Proposed Objectives of the Project:

1. Understand the feasibility of “genetically modified” (GM) labeling in agri-food value chains in developing countries.
2. Assess developing country consumers’ knowledge and familiarity with GM crops.
3. Measure consumers’ willingness to pay (WTP) for GM and non-GM varieties and whether this WTP is affected by GM information labels.

Summary of Results:
Because of the pandemic, we were not able to go to Bangladesh at the time. The original student (Carly Trachtman) who started the project, got paid for a year, and graduated after doing research because of the pandemic and then left. Instead, Hamza, Ahsan, and I analyzed the data that we obtained from Bangladesh. Ahsan got extra data, and we wrote a paper that

1. The yield, pesticide use, and cost effect of Bt Eggplant
2. This is the first study that shows Bt Eggplant get higher prices than non-Bt Eggplant (we collected data that we will use later. This data explains the differences in prices better).
3. We identified the factors that explain adoption

More importantly, we compared our results with two existing studies based on control trials. We show that yield and pesticide use are similar. They lack some of the other cost considerations and don’t have adoption results. Moreover, we show that actual field data can do what controlled experiments can do and provide much more information at lower research cost. The bottom line is that the highly regarded control trial is not the gold standard for empirical research in economics. Of course, they have an important role, but other studies using traditional econometrics and observed field data may be more appropriate in other situations.

The results will appear in European Review of Agricultural Economics in 2024. The paper was accepted.

Specific Objectives of the Project:
1. To conduct a study and provide data analysis on effects of severity of the pandemic measured in terms of county level health outcomes (deaths, deaths, hospitalizations) by day on the mobility and social distancing behavior across counties in California, and how mobility and social distancing changes when county level mitigation strategies and stay at home mandates are implemented using an econometric model and unique policy county level data merged with data on daily county level mobility measures (via a data sharing agreement between U C Berkeley and Unacast).
2. To develop an econometric model to measure the effects of the severity of the pandemic, the mobility changes, and the mandate effects on business revenues and rate of closures by county over time, investigating such patterns for different types of business.
3. Distinguish such patterns of effects by the county rural classification, and also as a function of availability of hospitals using a unique data set of hospital locations by county and characteristics of hospitals in terms of number of beds, trauma centers, and number of recent closures of hospitals, especially in rural counties.

Summary of Research to Date:
Results are preliminary. They rely on mobility data collected from Unacast by county by day, and the county level and city level extensive mandate and policy implementation data we collected. We also used womply business data by week, for less representative samples than county level data (we only realized this after we had access to the data). In addition, we define counties according to population, rural definition, and merge hospital census data with hospital characteristics and capacity into our analysis.

Objective 1: On average, mandates have a significant effect on mobility and encounters with other individuals. The measures we use are average distance traveled, distance to non-essential businesses, and the number of encounters (based on the number of cell phones in a radius to a certain cell phone by day) in the data. All three distance measures drop significantly even before the first policies are implemented, in all counties (first implementers and also for the later implementing counties) The incremental effect of mandates is a proportion of the initial drop.

Objective 2: There is a significant exit of businesses in the data during covid, resulting in tremendous sample selection that does not allow us to look at the effect on business revenues, as the businesses that survive in the sample during covid are the ones that are different from the ones that exit, so we do not have comparable samples throughout as the policies are implemented. Second, we find that almost all business sectors suffer the exit of small businesses. The effect is long-lasting and we do not have data long enough to investigate recovery. We plan to do so in a future project.
Objective 3: we did not have enough coverage of small business data in rural areas to investigate the differential effect of the pandemic and policies to investigate exit intensity. We do not find differential effects in mobility and human encounter responses to the mandates between rural and nonrural counties. We did not have enough power
(sample size) to distinguish the effect on mobility due to hospital capacity in rural versus non-rural counties.

Specific Objectives of the Project:

1. To analyze the extent to which the COVID-19 health crisis affected California farmers’ access to farm workers in 2020.
2. To analyze how COVID-19 affected farming decisions in California in 2020.
3. To analyze how the COVID-19 pandemic is likely to influence the use of labor-saving technologies in the future.

Summary of Results:
We surveyed a broad sample of fruit, vegetable, horticultural, and nursery farmers covering California and the entire United States, in conjunction with the California Farm Bureau Federation, the National Council of Agricultural Employers, and American Hort. Surveyed farmers reported significant labor shortages. For example, in a national survey, nearly two-thirds of the farmers reported having difficulty hiring all the workers they wanted to produce their main commodity in their highest revenue producing state during 2021. The average farmer who would have normally hired 100 workers but faced a labor shortage would have only been able to hire 82 during 2021.

