Effects of Climate Change on Crop Revenue in Kenya: Implications On Agricultural Production Joshua Laichena, E. Omosa and K. Musyoka Kenya Institute for Public Policy Research and Analysis Discussion Paper 355 2024 ii Effects of climate change on crop revenue in Kenya KIPPRA in Brief The Kenya Institute for Public Policy Research and Analysis (KIPPRA) is an autonomous institute whose primary mission is to conduct public policy research leading to policy advice. KIPPRA’s mission is to produce consistently high-quality analysis of key issues of public policy and to contribute to the achievement of national long-term development objectives by positively influencing the decision-making process. These goals are met through effective dissemination of recommendations resulting from analysis and by training policy analysts in the public sector. KIPPRA therefore produces a body of well-researched and documented information on public policy, and in the process assists in formulating long-term strategic perspectives. KIPPRA serves as a centralized source from which the Government and the private sector may obtain information and advice on public policy issues. Published 2024 © Kenya Institute for Public Policy Research and Analysis Bishops Garden Towers, Bishops Road PO Box 56445-00200 Nairobi, Kenya tel: +254 20 2719933/4; fax: +254 20 2719951 email: admin@kippra.or.ke website: http://www.kippra.org ISBN 978 9914 738 81 0 The Discussion Paper Series disseminates results and reflections from ongoing research activities of the Institute’s programmes. The papers are internally refereed and are disseminated to inform and invoke debate on policy issues. Opinions expressed in the papers are entirely those of the authors and do not necessarily reflect the views of the Institute. iii Abstract This study investigated the impact of climate and economic variables on crop production and the effects of climate change variation on crop revenues using a Ricardian model. The study used cross sectional survey data collected by Agriculture Sector Development Support Programme II (ASDSP II) in 2013 of 12,651 households across Kenya and the historical climate data from the Kenya Meteorological Services. Public Expenditure Review and Analysis for climate change adaptations and mitigation in agriculture (PERCC) methodology was used to identify climate relevant public expenditure in agriculture. This was done by using the UN Food and Agriculture Organization’s (FAO) Monitoring and Analysing Food and Agricultural Policies (MAFAP) data for the period 2013/14 to 2017/18. The findings of the study provide evidence that climate change significantly impacts crop revenue in Kenya, showing that farming in Kenya is sensitive to temperature and precipitation changes. The study further highlights the need for effective adaptation policies to reduce vulnerability and enhance resilience in the face of a changing climate. These findings are warnings that severe climate scenarios could lead to substantial losses in crop yields, particularly crop revenue, affecting the most vulnerable regions and posing a major threat to the agricultural sector, which is a key contributor to Kenya’s Gross Domestic Product (GDP). Based on these findings, it is recommended that the government focuses on various policy measures. First, strengthening climate adaptation measures to safeguard the agricultural sector by ensuring allocation and tracking of the funding for adaptation in the sector, since the sector is vital for Kenya’s GDP. Secondly, to mitigate the potential adverse effects of climate change, it is imperative to design and implement robust climate adaptation policies that prioritize enhancement of agricultural resilience through adoption of improved farming inputs, advanced technologies, and comprehensive training and extension services. Lastly, to improve the management and tracking of climate financing in the agriculture sector, the government needs to ensure that funds are effectively used for their intended purposes in both adaptation and mitigation efforts to achieve food and nutrition security, as outlined in the Bottom-up Economic Transformation Agenda (BETA) plan and the 4th Medium Term Plan (MTP IV). iv Effects of climate change on crop revenue in Kenya Abbreviations and Acronyms AEZs Agro-ecological Zones APSIM Agricultural Production Systems sIMulator ASDSP II Agriculture Sector Development Support Programme II ASTGS Agricultural Sector Transformation Growth Strategy BETA Bottom-up Economic Transformation Agenda CCCF County Climate Change Fund CCM Canadian Climate Model CPI Climate Policy Initiative DSSAT Decision Support System for Agrotechnology Transfer EDE Ending Drought Emergencies FAO Food and Agricultural Organization GCF Green Climate Fund GCM Global Circulation Model GDP Gross National Product GEF Global Environmental Facility GFDL Geophysical Fluid Dynamics Laboratory Model GHGs Green House Gases GoK Government of Kenya IPCC Intergovernmental Panel on Climate Change KALRO Kenya Agricultural and Livestock Research Organization KIPPRA Kenya Institute for Public Policy Research and Analysis KMS Kenya Meteorological Services KNBS Kenya National Bureau of Statistics LDCF Least Developed Country Fund LULUCF Land Use Land Use Change and Forestry MAFAP Monitoring and Analysing Food and Agricultural Policies MAM March April May MTP Medium Term Plan MRV Monitoring Reporting and Verification NAP National Adaptation Plan NAPA National Adaptation Plan of Action NASA National Aeronautics and Space Administration NCCAP National Climate Change Action Plan NCCRS National Climate Change Response Strategy NDC Nationally Determined Contribution OECD Organization for Economic Cooperation and Development OND October November December PEA Public Expenditure Analysis PERCC Public Expenditure Review for climate change adaptations and mitigation in agriculture SAPs Sustainable Agricultural Practices SCCF Special Climate Change Fund SEI Stockholm Environment Institute UN United Nations UNFCCC United Nations Framework Convention on Climate Change WMO World Meteorological Organization WOFOST World Food Studies simulation model v Table of Contents Abbreviations and Acronyms..................................................................................iv 1. Introduction................................................................................................1 2. Level and Composition of Spending in Support of Food and Agriculture.................................................................................................4 2.1 Background.................................................................................................4 2.2 Climate Change Public Expenditures in Support of Food and Agriculture Sector.........................................................................................................5 3. Literature Review......................................................................................12 3.1 Theoretical Framework.............................................................................12 3.2 Empirical Review on Climate Change Impact on Agriculture.......................13 3.3 Climate Financing Mechanisms................................................................14 4. Method and Model Specification...............................................................16 4.1 The Ricardian Method...............................................................................16 4.2 Empirical Model Specification...................................................................17 4.3 Kenya’s Planting Calendar by Region and Crop Stage.............................19 4.4 Measurement of Variables.........................................................................21 4.5 Data and Data Sources..............................................................................26 4.6 Descriptive Statistics................................................................................27 5. Results and Discussion.............................................................................29 5.1 Results of Estimation of Response of Crop Revenue to Climate...............29 5.2 Marginal Effects........................................................................................32 6. Conclusion and Policy Implications.........................................................35 6.1 Conclusion................................................................................................35 6.2 Policy Implications...................................................................................36 References..............................................................................................................38 Appendices.............................................................................................................45 Appendix 1: Assessment of climate finance flows to agriculture sector using MAFAP data methodology.....................................................................................45 Appendix 2: Ricardian regression results..............................................................47 vi Effects of climate change on crop revenue in Kenya List of Figures Figure 2.1: Public expenditures in support of food and agriculture sector and climate change related spending (Ksh billion), 2013/14- 2017/18.......6 Figure 2.2: Climate-relevant agricultural public expenditure positively supporting adaptation, 2013/14-2017/18 (Ksh billion)............................................7 Figure 2.3: Agriculture-specific/ agriculture-supportive public expenditures positively supporting adaptation (Ksh billions) 2013/14-2017/18........8 Figure 2.4: Average composition of agriculture specific public expenditure positively supporting adaptation (2013/14-2017/18)............................8 Figure 2.5: Composition of agriculture-supportive public expenditures positively supporting adaptation, 2013/14-2017/18 (average)..............................9 Figure 2.6: Average agriculture specific/agriculture supportive public expenditure with link to mitigation, Ksh billion (2013/14-2017/18)........................10 Figure 2.7: Average Composition of public expenditures enhancing mitigation, 2013/14-2017/18...................................................................................10 Figure 2.8: Average composition of public expenditures with adverse effects on mitigation, 2013/14-2017/18................................................................11 Figure 2.9: Comparison between budget and actual climate relevant expenditures (2013/14- 2017/18)................................................................................................11 List of Tables Table 2.1: Main actions related to agriculture and associated budget for NAP, 2015-2030..............................................................................................5 Table 4.1a: Kenya planting calendar by region and crop stage (maize).................19 Table 4.1b: Kenya planting calendar by region and crop stage (beans and other crops)...................................................................................................20 Table 4.1c: Counties distribution per region..........................................................21 Table 4.2: Definition and measurement of variables.............................................21 Table 4.3: Description of climate variables............................................................24 Table 4.4: Descriptive statistics for model variables.............................................27 Table 5.1: Response of crop revenue to climate variables, adaptation and farmer and socioeconomic variables...............................................................29 Table 5.2 Marginal effects of seasonal temperature and precipitation on crop yield per acre.................................................................................................32 Table 5.3: Forecasted effects of climate change on crop yields and revenue........34 1 1. Introduction Many climate models indicate that climate change is expected to negatively impact crop yields, revenue, and agricultural