A novel way of approaching farm-specific climate change adaptation

Abstract

The agricultural sector is one of the sectors directly dependent on the climatic conditions of its surrounding environment. With climate change, farmers need to adjust their current farm management practices to adapt to new climatic conditions in order to moderate climate change impacts or exploit beneficial opportunities. To provide insights into these adaptation decisions and processes, it is important to use climate change impact models that account for the numerous adaptation strategies farmers possess to displace no-longer-advantageous activities under new climate conditions. As a result, the microeconomic approach most commonly used to assess the impact of climate change on agriculture is the Ricardian method. The Ricardian method accounts implicitly for adaptation options by assuming that farmers maximize their profits by optimizing all variables within their control. The only variables that then influence yields are exogenous variables outside the farmer’s control. The method uses cross-sectional data, as it assumes that farmers today have adapted to their current environment. As such, by looking at how farmers behave in response to their current environment, one can understand how farmers respond to climate change by comparing them with farmers in other climates.

Nevertheless, the Ricardian method’s ability to capture adaptation has to be treated with caution because it does not model adaptation explicitly. Endogenous farm management variables are not explicitly modeled by the method, as these are assumed to be optimized. As a result, the user of the results of the Ricardian method gains no insight into how a farmer adapts, making adaptation invisible and undefined. This is unfortunate for adaptation practitioners who need to gain more insight into the adaptation decision and implementation process itself. The goal of this dissertation is therefore specifically to improve the Ricardian method to make its results regarding climate change adaptation more defined and explicit for policy use (European policy use in particular). This is done by addressing four weaknesses of the Ricardian method.

First of all, the Ricardian method ignores adaptation requirements that need to be in place before a farmer can adapt. The method gives a false feeling of certainty that “farmers will adapt in the most optimal way” even though this is not always realistic, as not all farmers have access to such adaptation strategies. This is because farmers do not always possess all the necessary adaptive capacity characteristics such as information, skills, technology, economic wealth, and institutions that help them to implement adaptation strategies. Chapters 2 and 3 both resolve this weakness but in different ways. Chapter 2 suggests clustering farmers or regions based on pre-existing historical conditions that are assumed to influence farmers’ ability to adapt. As such, all farmers in one cluster are restricted to uniquely rely on only the adaptation strategies available in that cluster. For European farms, when clustered in Eastern versus Western Europe, this implies that Eastern Europe only has access to the adaptation strategies available in Eastern Europe. Depending on the climate change scenario, this results in an almost 50 percent loss in Eastern European land values compared to a 2 to 32 percent loss for Western Europe. As an alternative, chapter 3 suggests explicitly capturing a measurement of adaptive capacity as an additional variable in the model. This leads to more detailed results as adaptive capacity also diverges within the clusters built in chapter 2. Chapter 3 confirms that not taking into account adaptive capacity leads to too optimistic results, as positive marginal effects of temperature decrease between 2.5 and 5 percentage points in regions with a lower adaptive capacity. Clearly, there is a positive relationship between adaptive capacity and the agricultural climate response, even though chapter 3 shows this relationship is non-linear.

A second weakness of the Ricardian method is that it only accounts for adaptation options that are in the dataset. It does not account for future technological improvements, as these do not yet exist in the data. This is unfortunate for estimations of the impact of climate change on agriculture as technological improvements will play a very important role in climate change adaptation. Chapter 2 shows that by splitting technological development into two types (development due to existing technologies in developed countries, and development due to future technologies), it is already possible to account for technological development in regions in transition, based on existing technologies in more advanced regions. As such, by broadening the dataset with more advanced adaptation strategies from other regions, one can improve the climate response by taking into account future technological development based on existing technologies. The results in chapter 2 show that if Eastern Europe were to apply and implement the same adaptation options as Western Europe by 2100, it could avoid a 50 to 69 percentage point decrease in land value, depending on the climate scenario.

Unlike chapters 2 and 3, which focus on adaptation in general, the second part of this dissertation focuses specifically on one adaptation strategy, irrigation, in order to better understand in which contexts it should be prioritized compared to other adaptation strategies. Irrigation is one of the primary mechanisms by which agriculture can respond and adapt to climate change. Comparing irrigated versus rain-fed agriculture reveals a third weakness of the Ricardian method: There is heterogeneity within the adaptation option that influences its overall effectiveness. Farmers do not simply make one adaptation decision. In the case of irrigation, they consider water management options across a spectrum that ranges from purely rain-fed farms to purely irrigated farms. In between the extremes, there are, among others, farmers that use supplemental irrigation on only part of their field, farmers that apply conservation practices to store water in the soil, farmers that add more surface- or groundwater to their fields, and farmers that irrigate on a very frequent basis. Chapter 4 shows that by taking into account such within-adaptation-option differences (either by means of subsampling, or by means of an interaction term), differences in the effects of marginal changes in climate on farmers at the extremes of the irrigation spectrum can rise up to 30 percentage points, depending on the size of the farm.

Finally, because (as shown in chapter 4) the type of adaptation significantly influences the climate response and a farm has numerous adaptation strategies to choose from, it is important to tackle a fourth weakness of the Ricardian method by revealing the farm adaptation decision process itself. Because each farmer makes numerous decisions at one time, in chapter 5, this dissertation provides a unique simultaneous irrigation-crop decision model to illustrate the farm adaptation decision-making process. The model shows that the irrigation choice is highly crop-specific and that the farm irrigation probability is highly influenced by climate. Specifically, the model reveals that climate and water constraints often hamper the use of irrigation as an adaptation tool. Southern regions, for instance, show decreases in irrigation probability of up to 7 percent in summer, when temperature marginally increases. This shows that those regions adapt through other means than irrigation (for instance by means of crop choice). As a result, the conditional climate response of the different farm adaptation responses differs significantly between irrigated and rain-fed farms. In general, irrigated crops are more resistant to higher temperatures than rain-fed crops in southern regions. However, the model shows there is a difference between large and small farms as small farmers are more dependent on water access before they can irrigate. In addition to the insights this model provides, the model also appears to be more robust when compared to traditional cross-sectional models that do not capture irrigation explicitly.

These four chapters prove that it is possible to adjust the Ricardian method to reveal adaptation more explicitly. This more explicit view is important for policy as it leads to more insight and results that are more robust. It shows that the climate change responses of Western and Eastern Europe could be similar, on the condition that policy, society, and behavior are devoted to bringing forth equal and optimal adjustment and adaptation conditions over both regions. Policy and institutions should increase adaptive capacity to facilitate climate change adaptation. The EU should ensure sufficient investment in water management infrastructure, as well as ensuring water regulations as climate change will limit the usage of adaptation strategies that are dependent rain water. Nevertheless, adaptation through more drought-resilient crops should also be further encouraged as crop choice is a beneficial alternative to irrigation in water-scarce regions. Finally, policy should not merely scale up adaptation strategies that work in one region to a larger region. There are clear differences in the way different adaptation options are implemented, and policy should allow the execution of different adaptation strategies.