Farmers reported incurring additional costs related to COVID-19, including purchasing additional personal protective equipment, extra cleaning and sanitation activities, and adding sanitation facilities for workers. Some farmers reported spending thousands of dollars on each employee. COVID-19 played a role in the labor shortage issues that California farmers reported, with over half of the farmers who experienced a shortage indicating that COVID exacerbated it. Farmers reported a number of factors related to COVID-19 that contributed to labor shortages, including direct exposure to the virus, shelter-in-place orders, and generous unemployment benefits that allowed workers to obtain more income by not working. Most farmers who faced labor shortages indicated that the labor shortages were worse in 2020 than 2019.

H-2A workers are increasingly common, consistent with a USDA Economic Research Service study (co-authored by Professor Rutledge) revealing that H-2A visa use is rapidly expanding (Castillo, Martin, and Rutledge, 2022). There is no sign of this trend slowing down. H-2A workers continued to constitute a small share of labor use among California survey respondents. However, the H-2A program in California is expanding rapidly (Martin and Rutledge, 2022).

A significant share of farmers reported using a new labor-saving technology to help them mitigate problems stemming from labor scarcity. More than a third nation-wide made changes to their product mix to reduce labor costs. One in three California farmers reported implementing new labor-saving technologies in 2020, and among these, the main labor-saving technology was used on an average of 66% of their main crop production in their main county. The most oft-cited reasons for using labor-saving technologies and mechanical harvest aids, respectively, were rising labor costs and the ability to harvest faster.

Overall, this project found that many farmers face significant issues stemming from a lack of labor, while many are struggling to navigate the situation. Farmers are clearly making efforts to mitigate production and profit losses resulting from labor shortages; however, this issue continues to be a major challenge for fruit, vegetable and horticultural and nursery farmers in California and throughout the United States.

As background for these studies, we conducted detailed reviews of the changing agricultural workforce in the US and abroad, why it is happening, and its implications (Charlton, Rutledge and Taylor 2021); the agricultural labor supply response (Hill, Ornelas and Taylor 2021); and the future of work in agri-food (Christiaensen, Rutledge and Taylor 2021).

Specific Objectives of the Project
Production of a dissertation investigating the effects of the Federal Milk Marketing Orders on international trade in dairy products.

Summary of Results
This project is investigating the effects of the U.S. Federal Milk Marketing Orders (FMMOs) on trade in dairy products. The FMMOs implement price discrimination in the beverage milk market and redistribute gains to milk producers through blend pricing. Higher returns for dairy farmers increase the quantity of milk supplied to the market, while high prices in the beverage milk market reduce quantity demanded, leading to increased quantity in the dairy product manufacturing sector. Over the past two decades the U.S. has become a net exporter of dairy products, and this project investigates the role played by the marketing orders in that transition.

The research funded by this grant focused on the movement of farm milk and trade in dairy products within the United States. The marketing orders set regional prices for beverage milk, influencing the price received by milk producers in that region. This paper uses recent developments in the trade literature to model the process by which milk is delivered to different regions across the U.S and how dairy products are distributed to buyers. The dairy product market will include domestic demand from across the U.S. as well as demand facing the U.S. from the international market. We use simulations to compare a baseline scenario with the FMMOs in play to counterfactual scenarios that remove the regional price discrimination and revenue pooling.

The model of interstate trade will be supported by a second paper focused on empirical estimation of dairy product demand. Whether U.S. dairy policies have an impact on the international dairy product market depends on whether the U.S. faces a downward sloping world demand curve for tradable dairy products. This paper will use an econometric model of U.S. net exports and exogenous shifts in marketing order policies to estimates the world demand curves for tradable dairy products.

This research will be extended to an international scope to investigate the impacts of FMMO policies on international trade more broadly. While the model of interstate trade allows dairy product manufacturers to export products to the international market, the impacts of the FMMOs on that market are not explicitly modeled. The international trade model with consider U.S. imports and exports in the context of the global market for dairy products. This paper will focus on the U.S. shift from net importer to net exporter of dairy products and the extent to which changes in FMMO policies caused that shift.

Alongside the main goals of this project, it was a natural extension to study the impacts of the COVID-19 pandemic on the dairy industry. In the early stages of the pandemic school and restaurant closures impacted demand for dairy products, and milk producers reacted to falling prices by reducing production. Milk and dairy product prices were volatile throughout 2020. We initially focused on explaining these impacts in the western states for a publication in the Western Economic Forum. We outlined the government response to the pandemic and payments received by the dairy industry in 2020. This study was expanded to use forecasting methods to evaluate the magnitude of the impacts relative to historical data. There are outstanding questions to investigate, such as the impact of government purchasing programs on the market for dairy products, so we are developing a follow-up project to study these questions.