production in Africa by up to 50 per cent by 2020 ( Mendelsohn and Seo, 2007). A range of Global Circulation Models predicts a warming of above 20C in most of East Africa and an increase in precipitation of 6.3 per cent, coupled with large regional variations in precipitation. Given a temperature rise of 2.50C the biophysical impacts of climate change will reduce food availability, undermine food security and expose between 55 and 65 million extra people to the risk of hunger by the year 2080 (Shi Zhen et al., 2020). The disastrous effects of global climate change are becoming more severe, and most of the damage is expected to occur in underdeveloped countries. Compared to developed and other developing nations, African nations are more vulnerable to the effects of climate change, because they are highly dependent on rain-fed agriculture and their limited institutional, financial, and technical capabilities for adaptation makes them less able to withstand adverse effects of climatic changes (Bruckner, 2012; Dessalegn and Akalu, 2015; Gemeda and Debela Hunde Feyssa, 2015). Climate change has a profound impact on agriculture, influencing crop revenue. As global temperatures rise, the frequency of extreme and unexpected weather events such as heatwaves, droughts, and floods increase, posing significant challenges to agricultural production, and therefore crop revenue (Jenkins et al., 2021). While some regions may experience marginal benefits in crop yields due to factors such as increased CO2 levels, which can enhance photosynthesis, these benefits are often offset by the negative effects of higher temperatures and altered precipitation patterns (Ibid; McLachlan and Zheng, 2020). For instance, studies have shown that in countries like China, India, Brazil, Egypt, Ghana, and Ethiopia, climate-induced changes in crop yields, and therefore crop revenue ,can result in varying economic outcomes (Jenkins et al., 2021). In some scenarios, GDP and welfare indicators may initially experience marginal increases due to improved yields; however, as warming levels rise beyond certain thresholds, these trends begin to reverse, resulting in higher consumer prices and decreased national economic welfare (McLachlan, van, and Zheng, 2020). It must be noted, however, that the relationship between climate change and crop revenue is complex, as it is influenced by both natural and socio-economic factors, and the overall effect is highly dependent on the specific regional climate impacts and the resilience of the agricultural systems in place (Ibid). Empirical studies have predicted how climate change affects yields, and therefore, crop revenue. In China, for example, a study assessing the economic effects of climate-induced crop yield changes showed that there were marginal benefits on GDP and welfare up to certain levels of warming due to projected increases in rice yields, which lowered domestic consumer rice prices. However, at higher warming levels, these trends began to reverse. In contrast, India faced negative impacts due to declining crop yields, leading to increasing consumer prices of domestic and imported rice and wheat, and declines in GDP and welfare, especially at higher warming levels (Wang, D., et al., 2021). Furthermore, a study by the National Aeronautics and 2 Effects of climate change on crop revenue in Kenya Space Administration (NASA) predicted that climate change might affect the production of maize (corn) and wheat as early as 2030 under a high greenhouse gas emissions scenario. Maize crop yields are projected to decline by 24 per cent, while wheat could potentially increase by about 17 per cent. The change in yields for maize and wheat is due to projected increases in temperature, shifts in rainfall patterns, and elevated surface carbon dioxide concentrations from human-caused greenhouse gas emissions (Müller et al., 2021). In the United States (US), climate change is boosting maize yields in parts of the US, Latin America, and Asia, but sharply reducing them elsewhere. It is also reducing the US soybean yields in southern and eastern states and expanding them to the north and west (Ray et al., 2019). A study in India and Pakistan, shows that extreme heat may lead to a decline in crop yields, which, combined with the banning of wheat and rice exports in India, posed a threat to international food markets and countries already affected by shortages of staple foods (WMO, 2023). The above studies illustrate the varied and significant ways in which climate change can affect crop revenue, with implications for both local economies and global food security. However, the specific impacts are highly dependent on regional climate conditions and the resilience of agricultural systems. At a country level, past studies have indicated that temperatures throughout Kenya have generally risen mainly near the large water bodies (King’uyu, Ogallo and Anyamba, 2000; GoK, 2010). Climate projections also show that the country will experience an annual temperature increase ranging from 10C to 3.5°C by the year 2050 (SEI, 2009). Climate change associated with unpredictable rainfall, reduced soil productivity through erosion, and increased evapo-transpiration is responsible for declining agricultural production (GoK, 2010). Kenya is already experiencing climate-change characterized by more frequent and intense extreme weather events, particularly drought and flood (GoK, 2012). The effects of climate change are frequently discussed in terms of rising temperatures and shifting precipitation patterns whereas it is more crucial how these changes affect agricultural production (Kiremu et al., 2022). Climate change also increases the risk of rural populations who derive majority of their livelihood from agriculture, while at the macro-scale; it increases the vulnerability of the economy, which is dominated by climate sensitive sectors such as agriculture (GoK, 2010). Due to its significant contribution to the GDP, growth in the agriculture sector is correlated with growth of the overall national economy. For instance, in 2013, the sector accounted for 65 per cent of the country’s exports, and more than 70 per cent of total national labour force (KIPPRA, 2014). Declining productivity is mainly attributed to unfavourable climatic weather conditions and high cost of farm input. Further, agriculture is sensitive to climate change due to the close natural link between climatic weather conditions and plant development. The most important climate variables that affect agricultural production are temperature, precipitation, atmospheric pressure and humidity, wind and sunshine and cloud cover. Agricultural production potential is determined by physical factors, primarily soil and climatic conditions, and a complex interaction of socioeconomic, cultural and technological factors. These factors include farm sizes, level of farming and livestock inputs and management practices such as soil 3 conservation and enhancement, and veterinary services. Furthermore, economic factors like market prices and access to credit, education and extension services also play a role in determining production potential (FAO, 2016). Under extreme weather conditions, resulting to severe drought in Kenya, the government is forced to allocate funds for emergency food relief, pointing to a close link between climate change, agricultural production and financing for mitigation and adaptation. Addressing climate-related effects require urgent action on a local and national scale. Responding to negative effects of climate change, and preparing for future climate change is estimated to cost US$500 million per annum, with adaptation cost alone increasing to between US$1 to US$2 billion per year by the year 2030 (SEI, 2019). The Kenya National Adaptation Plan (2015-2030) estimates the cost of adaptation in the agriculture sector alone to cost US$375 million and for all priority sectors in the plan to cost a total of US$36,136 million. These estimates are beyond what the country can mobilize domestically, and hence, leveraging domestic resources with international climate finances is therefore imperative. On the other hand, investments in climate mitigation and adaptation projects were estimated to cost US$2.4 billion in 2018, which came from both domestic and foreign sources. About 42.2 per cent (Ksh 102.7 billion) of the total climate finance tracked, came from domestic sources with domestic private and public sector sources accounting for 14 per cent (Ksh 34 billion) and 28.3 per cent (Ksh 68.8 billion), respectively (Odhengo et al., 2021). Domestic sources have recently become important as the government allocates more resources to climate activities. However, there are concerns on whether the available resources can address the climate change problem in agriculture. Although public expenditures are an important policy instrument in addressing climate change mitigation and adaptation, the linkages between agricultural production and climate finance in the sector for both mitigation and adaptation are poorly understood. Previous studies have concentrated on assessing the economic cost of climate change, from the demand side ignoring the supply side. This study aimed to bridge this gap by investigating how variations in climate change and other economic variables affect the crop yield and revenue in Kenya. The study aimed to answer the following questions: first, how do variations in climate change and other economic variables affect crop yield and, therefore, crop revenue in Kenya? Secondly, what are the effects of variations in climate change on crop yields in Kenya? Finally, what are the sources and composition of climate financial flows to the agriculture sector in Kenya? The study aimed to address three objectives, namely: examining the relationship between climate change and other economic variables on crop yields and crop revenue in Kenya; forecasting the effects of climate change variation on crop yields and, therefore, crop revenue; and identifying climate financial flows and their composition to the agriculture sector in Kenya. Introduction 4 Effects of climate change on crop revenue in Kenya 2. Level and Composition of Spending in Support of Food and Agriculture 2.1 Background The importance of agriculture in Kenya is re-emphasized through the Kenya Vision 2030, the Bottom-up Economic Transformation Agenda (BETA) and the Fourth Medium Term Plan (MTP IV - 2022/23 to 2027/28) in which agriculture, along with manufacturing, housing and healthcare, are given priority to ensure inclusive growth, job creation and food security for all. In line with this framework, the Agricultural Sector Transformation and Growth Strategy (ASTGS) guides the agriculture sector interventions until 2029. The Strategy identifies three key pillars of improved agricultural performance including, increasing the incomes of small- scale farmers, pastoralists, and fishermen; increasing agricultural production and value-added; and boosting household resilience to food insecurity as well as 'enablers' that are necessary to guarantee reaching the objectives defined within the pillars. The specific actions to be undertaken by the government are specified within 'flagships,' six of which are defined under each of the pillars and three belonging to the groups of enablers. This is further reinforced by MTP IV, which identifies various value chain development and agro-processing for prioritized crops and livestock. Climate change issues are addressed in economy-wide strategies that specify contributions of individual sectors, including agriculture. Kenya launched the National Climate Change Response Strategy (NCCRS) in 2010 and a National Climate Change Action Plan (NCCAP 2013-2017) in 2013. In 2018, the action plan was renewed into the National Climate Change Action Plan (NCCAP) for 2018- 2022. Kenya's third Action Plan on climate change (NCCAP 2023-2027) builds on the previous Action Plans and provides a framework for Kenya to deliver on its Nationally Determined Contribution (NDC) under the Paris Agreement and transition the country into a low