric insurance contracts such as those reading weather data enabling widespread adoption.
Farmer adoption of parametric insurance products had long been a struggle given that the payouts and come a long way since its introduction a few decades ago. For a long time, educating farmers about purchasing insurance based on data instead of damage they directly suffer was a difficult task. Indeed, even the concept of insurance for crops was new in most places and to this day, is non-existent in many countries. The datasets used for parametric insurance were also relatively new for both users and insurers making it difficult to price products and for farmers to get an intuitive sense for whether the data would match actual experience. For example, if there is a drought in a local area, will the data show that? Certainly, this was related to the basis risk issue discussed above.
Over the last decade, farmer comfort with parametric insurance has increased in many locations. In the United States, a parametric insurance product that pays out in case of deficient rainfall over pasture and grazing land has become a top 10 USDA program covering over $2.5bn of a low value crop. Other examples of rapid adoption of data driven weather insurance for crops abound from France to Colombia.
At Arbol, we believe these examples demonstrate that the time is right for adoption of parametric insurance to increase rapidly in agriculture. Over $1 trillion of crops area estimated to have no insurance globally each year and the insurance industry clearly needs new solutions to provide financial security to farmers from the vagaries of weather. Many countries rely on disaster aid and loan waivers to help their farm sector but such methods are much more expensive than insurance and often slow or plagued by corruption. Most importantly, such means to help farmers are reactive and not proactive. To truly help the agriculture sector plan for the future and grow, farmers need to be empowered to get proactive solutions rather than waiting for uncertain payments after a disaster.