Specific Objectives:
Chilean fruit production grew by 12% per year from 1997-2007, and the country is currently one of the world's leading fruit exporters. This project evaluates the role that Chilean agricultural land redistribution in the 1960's and 1970's played in the subsequent adoption of fruit production.

Summary of Results:
In the 1960's and early 1970's, progressive land redistribution efforts in Chile reassigned ownership of roughly 13 percent of the nation's agricultural land. Municipality-level analysis indicates that redistribution was associated with greater area devoted to vineyards and less to forestry by 1997, suggesting that land reform increased the intensity of fruit cultivation. These results were derived using digitized records of plots redistributed during land reform. Comparable data on
non-reform plots exists in undigitized format, and digitizing these records provides a path forward for future research that evaluates the effect of land redistribution at a more granular level.

Specific Objectives of the Project:
Re-examine the economic logic that the US Environmental Protection Agency used to justify the Navigable Waters Protection Rule; use machine learning and artificial intelligence algorithms to determine the scope of these water quality regulations.

Project Report/Summary of Results:
The first component resulted in a published publicly available report that received media coverage. It also resulted in a peer-reviewed journal article in Science.

The second component trailed a number of graduate students, and resulted in a preliminary set of results and series of meetings with senior staff at the US Environmental Protection Agency to convey the approach and initial results. This is an ongoing project that will involve research and policymaker outreach over multiple years as the EPA revises the Clean Water Act.

Proposed Objectives of the Project:
The objective of this project is to improve our understanding of the demand for plant-based alternatives and their competing relationship with conventional animal products including (i) growth over time in this segment, (ii) substitution patterns between animal products and their plant-based analogues, (iii) importance of price, availability, variety, and store location as drivers of demand for these products, (iv) socioeconomic and demographic characteristics of consumers of plant-based products, and (v) implications of the plant-based foods movement for California agriculture.

The aforementioned original objectives of the project have been pursued. An additional objective added as the research has unfolded is to understand the role that plant-based alternatives to traditional meats and dairy may play in global food production needs in the 21st Century. More specifically, to understand the extent to which plant-based food production can be an answer to the inefficient calorie conversion of grain and other inputs into human foods from animals. Growth of the plant-based sector may represent a market-based approach to reduce traditional meat and dairy consumption without coercive means such as meat taxes or taxes on emissions from livestock production.

Summary of Results:
This work comprises the dissertation research of Stamatina Kotsakou and is still an in-progress work with an estimated completion date of September 2023. The work involves extensive analysis of IRI data, both from the consumer panel and from retail scanner data from 2012 – 2021 using a mixed logit estimation framework with a control-function approach to address endogeneity. Results suggest that, despite the growth of household expenditures over time for plant-based substitutes, the demand for meat has not declined. Some evidence suggests that plant-based alternatives may in fact act as complements to traditional meats in the sense that a household with a vegetarian or vegan member may maintain its meat consumption and complement it with a convenient alternative for the non-meat eater.

Specific Objectives of the Project
The goal of this project is to analyze the efficacy, efficiency and equity of plausible regulatory policies that seek to reduce nitrogen emissions from agricultural soils.

Summary of Results
The research team used the funds to develop new understandings and create a research agenda around the use of biogeochemical models to perform policy analysis and an emissions inventory related to greenhouse gas emissions from synthetic fertilizer use in croplands.

We are actively collaborating with a social scientist at UC Berkeley to build a new tool based on the biogeochemical model, DayCent. We have (after some delay) gotten full approval to access the necessary survey data from the USDA. We have vetted our research questions with several experts. We are now building the computer code that will enable us to create the tool that will serve as the basis for our policy analysis and simulation. This ongoing work is being supported by a new Giannini grant, and we have identified several extramural grants to provide additional support.

In the meantime, we have some preliminary conclusions from our research pertaining to the magnitude of the greenhouse gas emissions costs from the use of fertilizers, which we are summarizing in a piece to be sent to ARE update this fall.

Specific Objectives of the Project:
The specific objectives of this project are twofold: i) estimate the basis risk (i.e., residual risk remaining after insurance) associated with the U.S. Department of Agriculture (USDA) Risk Management Agency’s (RMA) Pasture, Rangeland, and Forage (PRF) Insurance; and ii) determine the degree to which PRF Insurance mitigates forage-production-related risk for ranchers in California based on varying levels of risk aversion.