carbon climate resilient development pathway. Climate change adaptation was set as a key priority for the country recognizing the adverse socioeconomic impacts related to climate change and the increasing vulnerability of the different sectors. It was also recognized that adaptation and development goals need to complement each other to achieve sustainable development. The National Adaptation Plan (NAP), developed for 2015-2030, builds on the foundations laid out in the NCCRS and the NCCAP. It forms the basis for the adaptation component of Kenya’s Intended Nationally Determined Contribution (INDC) submitted to the United Nations Framework Convention on Climate Change (UNFCCC) Secretariat. The actions proposed in the NAP complement adaptation actions that are ongoing through various projects and programmes already implemented by the public and private sector. Related to the agriculture sector, the NAP proposes to reinforce the land reforms and ensure sustainable land use; protect the environment to secure livelihoods, health and ecosystem services, among others; enhance the resilience of the food crops, industrial crops and horticulture value chains; enhance the 5 Level and composition of spending in support of food and agriculture resilience of the livestock value chains; enhance the resilience of the fisheries value chains; and fast track the Common Programme Framework for Ending Drought Emergencies (EDE) 2012-2022. These priority actions related to agriculture are elaborated in Table 2.1. Table 2.1: Main actions related to agriculture and associated budget for NAP, 2015-2030 Sector Action in NAP 2015-2030 Budget (US$ m) Land reforms Mainstreaming climate change adaptation in land reforms 1.40 Environment Mainstreaming climate change adaptation in the environment sector 636.10 Crops Enhance the resilience of crop value chain 375.10 Livestock Enhance the resilience of livestock value chain 299.80 Fisheries Enhance the resilience of fisheries value chain 136.90 Common Programming Framework for EDE (2010- 2022) Fast track the implementation of the EDE common programme framework 2,118.30 Total budget allocated to actions related to agriculture 3,567.60 Source: Government of Kenya (2016) However, it is important to note the amount of funding required for climate proofing on the sectors to achieve the proposed actions for various sectors in the NAP. Building on NAP main actions, it is necessary to track the funding for climate change in the agriculture sector to monitor and track the funding that is directed towards adaptation and mitigation in the agriculture sector. In the next section, climate relevance funding is tracked using the FAO’s database on Monitoring and Analysing Food and Agricultural Policies (MAFAP) for the period 2013/14 to 2017/18. 2.2 Climate Change Public Expenditures in Support of Food and Agriculture Sector The adaptation aspect of climate change has been defined by various studies as the process of adjusting to the current and future effects of climate change (Barnes et al., 2020; Owen, 2020). On the other hand, mitigation measures of climate change has been defined as preventing or reducing the emissions of greenhouse gases into the atmosphere to reduce the adverse effects of climate change (Amelang et al., 2020). Most of the public expenditures in support of the food and agriculture sector in Kenya affect the sectors capacity to adapt to climate change (Figure 2.1). The 6 Effects of climate change on crop revenue in Kenya information gathered in the MAFAP database allows for classifying about 99 per cent of public expenditures as relevant to climate change adaptation and mitigation (referred to as climate relevant), on average in the considered period. The amount of climate relevant spending increased throughout the period of analysis from Ksh 114 billion (US$0.912 billion) to nearly Ksh 280 billion (US$2.24 billion), with an average annual growth rate of about 20 per cent. The climate relevant spending in agriculture constitutes, on average, 14 per cent of total government spending in all sectors of the economy. Figure 2.1: Public expenditures in support of food and agriculture sector and climate change related spending (Ksh billion), 2013/14- 2017/18 Source: Calculations based on MAFAP database, 2018 2.2.1 Climate change adaptation and mitigation definitions The Public Expenditure Review for climate change adaptations and mitigation in agriculture (PERCC) adopts FAO’s definitions of adaptation and mitigation. FAO defines adaptation as “the vital response to the adverse effects of climate change and the preparation for future impacts”. This includes an adjustment in natural or human systems in response to actual or expected climatic stimuli or their effects, which moderates harm or exploits beneficial opportunities as noted by the Intergovernmental Panel on Climate Change (IPCC, 2014). In agriculture, adaptation actions encompass technological responses, enhancing smallholder access to credit and other critical production resources, and strengthening institutions at local and regional levels. Specific responses consist of developing new crop varieties adapted to changes in carbon dioxide, temperature and drought, fostering the capacity for climate risk management, offsetting economic impacts of land use change, crop insurance, and information systems to support early warning and proactive planning. 7 Mitigation, includes all the “human interventions to reduce the emissions of Green House Gases (GHGs) by sources or to enhance their removal from the atmosphere by sinks (for example, forests, vegetation or soils that can reabsorb carbon dioxide (CO2).” Mitigation measures in agriculture include technological innovation and transfer, crop diversification, climate-smart agricultural practices to increase soil quality and decrease soil erosion (IPCC, 2014). Within climate-relevant spending, an evaluation of expenditure positively linked to climate adaptation was performed. Figure 2.2 shows that on average, 86 per cent of public expenditures are positively linked to climate change adaptation. This share was relatively stable throughout the years, while the underlying amount increased on average by 24 per cent per year from Ksh 105.3 billion in 2013/14 to Ksh 248.5 billion in 2017/18. Figure 2.2: Climate-relevant agricultural public expenditure positively supporting adaptation, 2013/14-2017/18 (Ksh billion) Source: Calculations based on MAFAP database, 2018 The average composition of public expenditures positively supporting adaptation was investigated to determine the type of adaptation-enhancing measures that are employed. First, a broad disaggregation between agriculture-specific and agriculture-supportive measures shows that a higher share of measures in favour of adaptation is spent on agriculture-supportive activities, ranging from 60 to 70 per cent in the periods examined. Agriculture-specific measures enhancing adaptation increased from Ksh 42.5 billion in 2013/14 to Ksh 66.2 billion in 2017/18 (Figure 2.3). Level and composition of spending in support of food and agriculture 8 Effects of climate change on crop revenue in Kenya Figure 2.3: Agriculture-specific/ agriculture-supportive public expenditures positively supporting adaptation (Ksh billions) 2013/14- 2017/18 Source: Calculations based on MAFAP database, 2018 Further disaggregation focusing on the average composition of agriculture-specific public expenditures positively stimulating the adaptive capacity of the agriculture sector (Figure 2.4) was done, where the largest share, 21 per cent was given to other off-farm infrastructure (including off-farm irrigation, which contributes the most, and feeder roads), followed by spending in other general support to the food and agriculture sector 15 per cent. Figure 2.4: Average composition of agriculture specific public expenditure positively supporting adaptation (2013/14-2017/18) Source: Calculations based on MAFAP database, 2018 9 The adaptation supportive measures spent 13 per cent on subsidies for improved quality inputs, including variable inputs such as improved seeds or fertilizers, which contributes the most, support to building on-farm capital and on-farm services provision, on agricultural research, and payments to consumers (cash transfers above all). Lastly, inspection, training, extension and technology transfers, other payments to producers, storage, marketing and school feeding programmes contribute smaller shares, generally renging from 2.0 to 5.0 per cent. The analysis of the average composition of agriculture-supportive public expenditures positively supporting adaptation (Figure 2.5) reveals that rural education takes the largest share (47%), while rural health and rural roads accounts for 19 per cent and 18 per cent, respectively. Rural water and sanitation and rural energy make up the remaining share (8%) each. Figure 2.5: Composition of agriculture-supportive public expenditures positively supporting adaptation, 2013/14-2017/18 (average) Source: Calculations based on MAFAP database, 2018 To examine the average composition of public expenditures linked to climate change mitigation, the agriculture-specific and agriculture-supportive spending are disaggregated. The analysis revealed that mitigation-related spending concerns mainly the agriculture-specific measures, while agriculture supportive measures account for a relatively lower share (Figure 2.6). Level and composition of spending in support of food and agriculture 10 Effects of climate change on crop revenue in Kenya Figure 2.6: Average agriculture specific/agriculture supportive public expenditure with link to mitigation, Ksh billion (2013/14-2017/18) Source: Calculations based on MAFAP database, 2018 Among the public spending positively supporting climate change mitigation, agriculture-specific measures 70 per cent on average, and the main contribution was 'other general support to the food and agriculture sector' category with a share of 46 per cent, followed by agricultural research with 12 per cent (Figure 2.7). Other off-farm infrastructure, variable inputs, training and capital contribute between 2.0 and 4.0 per cent. Agriculture-supportive measures accounted for the remaining 30 per cent, consisting of spending on rural energy (27%) and rural water and sanitation (3%). Figure 2.7: Average Composition of public expenditures enhancing mitigation, 2013/14-2017/18 Source: Calculations based on MAFAP database, 2018 Most of the public expenditures with adverse effects on climate change mitigation are agriculture-specific (94% on average) driven by expenditures in the livestock 11 subsector, and the main spending components are inspection activities, other general support to the food and agriculture sector and other off-farm infrastructure (32%, 24% and 19%, respectively). Marketing accounts for 12 per cent, while other payments to producers, input subsidies and technical assistance contribute 2-3 per cent each (Figure 2.8). Figure 2.8: Average composition of public expenditures with adverse effects on mitigation, 2013/14-2017/18 Source: Calculations based on MAFAP database, 2018 2.2.2 Budgeted climate relevant amounts versus actual spending The budgetary allocations for climate relevant expenditures were higher than actual expenditures in every year considered (Figure 2.9). The budgeted amounts for the climate relevant spending in agriculture were generally lower than revised budget allocations. With an actual average execution rate of about 77 per cent, the spending shows that an important part of the allocations (about 33%) was not disbursed for implementing the planned activities. Adaptation enhancing measures received an average of about 77 per cent of the total allocated fundings. Figure 2.9: Comparison between budget and actual climate relevant expenditures (2013/14-2017/18) Source: Government of Kenya, 2021 Level and composition of spending in support of food and agriculture 12 Effects of climate change on crop revenue in Kenya 3. Literature Review 3.1 Theoretical Framework There are two methods that have been developed during the last few decades to evaluate the impact of climate change on the agricultural sector: the production- function approach (Rosenzweig and Iglesias, 1999), and the Ricardian approach (Mendelsohn et al., 1994). The production-function approach relies on empirical or experimental production functions to predict environmental damage. This approach has been criticized for inherent bias, which tends to overestimate the damage, a bias sometimes referred to as the ‘dumb farmers scenario’, since it fails to consider the variety of adaptations that farmers might make in response to changing economic and climatic conditions (Mendelsohn et al., 1994). The Ricardian approach is a cross-sectional model used to study agricultural production by measuring climate change damage as a reduction in net crop revenue or land value. In addition, it takes into account the costs and the benefits of different adaptation techniques that farmers apply. The Ricardian approach is based on several assumptions. First, it is assumed that climate shifts the production function for crops. Second, there is perfect competition in both product and input prices (no public intervention in the market and no monopoly). Third, the land values have attained the long run equilibrium associated with each region’s climate. Fourth, market prices are unchanged because of the change in environmental conditions. Fifth, adaptation takes place by all means including the adoption of new crops or farming systems. And finally, the adaptation cost is not considered in the analysis. Conventional approaches to assess the impacts of climate change on agriculture rely on complex crop simulation models that compare crop yields under different climatic conditions (Torriani at al., 2007; Adams et al., 1999; Ochieng et al., 2016). Most models use calibrated crop models from controlled experiments in which crops are grown in a laboratory setting that simulates different climates and levels of carbon dioxide. By adjusting different climate parameters while keeping the farming methods unchanged, the effects of climate change on yields are established. Elbehri and Burfisher (2015) have categorized approaches for analyzing the impacts of climate change and variability on agriculture into two: pathway models and Ricardian models. Pathway models quantify biophysical effects of changes in temperature, precipitation and carbon dioxide on crop yields. They generate variables used to shock economic models. Two types of crop yield are identified: dynamic crop growth simulation and statistical yields. Dynamic crop growth models simulate incremental plant growth process and yields in response to changes in climate conditions. They include Decision Support System for Agro-technology (DSSAT), which has been applied by studies such as Thornton et al.(2011) and Nyang’au et al. (2014). Other dynamic models include the World Food Studies (WOFOST) simulation model (Alvaro et al., 2010) used for analysing the growth and production of field crops under a wide range of weather and soil conditions. FAO’s Aqua Crop (Miller et al., 2008) and the Agricultural Production Systems Simulator (APSIM) a wiki-modelling framework. 13 Statistical crop yields models describe empirical relations between observed crop yields and projected changes in climate variables. These have advantages over dynamic crop growth models since they take into account farmers adaptive behaviours and can be estimated at different time and spatial scales. These models have been used by Lobell et al. (2007, 2008, 2010) and Rowhani et al. (2011). Crop models are extremely data intensive and often predict severe yield reductions resulting from climate change because of the ‘dumb farmer effect’ whereby they ignore adaptation actions, which farmers would ordinarily take in response to a changing environment (Mendelsohn et al., 1994). Economic models have emerged to address this weakness. These are used to predict aggregate crop outputs, prices and net revenue using the yields from the agronomic models (Mendelsohn and Dinar, 1999). The Ricardian model has got its advantages and weaknesses. One of its fundamental advantages is its ability to incorporate private adaptations. Farmers adapt to climate change to maximize profit by adjusting the crop mix, planting and harvesting dates, and various agronomic practices. Another advantage of the model is its cost-effectiveness, as secondary data on cross-sectional sites can be relatively easy to collect regarding climatic, production, and socioeconomic factors. One of the weaknesses of the Ricardian approach is that it is not based on controlled experiments across farms. Farmers’ responses vary across space not only because of climatic factors, but also because of many socioeconomic conditions. Such non- climatic factors are seldom fully included in the model. Attempts have been made to include soil quality, market access and solar radiation to control for such effects (Mendelsohn et al., 1994; Kumar and Parikh, 1998). The other weakness of the Ricardian model is that it does not include price effects (Cline, 1996). It also does not account for price changes as it assumes that prices are implicitly constant (Elbehri and Burfihser, 2015). 3.2 Empirical Review on Climate Change Impact on Agriculture Despite the weaknesses, the Ricardian model has been used in many studies. (Ali et al., 2021; Hossain et al., 2019) employed the Ricardian model in estimating the economic effects of climate change on the net revenue from crop cultivation in Pakistan and Bangladesh. The findings demonstrated that crop revenue is susceptible to climate change and fluctuation in both countries. Net revenue losses are highly correlated with a rise in the yearly average temperature and a decrease in rainfall. While an increase in temperature is expected to hurt net revenue, an increase in precipitation is seen to have beneficial impacts. In Bangladesh for instance, marginal impact estimates indicate that temperature rise in agro-ecological zones (AEZs) with adequate irrigation systems was shown to be favorably influencing crop revenue. However, the effects will differ greatly depending on cropping season and location. The studies further revealed that not all the zones would be affected equally by expected climatic changes. In Sub-Saharan Africa, similar studies have been undertaken, some supported by the World Bank (Benhin, 2008; Dessalegn and Akalu, 2015; Gbetibouo and Literature review 14 Effects of climate change on crop revenue in Kenya Hassan, 2005; Gemeda and Debela Hunde Feyssa, 2015; Herrero et al., 2010; Kogo et al., 2021; Mano and Nhemachena, 2007; Molua et al., 2007; Mulwa et al., 2016; Ochieng et al., 2016; Mariara and Karanja, 2007), on estimating economic impacts of climate change on agriculture for various countries. The studies also investigate whether there are any significant differences in the effects between irrigated farms and dryland farms and between large-scale and small-scale farms. A synthesis of these studies and a similar one in South America by Mendelsohn (2009) concluded that farmers in the tropical and sub-tropical regions suffer greater losses from even marginal changes in climate than those in temperate regions. The studies also conclude that climate change is expected to have mixed impacts with some of the AEZs gaining and others losing. Climatic variables (temperature and precipitation) have significant effects on crop revenues. The analysis indicates that crop revenues are affected negatively by increases in temperature and positively by increases in precipitation. In South Africa for instance, (Benhin, 2008) found that the differences between the impacts on large-scale farms and small-scale farms were not very clear-cut, because they are overshadowed by the impacts of whether a farm is irrigated or not. The results also revealed seasonal differences in the impacts. An increase in temperature will affect crop net revenues negatively in the summer farming season but positively in the winter season. The studies also conclude that under current conditions, farmers practicing irrigated agriculture experience less damage than dryland farmers showing that seasonality and moisture availability frequently constrain agriculture. 3.3 Climate Financing Mechanisms There is no internationally agreed definition of climate finance (Bucher et al., 2011) but broadly, it refers to financial support for mitigation and adaptation activities, including capacity-building, research and development, and broader efforts to enable the transition towards low-carbon, and climate-resilient development. In addition, climate financing as defined by the United Nations Framework Convention on Climate Change (UNFCCC) refers to local, national, or transnational money from public, private, and alternative sources aimed at supporting climate change mitigation and adaptation efforts. Furthermore, Climate Policy Initiative (CPI), 2013 defines climate finance as specific capital flows from developed to developing countries with direct or indirect greenhouse gas mitigation or adaptation objectives towards a low-carbon and climate-resilient development pathway. Climate finance at the international arena emerged following the Copenhagen climate talks in 2009. The Paris Agreement reinforced industrialized countries' commitment from the 2009 Copenhagen Accord to mobilize US$100 billion per year by 2020 to fund climate change mitigation and adaptation in developing countries. Climate finance is still essential to spur mitigation and adaptation efforts in developing nations with severe climate change consequences. Strong financial structures that comprise systems, initiatives, and programmes supporting mitigation and adaptation measures are necessary for an effective response to climate change problem. 15 As implementation of the Paris Agreement on climate policy gathers momentum worldwide, it is becoming increasingly clear that developing African countries will need efficient and enabling financial and technical help as finance is a key factor in determining a country’s response to climate change (Odhengo et al., 2019). Since 2010, many developed countries have channeled climate financial support through bilateral and multilateral agencies, and other international organizations. In the last few years (2010-2018), many funds have emerged to finance climate change activities in developing countries. Some estimates show that annual global climate finance may be as high as US$359 billion (Buchner et al., 2019). Sixty-two per cent (62%) of these funds were from the private sector. Overall, the climate finance flows are lower than required investment needed to shift global economy to a below 2°C pathway. Climate funds flow through several mechanisms such as multilateral funds and other climate finance initiatives through bilateral institutions. These support low-carbon and climate-resilient projects through diverse economic and financial instruments, including policy incentives; risk management; grants; low-cost debt; and capital instruments, such as project-level market rate debt, project-level equity, and balance sheet financing. Similarly, multiplicity funding avenues complicate the process of accessing the climate finances, while at the same time making the process of monitoring, reporting and verification (MVR) difficult. Under the UNFCCC, two funds, the Least Developed Country Fund (LDCF) and the Special Climate Change Fund (SCCF), exist to support mitigation and adaptation needs of developing countries. LDCF addresses the special needs of the Least Developed Countries by financing the preparation and implementation of the National Adaptation Programmes of Action (NAPAs). The NAPAs outline country priorities for adaptation actions in sectors and resources that are central to human and socioeconomic development. More than US$10 million is available for NAPAs and countries can apply for full cost funding of up to US$200,000 (World Bank, 2013). The SCCF was established in 2001 to support activities, programmes and measures under four financing windows namely: adaptation to climate change; technology