Summary of Results:
Agricultural producers who are reliant upon rangelands and pasturelands are some of the most vulnerable to weather-related risk given their dependence on climate-sensitive resources. Index-based insurance products, like the Pasture, Rangeland, and Forage Insurance Program, are considered to be an adaptation strategy for mitigating climatic risks. Given that indices are imperfect predictors of losses, residual (basis) risk persists. This study quantifies basis risk and assesses insurance contract quality for nearly 63 million acres of rangelands in California. Basis risk can be summarized by considering false negative probabilities – the probability that an insured producer suffers a loss without receiving an indemnity payment. On California rangelands the false negative probabilities associated with one (both) of insured time intervals failing to indemnify a loss range from 31% - 46% (14% - 25%). When indemnities were paid, in 36% of the cases the payments are not sufficient to compensate for forage-related losses; the average shortfall in indemnity-related compensation ranged from $1.74/acre to $2.73/acre depending upon the location and underlying value of forage.

Specific Objectives of the Project
We quantified the short-term impact of tariffs and the corresponding relief payments on soybean futures prices.

Summary of Results
Due to the 2018–2019 trade wars, U.S. agricultural and food products have suffered eight waves of retaliatory tariffs from Canada, China, Mexico, the EU, and Turkey. The COVID-19 pandemic further isolated the economies in 2020. We studied the futures market effects of these tariffs on various U.S. crops, particularly soybeans.

We estimated reduced-form regressions of real futures prices or spreads on retaliatory tariffs and a set of event indicators to quantify the short-term impact of retaliatory tariffs and the corresponding relief payments on soybean prices. We controlled for the COVID-19 epidemic, related U.S. government direct payments, weather shocks, and information from USDA reports. Our analysis used price data from the Barchart website, tariff data from official documents published by the tax bureaus of each country and the World Trade Organization (WTO) Tariff Download Facility (TDF) database, USDA World Agriculture Supply and Demand Estimation Reports, and weather data from the Google earth engine from 2004 to 2020.

We found that a 25% increase in a retaliatory tariff, holding projections and weather variables constant, decreased the real futures price by 14.25% while the tariff was in place. The effect on the futures price spread grew with the length of the spread. It reached its peak at a one-year spread. Thus, the price pass-through of the tariff increase was large, and farmers suffered from the retaliatory tariff in the short run.

Specific Objectives of the Project:

• Assess water utilities wildfire risk, with a specific focus on rural and agricultural communities;
• Establish an association between water utilities’ vulnerability level to wildfires with water utilities’ and communities’ characteristics; and
• Estimate short- and long-run effects of wildfires on households’ bottled water and water purifying products purchases, with a specific focus on rural and agricultural communities.

Summary of Results:
The project is completed for the first and second objectives, and a paper is submitted to the Journal of Water Resources Management. For the third objective, data is collected and merged/cleaned, but we are still working on refining the results. We do not have a working paper for this objective yet. Summary of the results from the first and second objectives are listed below.

Wildfires have occurred more frequently and are more devastating in California, and quantifying their impact on water utilities, which potentially may lead to water availability and water quality threats, is essential. This is especially important for water utilities whose characteristics are susceptible to wildfires. Due to the unpredictable nature of wildfires, drinking-water utilities face a considerable challenge in developing plans and strategies for managing floods and treating polluted water. Information and tools are needed to help water storage and treatment managers better prepare for the impacts of wildfires.

Our study quantifies these impacts by measuring the effects of wildfire on each water utility service area, based on the exposure frequency and the extent of acres burned by wildfires that occurred in each water utility service area, and by calculating the severity of the wildfire. Our quantitative models take into account the nature of the censoring and selection biases on wildfire data and show an association between water utility characteristics and the level of vulnerability to wildfire risks.

As a result of the cross-sectional estimation of the OLS, Tobit, and Heckman models, we found that wildfire severity increased in areas of (1) government-owned utilities vs. private-owned utilities; (2) utilities relying on surface water vs. those relying on groundwater; (3) utilities relying on local water sources vs. those relying on purchased water; (4) utilities located in Southern vs. Northern California; (5) utilities located on the coast vs. inland California; and (6) utilities in highly populated areas vs. non-populated areas.