transfer; mitigation in selected sectors including: energy, transport, industry, agriculture, forestry and waste management; and economic diversification. In addition, the fund supports countries to prepare their (initial) National Communications to the UNFCCC and strengthen implementation of related adaptation activities. The SCCF is characterized by lack of adequate and predictable resources amid a growing demand (GEF, 2013). Kenya has the potential to access a variety of climate financing tools and windows like different bilateral funds that are outside the framework of the UNFCCC (Odhengo et al., 2019). Other opportunities include those that are part of the UNFCCC multilateral funds framework like the Green Climate Fund (GCF), Global Environment Facility (GEF), and Adaptation Funds. The country could also take advantage of domestic sources of climate finance such as corporate investments and national budgetary allocations. Literature review 16 Effects of climate change on crop revenue in Kenya 4. Method and Model Specification 4.1 The Ricardian Method This study applied the Ricardian method developed by Mendelsohn et al. (1994) to measure the value of climate in US agriculture. This analysis is based on the assumption of a direct cause and effect relationship between climate events and farm value. Ricardian approach is preferred to the traditional estimation methods, given that instead of ad hoc adjustments of parameters that are characteristic of traditional approach, the technique automatically incorporates efficient adaptations by farmers to climate change. Furthermore, Ricardian analysis incorporates the substitution of different inputs and the introduction of alternative activities that each farmer has adopted in light of the existing climate (Kurkurlasuriya et al., 2004). The analysis of climate change impact on agriculture applying the Ricardian approach uses net crop revenue as a dependent variable, a more robust measure given concerns about equilibrium as it measures what the farmer currently receives without any concerns for future returns, discounting, capital or labour markets. The Ricardian method is superior to other methods as it accounts for changes in management practices and adaptation measures taken by the farmers according to the changing local climatic conditions to maximize their outputs and farm incomes, through pursuit of certain adaptation measures such as crop diversification, mixed cropping, irrigation, different planting dates, water and soil conservation practices (Mariara and Karanja, 2007; Mendelsohn 2014). Furthermore, the Ricardian method relies on cross-sectional data with the flexibility to consider all major enterprise activities. Lastly, the Ricardian method is relatively simple to implement compared to other methods. However, it also has some weaknesses. The first shortcoming is the likelihood of omitted variable bias, which exists in all cross-sectional analyses. It does not account for CO2 fertilization effects. Moreover, the method cannot capture the transition cost of a sudden adaptation against climate. However, in the real world a swift adaptation to new technology is barely practiced. The basics of the Ricardian model were built on the simple profit function (V) which is the difference between total revenue (PQ) and total cost (PX). The basic model can be stated as Equation 1 (Mendelsohn et al., 1994, 2014). V= ∑Pi Qi (F,X,G,M,R)- ∑PxX 4.1 Where V, is the net crop revenue; Pi is the market price of crop i, Qi is the output of crop i, F is a vector of climate variables, X is a vector of purchased inputs (other than labour and land), G is a set of socioeconomic variables such as household size, and average education, M is a set of farm characteristics such as farm size, access to credit, and access to extension services, R is a dummy variable representing adaptation measures and Px is a vector of input prices. Given the characteristics of the farm and market prices, it is assumed that the farmer will select input X that maximizes the net crop revenues (Mendelsohn, 2014). 17 Method and model specification 4.2 Empirical Model Specification A quadratic formulation of climate is the foundation of the standard Ricardian model (Mendelsohn and Dinar, 2009) and as a result, the net crop revenue can be stated as shown in Equation 2: V=βI+β2 F+β3 X+β4 G+β5 M+β6 R+u 4.2 Where V is the net crop revenue, F is a vector of climate variables, X is a vector of purchased inputs (other than labour and land), G is a vector of household socioeconomic variables, M is a vector of farm characteristics such as farm size, access to credit, and access to extension services, R is a dummy variable representing adaptation measures to capture farmers’ adaptation to climate change (for example, use of irrigation), β represent the coefficient of the explanatory variables and u is the error term. The Ricardian model regresses the natural log of net crop revenue per acre against long term climate variables and other control variables. Then, projecting in the future the estimated relationship between economic performance and long run climate variables allows us to account for climate impact (De Salvo et al., 2013; Mendelson et al., 1994). Most of the Ricardian model applications use seasonal climate variables (Mendelsohn and Dinar, 2009) since seasonal differences in temperature and precipitation have a significant impact on farmland productivity. This study used seasonal data, annual average temperature and average monthly precipitations for the cropping seasons and harvesting seasons, like studies by De Salvo et al., 2013; Fleischer et al., 2011. We could not use soil characteristics as a control variable because it was not easily available for each farm included in the sample. We, however, used the use of irrigation to cater for adaptation measures taken by the farmers. For temperature and precipitation, linear and quadratic terms are introduced to capture the level and non-linearity for climate effects. The temperature and precipitation hot season – defined as the average temperature and precipitation for March to May (long rain) and October to November (short rain) and temperature and precipitation cold season – defined as the average temperature and precipitation for June-August (cold season) because this is the standard specification employed in the pioneering Ricardo study to capture seasonal effects (Mendelsohn et al., 1994). The quadratic term is included in the model to capture the non-linear associations of climate variables and net crop revenue, which will indicate how the marginal effect will change as one moves away from the mean (Mendelsohn, 2014). If a positive coefficient for the quadratic term is obtained, the function assumes a U-shaped form (convex function), whereas if the value is negative, the function assumes a hill-shaped form (concave function). Based on the reviewed literature and previous cross-sectional analyses, it is expected that net crop revenue will have a hill-shaped relationship with temperature, implying that each crop has a known temperature at which it grows best throughout the seasons. Equation 2 can be re-written as Equation 3, which assumes a quadratic relationship between log of net crop revenue and climate variables to reflect the non-linear relationship that is consistent with other studies. 18 Effects of climate change on crop revenue in Kenya Vlog=β0+β1 Th+β2Th 2+β3 Tc+β4Tc 2+β5 Ph+β6 Ph 2+β7Pc+β8Pc2-β9 X+β10 G+β11 M++β12 R +μ 4.3 Where: Vlog represents logarithm of net crop revenue (acre) – measured as (average price in Ksh of 90kg bag of crop in 2013); T represents the mean temperature measured in degree Celsius; P represents the mean precipitation measured in millimetres; h represents the hot season (January to February, March to May and October to December for 2013); c represents the cold season (June to August for 2013); X represents a vector of purchased inputs (other than labour and land); G represents a vector of household socioeconomic characteristics; M represents a vector of farm and farmer characteristics; R represents dummy representing irrigation as an adaptation measure; βi are the coefficients of the variables; β0 is the constant term; and μ is the error term. To facilitate interpretation of the climate impact, we calculate the marginal effects of climate variables at sample means. From Equation 3, we can derive the marginal impact of a hot season temperature (Th) and hot season precipitation (Ph) on net crop revenue per acre evaluated at the mean as follows: MITh=dV/dTh = [β1+2β2 Th] 4.4 MIPh=dV/dPh = [β5+2β6 Ph] 4.5 The empirical process begins by estimating the response of crop revenues to climate variables only. The second model adds farm characteristics to the first model to add spatial heterogeneity in net crop revenue while the third model adds social economic variables and adaptation measures to the second model. Econometric analysis with cross-sectional data is usually associated with the problem of multicollinearity, heteroskedasticity, and the effect of outliers in the variables. The study implemented a quantile regression, which reduces the weight of the outliers and correcting for heteroskedasticity. Multicollinearity was reduced by dropping the variables that proved correlated with each other. This method provides a complete picture of the relationships between the outcome of y (Net Crop Revenue/Acre) and the regressors xi (climate variables and control variables). Moreover, quantile regression permits us to study the impact of such variables on both the location and scale parameters of the model, allowing a richer understanding of the data (Cameron and Trivedi, 2005). In our case we ran a regression, choosing the median as a statistical indicator of central tendency since a median regression is more robust to outliers than a mean regression. Finally, 19 it does not need any assumption about the parametric distribution of regression errors, since median regression is a semi-parametric approach. 4.3 Kenya’s Planting Calendar by Region and Crop Stage Tables 4.1a to 4.1c shows the planting calendar for various crops by region and crop stage. The details for specific crop and season by county and region are presented in Annex 2.8. Tables are important in showing how the four climatic variables (in Table 4.2) used for Ricardian analysis were constructed as indicated in Table 4.3. Table 4.1a: Kenya planting calendar by region and crop stage (maize) Region Crop type Planting period Reproductive period Ripening period Harvesting period Long rain season Western Maize February- March March-June July-August September Noth Rift Maize March-May May-July August- October November South Rift Maize February- May May-July July- October September- December Central Maize March-April April-June July-August September Eastern Maize March-April April-May June-July August Coast Maize March-April April-June June-August September North Eastern Maize March-April April-May June-July August Short rain season Western Maize August- September September- December January- February March Noth Rift Maize No short rain season South Rift Maize August- September September- December January- February March Central Maize October- November November- December January- February March Eastern Maize October- November November- December January- February March Coast Maize October- November November- December January- February March North Eastern Maize October- November November- December January- February March Method and model specification 20 Effects of climate change on crop revenue in Kenya Table 4.1b: Kenya planting calendar by region and crop stage (beans and other crops) Region Crop type Planting period Reproductive period Ripening period Harvesting period Long rains season Western Bean Green grams Cowpeas February- April March-May April-June July Noth Rift Beans April-May May-June June-July August South Rift Beans March- April April-May May-June July Central Beans March- April April-June June- August September Eastern Beans Green grams Cowpeas March- April April-May June-July August Coast Beans Green grams Cowpeas March- April April-May June-July August North Eastern Beans March- April April-May May-June July Short rains season Western Bean Green grams Cowpeas August- October September- November November- December January Noth Rift Beans September- October October- November November- December January South Rift Beans September- October October- November November- December January Central Beans October- November November- December January- February March Eastern Beans Green grams Cowpeas October- November November- December December- January February Coast Beans Green grams Cowpeas October- November November- December January- February March North Eastern Beans October- November November- December December January 21 Table 4.1c: Counties distribution per region Region Counties in each region Western Nyamira, Kisii, Migori, Homa Bay, Kisumu, Siaya, Kakamega, Bungoma, Busia and Vihiga Noth Rift Trans Nzoia, Uasin Gishu, Elgeyo Marakwet, Nandi, West Pokot, Baringo, Samburu and Turkana South Rift Kajiado, Narok, Bomet, Kericho and Nakuru Central Nyandarua, Nyeri, Laikipia, Kiambu, Murang'a and Kirinyaga Eastern Kitui, Machakos, Makueni, Embu, Meru, Tharaka Nithi, Isiolo and Marsabit Coast Taita Taveta, Kwale, Mombasa, Kilifi, Tana River and Lamu North Eastern Garissa, Wajir and Mandera 4.4 Measurement of Variables The variables used in the analysis of the impacts of climate change on net crop revenues are defined in Table 4.2. Table 4.2: Definition and measurement of variables Ricardian model variables Variable name Definition Measurement Expected sign V=V_log Net crop revenue per acre Logarithm of net crop revenue per acre Calculated as the total value of crop (Kgs) less input cost (using 2013 prices of a 90 Kg bag of maize in Ksh (5,000 - approx. 55.5/Kg)) + F Climate variables Th Temp_Hot season January- February temperatures Average degree Celsius ± Th2 Temp_Hot season sq January- February temperatures Sq Average degree Celsius ± Tc Temp_Cold season June-August temperatures Average degree Celsius ± Tc2 Temp_Cold season sq June-August temperatures Sq Average degree Celsius ± Method and model specification 22 Effects of climate change on crop revenue in Kenya Ph Preci_ Hot season March-May and October- December Precipitation Average in mm ± Ph2 Preci_Hot season sq March - May and October – December Precipitation Sq Average in mm ± Pc Preci_Cold season June - August Precipitation Average in mm ± Pc2 Preci_Cold season sq June - August precipitation Sq Average in mm ± M Farm and farmer characteristics M1 Access to Extension Service Farmers who accessed extension service Dummy-(Yes=1; No=0) + M2 Access to credit Farmers who accessed agricultural credit facilities Dummy-(Yes=1; No=0) + M3 Livestock_ ownership Livestock rearing of by H/H Dummy-(Yes=1; No=0) + M4 Priority crops Main crop grown (Maize) 1 = Maize; 2 = Beans; 3 = Other crops + M5 Total area under crop Land used for farming Acres ± G Household characteristics G1 Gender Gender of H/H head Dummy (M=1; F=0) + G2 Age Age of H/H head Count (in years) + G3 Education level Average years of education of the H/H head Count + G4 Occupation Main occupation of H/H head Categorical + 23 G5 Household size No. of persons in a given household Count/No ± G6 Poverty_ index Population living below poverty index Percentage (2005), county based ± R Adaptation measure R Access to irrigation Irrigation use at any point of the year Dummy (Yes=1; No=0) + Source: Authors’ compilation based on the literature review 4.4.1 Dependent variable The dependent variable for the analysis was the net crop revenue per acre. However, there were negative net crop revenue values derived from several observations. Besides, regression using these values produced large coefficients. To avoid this, we took the smallest value (negative +1) of the net crop revenue per acre and added across all the observations and then took the logarithm of the new net crop revenue value for all the observations. The logarithm of the net crop revenue per acre (V_log) was used as a proxy for the dependent variable (net crop revenue per acre). To arrive at the net crop revenue values, the gross crop revenue was calculated using the average of the 2013 value of maize per 90 kg bag (equivalent to Ksh 55.5 per Kg) and the cost of all inputs derived from the survey. The net crop revenue was then calculated by substracting the cost of all inputs (for example, fertilizer, seeds, and pesticides) used by the farmers in the survey. To obtain the net crop revenue per acre, we divided the new total crop revenue values by the farm size (total area under crop cultivation) for all observations. 4.4.2 Explanatory variables The relationship between climate change and net crop revenue is complex and is confounded by other factors including social economic and farm characteristics as described below: (a) Climate Variables The climate variables used are temperature and precipitation. The two variables were split into hot and cold seasons as described in Table 4.3. Method and model specification 24 Effects of climate change on crop revenue in Kenya Table 4.3: Description of climate variables Climate variable Calendar period Crop stage and weather condition Temperature – Hot season January-February Ripening and harvesting for the short rains Temperature – Cold season June-August Vegetative and ripening for the long rains Planting for the short rains Precipitation – Hot season March-May and October- December MAM – Planting for the long rains and harvesting for the short rains OND – Planting for the short rains Precipitation – Cold season June-August Vegetative and ripening for the long rains Planting for the short rains (b) Farm and farmer characteristics • Access to extension services: Agricultural extension services play a crucial role in enhancing farm productivity and income. They are the main conduit for disseminating information on farm technologies, supporting rural adult learning, and assisting farmers in developing their technical and managerial skills. Studies show that access to extension services can have a positive impact on crop revenue through increased productivity, improved farm management, and reduced use of inputs. • Access to credit: Access to agricultural credit can have a significant positive impact on crop revenue by enabling the purchase of better inputs and technologies, improving farm management, and increasing income. Studies have shown that access to agricultural credit increases the yields. • Livestock ownership: Livestock ownership can provide an alternative source of income and serve as a buffer against crop failures due to climate variability. It can also enhance crop revenue by using manure as fertilizer and integrating crop-livestock systems (Waseem et al., 2023; Mzyece, and Ng’ombe, 2020) in the rearing of livestock. • Priority crops: The crops were grouped into three categories; maize, beans and other crops based on the first two choices of priority crops reported by farmers in the survey. • Total cropped area: The total area under cultivation can influence crop revenue, as larger areas may lead to higher production. However, this relationship is also affected by the quality of land, availability of irrigation, and other factors (Matthew, 2022; OECD, 2024; Emran et al., 2021). (c) Household characteristics • Gender of household head: The gender of the household head can influence agricultural practices and access to resources. Studies have shown that female-headed households may have less access to land, credit, and extension services, which can affect their ability to adapt to climate change and maintain crop revenue. Gender implications of agricultural commercialization can lead 25 to changes in women’s decision-making agency and control over resources, impacting crop revenue (Berhane et al., 2022). • Age of household head: The age of the household head can have a significant impact on crop revenue. However, the relationship is not always straightforward and can depend on various factors as some studies have shown. Some studies have found a negative association between the age of the household head and crop revenue. As the age of the household head increases, the likelihood of crop diversification decreases, which could potentially lead to lower crop revenue. This could be due to older farmers being less likely to adopt to new technologies or practices. • Average education of the household head: Education plays a significant role in the ability to understand and implement adaptation strategies for climate change. Educated farmers are more likely to adopt innovative farming techniques and manage resources efficiently, potentially leading to increased crop revenue (OECD, 2019). • Occupation of household head: The occupation of the household head can significantly impact crop revenue, but the direction and magnitude of this impact can depend on various factors, including the nature of the occupation and the specific circumstances of the household. If for example the household head’s primary occupation is farming, they may have more time and resources to devote to farming activities, which could potentially lead to higher crop revenue. If the household head has a non-farming occupation, they might have additional income that can be invested in farming activities, such as purchasing better quality seeds or hiring labour, which could increase crop revenue. However, they might also have less time to devote to farming, which could potentially decrease crop revenue. • Household size: Larger households may have more labour available for farming, but they also have more mouths to feed, which can affect the allocation of resources and crop revenue. The social demographics of households, including size, can influence decisions to adopt sustainable agricultural practices (SAPs), which are crucial for adapting to climate change (Berhane et al., 2022). • Poverty index: Poverty can limit a household’s capacity to invest in climate- resilient agricultural practices, affecting crop yields and revenue. Climate- induced crop yield changes can have significant poverty implications, with impoverished households experiencing increased vulnerability (Thomas et al., 2010). (d) Adaptation measure Use of irrigation can significantly influence crop revenue in various ways. First, by providing a consistent water supply, irrigation helps crops grow more efficiently, leading to higher yields. This is especially important in regions with irregular rainfall (Olamide et al., 2023). Secondly, adequate irrigation ensures that crops receive the right amount of water at the right times, which can improve the quality of the produce thereby translating to higher market prices. Thirdly, with reliable Method and model specification 26 Effects of climate change on crop revenue in Kenya irrigation, farmers can grow a wider variety of crops, including those that are more water-intensive but potentially more profitable. Fourthly, irrigation reduces the risk of crop failure due to drought, providing a more stable income for farmers (Ibid). Overall, the use of irrigation can lead to higher productivity, better crop quality, and increased revenue for farmers. 