Our findings can potentially inform public land managers and water utilities by identifying which water utilities are most vulnerable to wildfires based on their characteristics. They also show the potential characteristics of water utilities that are highly likely to experience changes in potential water availability and quality degradation immediately following a wildfire. Our research reveals that utilities more vulnerable to wildfires will require more strategic management decisions. It also suggests that the countermeasures on wildfires should be different depending on the characteristics the utility has. For example, based on an assessment of wildfire severity, by identifying the utilities and their locations of greatest risk, the relevant institutes could set costs associated with wildfire damage and mitigation activities.

Wildfire management and reducing the risk of wildfires are problems not only for the U.S. Forest Service or other public land management utilities but many other entities, such as water utilities. To help solve this pressing issue, partnerships are needed to help identify landscapes with hazardous vegetation and implement effective land management strategies by federal, state, local municipalities, communities, and nongovernment utilities. From this point of view, our study findings may help inform public land managers about possible shared stewardship partnerships with water utilities to leverage resources and expertise to reduce hazardous vegetation on shared landscapes, thus reducing wildfire risk.

Specific Objectives of the Project
1. To quantify the extent to which decreases in the supply of farm workers affect the quantity of fruit and vegetable crops produced in California.
2. To determine how reductions in the supply of farm workers affects the number of acres of fruits and vegetables that can be harvested in California.
3. To quantify how much farm revenue is lost because of reduced access to farm workers in California.

Project Report/Summary of Results
Recent studies point to a decline in the U.S. farm labor supply driven by demographic and structural changes in Mexico, increased U.S. border security measures, and a decline in the number of farm workers willing to engage in follow-the-crop migration, which has reduced the geographic reach of local farm labor markets. A smaller farm labor supply has the potential to reduce access to safe and healthy produce, increase the nation’s reliance upon foreign producers, and reduce the profitability of U.S. farm operations. In order to examine the extent to which changes in the farm labor supply may affect crop production, we estimate panel regressions using a rich set of production and employment data from California counties.

This study brings together three datasets pertaining respectively to crop production, farm employment, and weather. Our empirical strategy deploys fixed-effects panel regression
models at the crop-county-year level of aggregation, where the regressor of interest measures county-year farm employment during the peak harvest season. The identifying
variation comes from differences across counties in the evolution of employment about smooth county-level trends, net of weather effects. The fact that crop employment, an equilibrium value, is used as the explanatory variable in place of the underlying yet unobserved labor supply variable causes important identification challenges. We use an equilibrium displacement model to gain insight into the bias likely to affect our empirical estimates. Our regression results reveal statistically significant upper bounds for the
effects of labor supply shifts on the production of hand harvested fruit and vegetable (FV) crops, but, as expected, not on that of mechanically-harvested nut or field crops.

Our empirical results indicate that a ten percent decrease in the farm labor supply (in terms of the number of workers willing to supply labor at a given price) causes at most a 4.2% decrease in hand-harvested FV production in the top 10 producing counties, which together produce 86% of the value of all labor-intensive crops in the state. Reduced production is primarily channeled through a decrease in the number of acres harvested (- 2.8%), although we also uncover small yield effects (-1.4%). Impacts on the total value of production appear to be concentrated in the top 5 counties (Monterey, Fresno, Tulare, Kern, and Ventura). There, a ten percent decrease in the farm labor supply causes at most a 5.5% decrease in production value.

Although the bounds for FV production are economically meaningful, they indicate that the impacts of a declining farm labor supply will likely be limited in the foreseeable future. These effects are perhaps best exemplified by focusing on the top 5 producing counties, which produce 67% of the value of all labor-intensive crops in the state. Recent estimates suggest that the U.S. farm labor supply is shrinking by about one percent each year. A decline in the farm labor supply of that magnitude in the top 5 counties could cause a loss of 60,000 tons of hand-harvested FV each year. Under the assumption that the decrease in hand harvested FV production is not replaced by the production of other crops, production value losses of an additional 0.55% per year for the 69 crops we consider in those counties could add up to as much as $3.8 billion, or 3.0% of the total value of production, over the course of a decade. Our analysis also suggests that harvest mechanization could provide an alternative to the use of hand-harvest labor in a time of labor scarcity, provided that technologies are advanced enough to prevent unacceptable damage to fragile FV crops.

Specific Objectives of the Project
To summarize payments under a variety of federal programs to US farmers, and to relate the incidence of these payments to different farm-household characteristics, including county, acreage, crops, and need, as measured by county-level estimates of the distribution of farm-households' incomes, expenditures, and marginal utility of expenditures.

Summary of Results
We were able to obtain a complete record of all federal payments to US producers under all of the programs administered by the USDA over the last two decades via the Freedom of Information Act. We were subsequently able to link these payments to producer characteristics, aggregated to the zip-code level, available in the Agricultural Resource Management Survey (ARMS) through 2019.