4.5 Data and Data Sources The primary data were collected through a national survey conducted in 2013 by the Agriculture Sector Development Support Programme (ASDSP) covering a sample of 12,651 farm households randomly selected from all the 47 counties of Kenya. The survey considered the cropping seasons of 2013. The questionnaire was developed to collect detailed information on the socioeconomic characteristics of the sampled households including basic household information (such as gender, farmer's age, education, and household size). The farm characteristics include farm size, access to farm credit, distance to market, access to extension services, and access to irrigation. Furthermore, the questionnaire also collected information relating to crop types, growing seasons of different crops, costs and amounts of inputs used such as seeds, pesticide, and fertilizer. This study settled on the analysis of maize production to cover the first objective on examining the relationship between net crop revenue and climate change and other economic variables and the second objective on forecasting the effects of climate change variation on net crop revenue. The production values of maize used are as reported by the farmers. To calculate the gross and net revenues, a 90 kg bag of maize was converted into Ksh at the 2013 price of Ksh 5,000. To get the net revenue per acre, the net revenue was divided by the total area under cultivation per each farmer. The climate data is based on average for maximum and minimum temperature and precipitation for the growing seasons; January-February, March-April-May (MAM) and October-November-December (OND) for the hot season and the period June-July-August (JJA) for the cold season for temperature and precipitation. The data was obtained from the Kenya Meteorological Services (KMS). The poverty index used was for 2005, given that the survey data was for 2013 and was obtained from Kenya National Bureau of Statistics (KNBS) survey of 2005. For the third objective on climate financial flows and composition to the agriculture sector in Kenya, we used FAO MAFAP 2013/14 to 2017/18 database. Flow for climate change financing in agriculture covers mitigation and adaptation activities. A quantile regression was implemented to examine the impact of a set of climate variables and other control variables on the outcome variable, the net crop revenue per acre. This method was chosen because it provides a complete picture of the relationship between the outcome (net crop revenue/acre) and the regressors (climate variables and control variables). Furthermore, quantile regression allowed us to study the impact of these variables on both the location and scale parameters of the model, enabling a deeper understanding of the data (Cameron and Trivedi, 2005). In the case of this study, we ran the regressions choosing the median as a statistical indicator of central tendency since a median 27 regression is more robust to outliers than a mean regression. It also does not need any assumption about a parametric distribution of regression errors, since median regression is a semi-parametric approach. STATA software was used to implement the analysis in this study. 4.6 Descriptive Statistics 4.6.1 Descriptive statistics for the model variables Table 4.4 presents the descriptive statistics (mean, standard deviation, minimum, and maximum) for the dummy and continuous socioeconomic and farm variables used in the study. Thus, the mean value of net crop revenue per acre of land is Ksh 41,344.90. The mean age of household heads is approximately 50 years, while the maximum age is 128 years. The average family size is approximately six (6) members in each household. The average total cropped area is approximately 2.56 acres, while the maximum is 58 acres. The average net crop revenue per farmer is Ksh 52,135.19, while the maximum is Ksh 7.48 million. The average quantity of maize harvested by farmers is 1,271.09 kgs. It is important to note that there is wide variation in net revenue, cropped area, age of household heads, size of the household, and amount harvested. Table 4.4: Descriptive statistics for model variables Variable Obs Mean Std. Dev. Min Max Net crop revenue per acre 9,663 41,344.90 390,139.40 -97971.15 22800000.00 Temperature – hot season 12,651 29.46 4.49 17.90 42.00 Temperature – Cold season 12,651 14.98 5.41 5.50 30.30 Precipitation – Hot season 12,651 185.54 52.90 66.55 264.87 Precipitations Cold season 12,651 70.79 55.70 1.48 172.77 Access to extension services 12,651 0.19 0.39 0 1 Access to agricultural credit 12,651 0.04 0.19 0 1 Livestock ownership 12,651 0.83 0.37 0 1 Priority crops 9,826 1.56 0.89 1 3 Total area under crop 9,762 2.56 3.95 0.01 58.00 Method and model specification 28 Effects of climate change on crop revenue in Kenya Gender of household head 12,651 0.83 0.37 0 1 Age of household head 12,470 50.01 14.95 19 128 Education of household head 12,477 7.44 5.09 0 18 Main o household ccupation of head 12,414 2.61 1.53 1 7 Household size 12,651 5.63 2.62 1 20 Poverty index 12,651 46.13 17.59 12.10 92.90 Use of irrigation 12,651 0.05 0.21 0 1 County No. 12,651 26.61 13.05 1 47 Source: Authors' calculations based on ASDSP survey data and KNBS Temperature and rainfall climate variables were calculated based on the dry season (January-February) and the main growing seasons (average of the March- April-May (MAM) for hot season and June-July-August (JJA) for cold season using 2013 climate data from Kenya Meteorological Services). Table 4.4 presents the mean, standard deviation, minimum and maximum values of all the variables used in the analysis. As can be seen from the table, for climate variables, the maximum average temperature recorded for the hot and cold seasons was 420C and 300C respectively, whereas the minimum average temperature recorded for the same was 17.90C and 5.50C, respectively. The standard deviation values of temperature for all seasons are approximately the same, indicating that the seasonal temperature is evenly distributed. Regarding the rainfall, the minimum average rainfall recorded for the hot and cold seasons were 66.55 mm and 1.48 mm, respectively whereas the maximum average rainfall for the hot and cold seasons were 264.87 mm and 172.77 mm, respectively. As indicated by the standard deviation column, there was no variation in the rainfall distribution across the hot and cold seasons each year. The total cropped area varied from a low of 0.01 acres to a high of 58 acres. The standard deviation values of the land under crops were large, indicating that there were variations in the land under crops. The mean age of household heads was approximately 50 years, with the youngest being 19 years and the eldest being 128 years. The standard deviation of household heads was large, indicating that there were variations in the age of household heads. There were also variations in the level of poverty across the households as the standard deviation indicates. 29 5. Results and Discussion 5.1 Results of Estimation of Response of Crop Revenue to Climate The estimation of the Ricardian function specified in Equation 3 follows six (6) model variations, with results presented in Table 5.1. The six (6) models were estimated to control for various explanatory variables. Models 1 and 4 focus on the climate factors (precipitation and temperature), with Model 1 controlling for use of irrigation (adaptation measure), farm and farmer characteristics as well as socioeconomic variables. In Models 2 and 5, the use of irrigation is introduced with model 2 controlling for farm and farmer characteristics and socioeconomic variables. Lastly in Models 3 and 6, other socioeconomic and farm and farmer characteristics, household size, poverty index, total cropped area, and livestock dummy are added, with Model 3 excluding squared terms for temperature and precipitation. Such an exercise is common in the literature of the Ricardian method (see Seo and Mendelsohn, 2007). The four main climate variables show significant effects on crop revenues in all the six models, as does the square of the temperature and precipitation except for the squared term of the precipitation for the cold season in Models 4 and 5 and the squared term of the temperature for hot season in Models 4, 5 and 6. A test of joint significance suggests that the combined effect of the four main climate factors is significant (F = 81.93; p < 0.00). This indicates that the estimates are robust across the six models. Table 5.1: Response of crop revenue to climate variables, adaptation and farmer and socioeconomic variables Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Log net crop revenue/ acre Coef. Coef. Coef. Coef. Coef. Coef. Temperature – hot season -0.021*** (-4.350) -0.000*** (-4.750) -0.032*** (-7.280) 0.002*** (3.050) 0.003*** (2.090) 0.047* (1.910) Temperature – Cold season 0.052*** (13.610) 0.049*** (12.710) 0.023*** (6.070) 0.105*** (5.180) 0.098*** (4.880) 0.085*** (4.510) Precipitation – Hot season -0.004*** (-7.830) -0.004*** (-7.830) -0.007*** (-12.770) 0.012*** (3.540) 0.009*** (2.830) 0.008** (-2.560) Precipitation – Cold season 0.007*** (12.600) 0.007*** (13.030) 0.006*** (11.460) 0.005** (2.500) 0.005*** (2.810) 0.012*** (6.510) Temparature – Hot season Sq 0.000 (-0.590) 0.000 (-0.660) 0.000 (0.380) Temparature – Cold season Sq 0.002*** (2.720) 0.001** (2.440) 0.002*** (3.220) 30 Effects of climate change on crop revenue in Kenya Precipitation – Hot season Sq 0.000*** (-5.090) 0.000*** (-4.340) 0.000 (-0.190) Precipitation – Cold season Sq 0.000 (1.300) 0.000 (1.090) 0.000*** (-3.290) Access to extension services 0.147*** (3.180) 0.129*** (2.740) Access to agricultural credit 0.145 (1.440) 0.153 (1.490) Keep livestock 0.006 (0.130) 0.002 (0.040) Priority crops 0.247*** (11.210) 0.252*** (11.090) Total area cropped -0.113*** (-24.170) -0.113*** (-23.480) Gender of household head 0.183*** (3.670) 0.187*** (3.670) Age of household head -0.003** (-1.960) -0.003* (-1.920) Average education level of household head 0.016*** (3.830) 0.016*** (3.780) Occupation of household head -0.046*** (-3.630) -0.048*** (-3.700) Household size -0.009 (-1.220) -0.010 (-1.360) Poverty index -0.011*** (-7.330) -0.013*** (-7.840) Use of irrigation 0.842*** (8.990) 0.516*** (6.120) 0.777*** (8.520) 0.486*** (5.610) Cons 11.155*** (62.380) 11.121*** (61.490) 12.035*** (54.130) 9.989*** (15.600) 10.057*** (15.840) 12.977*** (18.980) Observations 7,858 7,858 7,452 7,858 7,858 7,452 Pseudo R2 0.033 0.041 0.107 0.038 0.044 0.111 * Significant at ≤ 10% ** Significant ≤ 5% *** Significant ≤ 1%. (t-values in parentheses) 31 Results and discussion The coefficients of the linear term of the hot season precipitation in Models 1, 2, 3 and 6 are negative, implying negative effects of increases in precipitation during this season. In Kenya, this can be attributed to the fact that this is the period (MAM) of long rains and any further increase in precipitation during this period could lead to flooding and destruction of crops. The cold season precipitation in all the models has a positive relationship with the crop revenues. This period (June- August) is crucial and is required for ripening and flowering of crops. The variables for the hot season temperatures are negatively related to crop revenue, while cold season temperature variables in the models are positively associated with crop revenue. High hot season temperatures are harmful to crop production and, therefore, reduce crop revenue while high cold season temperatures are beneficial and positively related to crop revenue. This is because the hot season temperature (March-May) is the planting period followed by formative crop growth in May and June, while cold season (June-August) is the period for ripening and maturity of crops. High hot season temperatures, therefore, slow down or destroy crop growth, while higher cold season temperatures in June-August are crucial for ripening and harvesting. For the non-linear terms for temperature variable in the hot season in Models 4, 5 and 6 are positive and significant. Furthermore, the squared terms for temperature in cold season in Models 4, 5 and 6 are significant and positive. The results implies that increases in hot season temperatures tend to benefit crop revenue with diminishing marginal benefits up to a maximum turning point after which further increases in temperature lead to negative effects on the crop revenue (Mendelsohn et al., 1994; Seo and Mendelsohn, 2007). Similarly, the non-linear terms for precipitation variable in hot season in Models 4, 5 and 6 are positive and significant, indicating that increases in hot season precipitation tend to benefit crop revenue with diminishing marginal benefits up to a maximum turning point after which further increases in precipitation will have negative effects on crop revenue. Adding access to irrigation variable in the 2nd and 5th models helped to capture adaptation by the farmers in the face of climate change. The coefficient of irrigation use had a positive impact on crop revenue, implying the