Our project then foundered, as the USDA switched from using the NORC Data Enclave for managing confidential data to a new platform (the ADRF). The new platform is technically superior, but we have had to re-apply for permission to get ARMS data identified at the zip code level.

We remain excited by this project, and pending approval for data access from NASS we hope to return to it.

Specific Objectives of Project
On November 20, 2018, health agencies in the U.S. and Canada issued a food safety alert, advising consumers, retailers, and restaurants not to eat, serve, or sell any romaine lettuce or mixed salads containing romaine due E. coli. The Leafy Greens Marketing Agreement (LGMA) and other industry produce associations joined in this call on the same day, urging an industry-wide voluntary withdrawal of all romaine lettuce in marketing channels and inventory. On November 26, the United States Centers for Disease Control (CDC) and Food and Drug Administration (FDA) updated warnings to specify avoidance of romaine lettuce harvested from central and northern California. On January 9, 2019, the CDC declared the outbreak over, and Adam Bros. Inc. farm in Santa Barbara County was identified as the source of the outbreak.

Food-safety issues capture the attention of the media and consumers. Regulators face the challenge of balancing human health and welfare against the economic consequences of alerts and recalls. While farmers, handlers, food-service firms and retailers are all affected economically by a food-safety incident, the magnitude of financial losses and their determinants and distribution have been rarely studied over the entire supply chain. We use a unique combination of public and proprietary datasets to estimate the direct and indirect effects of this highly publicized outbreak on sales of romaine as well as other leafy greens for all of these supply chain actors. We further provide an estimate of the overall social welfare loss and discuss the magnitude of economic losses, incentives faced by supply chain actors, and implications for government regulation.

Project Report/Summary of Results
We decomposed damages into those associated with changes in prices for product that was sold during the outbreak and its aftermath, and those associated with romaine that could not be harvested, processed, and sold due to the outbreak and its aftermath. USDA, Agricultural Marketing Service data on farmgate prices, proprietary wholesale price data for the food service and retail marketing chains, and Nielsen retail scanner data allowed us to estimate changes in prices and quantities along the supply chain associated with the outbreak and its aftermath. We then use these results to estimate romaine prices and quantities in the ``but-for world'' that would have unfolded had the incident not occurred. The comparison of prices and quantities in the real-world and but-for-world scenarios for growers, handlers, and retailers reveal who benefited and who lost as a consequence of the incident. Recent changes in the structure of the produce industry and data limitations led to an estimation of damages under several scenarios and whenever possible, we consider additional losses or offsetting gains in other leafy greens categories.

For both romaine leaf and hearts, our regression results indicate spot-market prices were higher through the first six weeks of the outbreak than in the “but-for world,” indicating growers with romaine that was safe to sell during the outbreak was sold at a substantial premium. Price effects were negative during the remainder of the study period for romaine hearts and through week 1 of the post-outbreak period for romaine leaf, potentially due to decreases in demand for romaine and/or increases in the supply of safe romaine as warnings were removed from various growing regions. Further regressions found positive effects for contract prices paid to growers.

Regressions estimating quantity and price effects for food service wholesalers indicate that, relative to the “but-for world,” there were large decreases in the quantity of romaine sold to food service providers in the first week of the outbreak followed by smaller decreases through week 10 of the post-outbreak period. Similar to growers, wholesalers were initially able to capture a premium for safe romaine, but premiums were both smaller and more short-lived than for growers. Average prices paid to wholesalers increased weeks 2 through 4 of the outbreak by 24% to 41%, followed by moderate decreases in price through to the end of the post-outbreak period.

On the retail side, we estimated price effects for both wholesalers and retailers and quantity effects for retailers. For wholesalers, there is a negative price shock in week 2 of the outbreak in all categories followed by nonexistent or small and positive price increases relative to a no-outbreak scenario. We found modest or nonexistent retail price effects, with the largest effects being for romaine hearts. This is consistent with the fact that retail base prices are relatively stable in the short term. We find decreases in retail sales relative to the “but-for world” associated with nearly every week of the study period. The decreases are most pronounced in week 2 of the outbreak. While decreases in sales associated with the early weeks of the outbreak can be partially attributed to retailers removing unsafe product from shelves, the persistent decreases in sales throughout the post-outbreak period indicate consumers shifted consumption away from products containing romaine.