importance of adaptations to counter the effects of climate change through irrigation. This conforms to a prior expectation that the use of irrigation helps farmers to cushion themselves by increasing crop production, which increases their crop revenues. Finally, the impact of farm and farmers characteristics and socioeconomic variables were tested. The introduction of these variables in Models 3 and 6 raised the pseudo R2 from 0.041 to 0.107 for Model 3 and 0.038 to 0.111 for Model 6. Most of the household level variables have a significant impact on crop revenue. The results for Model 3 suggest that among the control variables, the average education level and gender of the household head, access to extension services, access to agricultural credit and priority crops planted are positively associated with crop revenue, while age and occupation of the household head, household size, poverty index and total cropped area are negatively associated with crop revenue. The livestock ownership dummy has a positive impact on crop revenue though not significant, which implies complementarity between crop farming and 32 Effects of climate change on crop revenue in Kenya livestock keeping. Furthermore, access to agricultural credit, livestock keeping, and farm size variables are however not significant in determining crop revenue in both models. One interesting result, however, stands out. A significant and negative effect of total cropped area is unexpected. This is because an increase in the area under crop would boost yields and, therefore, crop revenue. However, this is not surprising because several factors can cause an increase in total cropped land to be negatively associated with crop revenue. These may include land degradation, where expanded cropped land often involves using marginal or less fertile land, which can lead to lower yields and, therefore, reduced revenue; resource dilution, where when more land is cultivated, the resources (such as water, fertilizers, and labour) available per unit of land may decrease, leading to less efficient production and lower overall revenue; increased costs related to managing larger areas of land, which can increase operational costs for labour, machinery, and inputs; or climate risks, where larger areas of cropped land may be more exposed to climate risks such as droughts, floods, and pests and diseases. This can substantially reduce yields and revenue (Kaiser, 2021). 5.2 Marginal Effects In the context of climate and agriculture, the Ricardian model estimates the effects of temperature and precipitation on crop yields per acre by calculating the marginal effects of the climate variables (detailed results are presented in Annex 2.7). This is done by differentiating the crop yield per acre function with respect to each variable (temperature and precipitation). This allows for determination of how a small change in each variable affects crop yield per acre. The marginal effect depends on the region’s specific conditions and annual average temperature. To facilitate interpretation of the climate impact on crop yield per acre, the marginal effects of these variables at sample means were calculated (Table 5.2). The marginal effects for cold season temperatures are positive, but hot season temperatures have larger negative effects on crop yields. The seasonal marginal effects with respect to hot season temperature are statistically significant and thus different from zero confirming that global warming is likely to have devastating effects on agricultural production unless farmers take adaptation measures to counter the impact of climate change (Kurukulasuriya et al., 2004). Table 5.2 Marginal effects of seasonal temperature and precipitation on crop yield per acre Seasons Marginal effect Std. Err. P>z Temperature hot season (January- February) -29.19 0.004 0.000 Temperature cold season (June-August) 15.21 0.004 0.000 Precipitation hot season (March-May and October-December) -19.31 0.001 0.000 Precipitation cold season (June-August) 68.58 0.001 0.000 33 Results on Table 5.2 indicate that the marginal effect of increases in precipitation and temperature in the two seasons are significant. The higher hot season temperatures have negative effects on the crop revenue implying that a further increase in temperatures would be harmful to agricultural production whereby a further increase in hot season temperature by 10C would lead to a reduction in crop yield by 29.19 kgs per acre. Increases of the cold season temperatures are beneficial to crop yields whereby a further increase in cold season temperature by 10C will increase the crop yield by 15.21 kgs per acre. Furthermore, increase in precipitation has positive effects on crop yields in cold season when Kenya experiences short rains, whereby an increase in cold season precipitation by 10C would lead to an increase in crop yield by 68.58 kgs per acre. The climate conditions in the hot season temperature and precipitation are the main cropping seasons and any change in climate variables affects crop yields and, therefore, crop revenue. 5.3 Forecasted Effects of Climate Change on Crop Revenues To estimate the changes in crop yields due to changes in climate, two Global Circulation Models (GCM) results were used: The Canadian Climate Model (CCM) and the Geophysical Fluid Dynamics Laboratory Model (GFDL), which predicts 3.50C and 40C changes in temperature by the year 2030, respectively, for Kenya. Both models also predict a 20 per cent change in precipitation over the same period (Kabubo-Mariara and Karanja, 2006). By analyzing variations in temperature and precipitation, it is possible to understand how climatic variables impact crop yields. The study examined the impact of adverse changes in temperature and rainfall on crop yield. Using the estimated marginal effects in Table 5.2, the study calculated the changes in crop yield for each climate scenario throughout Kenya. The change in crop yield per acre was then multiplied by the number of acres of cropland per farmer to get an aggregate national effects of climate change on crop yield. This value was then summed across all the farmers to get the total effects at the national level. The change in crop yield per acre per farmer was derived using the formula: Aggregate climate effect Fi= ∆Yi*Si 5.1 Where ∆Yi = Change in crop yield per acre Si = Total cropped area per farmer i Fi= Farmer i GFDL and CCM climate forecast of temperatures and precipitation assumes the changes in climate are uniform across the country. These models show that a 3.50C increase in temperature will result in a crop yield reduction of 1.13, 90 kgs bags per acre per farmer and 28,318.57, 90 kgs bags for all farms in Kenya, equivalent to Ksh 141.59 million losses for all farms in Kenya (using the 2013 price of Ksh 5,000 for a 90 kgs bag). Furthermore, a 40C increase in temperatures would result in a reduction of crop yields of 32,364.08, 90 kgs bags for all farms in Kenya while a 20 Results and discussion 34 Effects of climate change on crop revenue in Kenya per cent reduction in rainfall would lead to a crop yield reduction of 109,448.30, 90 kg bags for all farms in Kenya, equivalent to Ksh 547.24 million. The combined effects of a change in adverse climate variables (temperature and rainfall) can be detrimental to crop production (Mendelsohn, 2014; Shrivastava et al., 2011). An increase in temperature of +3.50C, combined with a 20 per cent reduction in rainfall will cause crop yield reduction of 109,550.28, 90 kgs bags of maize valued at Ksh 1,666.85 million nationally. Similarly, an increase in temperature of 40C, combined with a 20 per cent reduction in rainfall will cause a crop yield reduction of 340,465.48, 90 kgs bags of maize valued at Ksh 547.75 million. Table 5.3: Forecasted effects of climate change on crop yields and revenue Climate scenario for CCC and GFDL Models Impact (No. of 90 kgs bags) Price of a 90 kgs bag in 2013 (Ksh 5,000) Average loss (No. of 90 kgs bags) per farmer Total loss (No. of 90 kgs bags) - National Total loss (Ksh million) - National +3.50C 1.13 28,318.57 141.59 +40C 1.29 32,364.08 161.82 20% rain reduction 4.38 109,448.30 547.24 +3.50C plus 20% reduction 1.13 109,550.28 547.75 +40C plus 20 % reduction 1.29 109,564.85 547.82 Source: Authors' calculation using Ricardian median regression Thus, a further reduction in precipitation and an increase in temperature in the country will make farming unproductive. As such, effective adaptation strategies need to be considered early enough by Kenyan farmers. These findings are consistent with the studies by Mulwa et al. (2022) and Owa and Ouda (2007). 35 6. Conclusion and Policy Implications 6.1 Conclusion Climate change is widely acknowledged as a global concern due to its significant effects on human life, given its multiple impacts on livelihoods. Therefore, it is a matter of great concern for economists and policymakers. This study aimed to measure the effects of climate change on net crop revenue in Kenya by employing a Ricardian method, using cross-sectional data from 12,651 randomly selected farm households across all 47 counties of Kenya. The findings show that climate variables have a significant impact on net crop revenue per acre across Kenyan farm households in all the models, indicating both positive and negative effects of precipitation and temperature. Net crop revenue is likely to decrease during cold season temperatures and hot season precipitation, the major harvesting seasons in Kenya. The results are consistent with Ricardian studies in Europe (Van Passel, Massetti and Mendelsohn, 2016; Fabian et al., 2018) and crop studies (Olbrisch et al., 2011). The results highlight the importance of seasonal climate changes when measuring impacts and considering climate adaptation policies. Notably, the seasonal variation will be significant if climate change models predict different effects by season. Climate impacts are also likely to vary considerably across Kenya’s regions because the climate is heterogeneous. The study results provide evidence of the economic impacts of climate change on crop revenue in Kenya and suggest that Kenyan farms are indeed sensitive to climate variables such as temperature and precipitation. Furthermore, the results enhance the understanding of the nature of climate change effects in Kenya. This is crucial for predicting the possible effects of future climate change on agricultural production, which is fundamental to the country's economy. Such knowledge is key to designing adaptation policies, which are vital for climate change preparedness. The results also reveal that since Kenya’s agriculture is vulnerable to climate change, if severe scenarios of climate change, as predicted, materialize, farmers in Kenya could greatly lose their crop yields and, therefore, their crop revenue. This would be a stunning blow to the agricultural sector in Kenya, which is the main contributor to GDP. The impact would be even more devastating to Kenya's most vulnerable ASAL regions. In terms of climate financing for the agriculture sector, the examined expenditure pattern aligns broadly with national adaptation and mitigation objectives. Adaptation-related measures enhance the resilience of the agricultural sector by promoting the use of improved inputs and better technologies, providing training and extension services, conducting agricultural research, and supporting food consumption. The allocated public resources are well-matched with the objectives of the national climate change adaptation plan. However, there is a disparity between the allocated funds and the actual disbursement of climate finance for mitigation and adaptation in the agriculture sector. Furthermore, it was also not possible to establish the role of the fund's measures in climate change adaptation 36 Effects of climate change on crop revenue in Kenya or mitigation; therefore, the data was marked as “not determined” and was not used for classification. There are caveats that readers should keep in mind when interpreting the results of this study. First, the cross-sectional analysis is vulnerable to