Based on the econometric results, we estimated that total romaine industry damages range from $52.7 million (grower contact prices vary, and buyers bear all pipeline damages) to $177.1 million (grower contract prices are fixed, and handlers bear all pipeline damages). Based on information regarding the extent of pipeline damage-sharing in each channel and contract price provisions, our most representative estimate of damages is $105.3 million. Cross-commodity effects for iceberg lettuce provided some offsetting gains. Food service firms can change the composition of their orders, like consumers in the store, but unlike retailers ordering bagged salad mixes. Iceberg lettuce producers obtained an estimated $11.5 million in benefits, likely due to such substitutions.

In addition to industry losses, there were losses sustained elsewhere in the economy. Particularly, we considered the effect of the outbreak on the welfare of consumers of romaine and suppliers of inputs to the romaine industry. Our estimated additional social losses range from $48.5 million to $59.9 million depending on the elasticity of demand for romaine lettuce.

In summary, our analysis provides two striking insights. First, the greatest losses occur in the retail marketing channel at the handler-retailer level of exchange, with an overwhelming share of losses in the retail channel. Second, growers gain from the food-safety incident, netting $4.3 million in our most representative estimate. These imply that as long as LGMA membership is voluntary and some handlers choose not to join, a subset of industry players are likely to be disproportionately imposing costs on the entire industry. Second, growers, who are in almost all cases the source of a food-safety incident, do not have a direct financial incentive to improve their practices to reduce the chance of an incident.

Specific Objectives of the Project
The objectives of this proposal are to empirically estimate the effect of the legalization of cannabis cultivation on agricultural labor, agricultural land prices, agricultural land use and water quality in California.

Summary of Results
Funding from the Giannini Foundation allowed us to begin to assemble a panel dataset comprised of satellite imagery data at a 30mx30m for agricultural land parcels in Humboldt and Mendocino counties spanning the years 1995 to 2020. We also constructed data on the cannabis laws in each county of California, including activities that are permitted and the date that these laws took effect. In our next stage of research, we will designate water bodies, forested land, roads and cannabis using a machine learning approach.

This research is ongoing. It quantifies the effectiveness and distributional equity of US maritime fuel emissions standards on air quality, racial exposure disparities, human health, and behavior. Defining where and to what extent the ECA policy affected air pollution for the on-land population is a first-order challenge. We quantify exposure to the policy based on the predictions of an atmospheric aerosol transport model. We leverage variation in (i) the timing of the regulation and (ii) the intensity of the regulation across locations, using a differences-in-differences design to measure the effects of the ECA. In preliminary results, we find the introduction of US maritime emissions control areas significantly decreased fine particulate matter, low birth weight, and infant mortality. Yet, only about half of the forecasted fine particulate matter abatement was achieved by the policy. Since maritime emissions occur off-shore, these abatement benefits were distributed more evenly across socioeconomic groups than those of land-based sources like ports. We show evidence consistent with behavioral responses among ship operators, other polluters, and individuals that muted the policy's impact, but were not incorporated in ex-ante models.

Specific Objectives of the Project
Cover crops have been proposed as a strategy to improve soil health in many crops, conserving water and mitigating effects of climate change (Pathak et al., 2018). In almonds, it has been suggested that bee-friendly cover crops could improve the health of honey bee colonies, adding another potential benefit to this conservation practice. Despite the numerous benefits of cover crops in almonds, less than 6% of growers use cover crop practices (Almond Board of California, 2015). Concerns regarding the economic return from cover crops are one of the top reasons for low cover crop adoption in the United States (CTIC and SARE, 2017). Because honey bee nutrition may be improved by cover crops, the use of cover crops may also provide an economic return to the beekeeper. It has been assumed that beekeepers will be willing to accept a discounted almond pollination fee if cover crops are grown in an almond orchard (DeVincentis et al., 2020). The objective of this project is to determine the size of discount beekeepers would be willing to accept for different cover crop mixes in order to facilitate wider adoption of cover crops in almonds.

Summary of Results
Using an online survey and a discrete choice experiment, we investigate and monetize the value of cover crops and other pollination contract components to commercial beekeepers who annually participate in almond pollination services. The analyzed sample of 81 commercial beekeepers represents approximately 19% of hives demanded for 2020-21 season.

We find that beekeepers value certain cover crop mixes, additional pesticide protections and up-front payment in their pollination agreements. Overwhelmingly beekeepers were willing to accept lower pollination fees for cover crop mixes that contain Brassica plants (e.g., mustards and canola) that bloom prior to and during almond bloom and were not willing to accept lower fees for a cover crop mix composed of legumes (e.g., clovers) that would not begin blooming until toward the end or after almond bloom.

Pesticide protection had the highest economic value to beekeepers: they were willing to give up $8 per colony in additional revenue to have a guarantee that almond growers would not tank mix pesticides and only apply fungicides at night when bees are not flying. For reference, the average pollination fee beekeepers reported receiving in 2021 was $189 per colony. Beekeepers were willing to give up $6.60 in revenue per colony if the grower plants the Brassica cover crop mix, and $5.60 per colony if the grower plants a cover crop mix that contains multiple species of brassicas, legumes and grains. If the beekeeper was paid 40% of their total pollination fee in advance of almond bloom, they were willing to give up $3.50 per colony in revenue. Assuming the preferred contract is adopted, beekeepers would be willing to accept a pollination fee per colony that is $18 lower than the standard agreement without those options, constituting a 10 percent reduction in the 2021 average fee reported by beekeepers. Using the industry standard of two hives per acre, this translates to a discount to the grower of $36 per acre.

Our findings show that there is potential for almond growers and beekeepers to work together to make mutually beneficial improvements in their almond pollination contracts. Importantly, two key improvements relate to additional pesticide restrictions and cover crops, practices likely to improve honey bee colony health and potentially lower abnormally high colony loss rates beekeepers have experienced over the last decade. High colony loss rates can severely impact beekeeping revenues from almond pollination, harming the economic sustainability of the U.S. commercial beekeeping industry. Thus, improved colony health will lead to a more stable supply of pollination services for almonds and other crops that bloom after almonds.

Specific Objectives of the Project:
• Complete a report that presents statistics on California’s organic agricultural production and the structure of the organic industry at the farm level using CDFA organic registry data
• Analyze the evolution of the industry’s structure utilizing the registry data and other data on organic acreage in California

Project Report/Summary of Results:
California’s organic agricultural production generated over $3 billion in farmgate sales in 2016, the most recent year in which detailed production information was available. Milk, strawberries, carrots, winegrapes and table grapes had the highest sales. Over 3,100 organic operations registered with the state. Over half of these operations produced fruit and nut crops. A majority of organic operations are in coastal regions. Organic operations have grown larger over time.

Specific Objectives of the Project:

The initial goal of the project was to assess the optimality of public reforestation projects in the context of both climate change mitigation and adaptation efforts in California and the U.S. We aimed to analyze the tradeoff between the climate mitigation value of afforestation and its cost in terms of wildfire risks. During the project, the objective became more specific. The focus shifted to the CARB’s Compliance Offset Program U.S. Forest Projects. The main final objective of the paper was to investigate whether the program’s design suffers from adverse selection that could lead to the program’s inefficiency.

Summary of Results:

In the program under consideration, an individual or an organization can get carbon offsets for engaging in one of the three activities: reforestation, improved forest management, or avoided conversion of forested land to other uses (such as residential areas or agricultural land). The main aim for each of them is to achieve a sustainable long-term growth of forests and maintain the stocking of trees at the project's designated land at a high level. The Compliance Offset Program does consider that wildfires might affect the forest projects and hence reverse the carbon sequestration. It treats the wildfire risk just like any other risk that can interrupt the project. To insure against such risks, the Air Resources Board maintains a Forest Buffer Account. Each project operator must contribute a part of their awarded carbon offsets to the Account. If an unexpected wildfire burns down the trees in the project area, the project operator will not lose the already awarded credits. The Account is insurance against such cases for both project operators and for the Program itself.

Our concern about this insurance-like buffer account was the potential adverse selection due to the essentially flat wildfire risk suggested by the program. We created a simple economic model of risk-neutral agents deciding whether to participate in the program. The model predicts that the potential forest projects with higher wildfire risks might be more likely to opt into the program because it provides partial insurance against wildfires. This adverse selection of the projects would make the Compliance Offset Program inefficient. First, the expected carbon sequestered, or the number of offsets supplied to the Cap-and-Trade market might not be optimal. For instance, fewer projects might participate, making carbon sequestration too low. In health insurance, it would correspond to a case when the equilibrium premium is too high, resulting in the lack of low-risk buyers. Second, the overall program will not be cost-effective because it could use more low-risk projects to sequester the same amount of carbon. Third, if high-risk projects increase the wildfire risks of the areas around them even more, it might put communities in danger and create other welfare effects.

To see empirically whether adverse selection is present in the program, we used open access data on the U.S. Offset projects, available on the ARB's website. At the time of the empirical analysis, 147 projects participated, but the program is ongoing, and more projects join each year. In our analysis, we used 79 projects because of the access to the GIS data for the other 68 of them. We compared the estimated fire risks of the projects' locations to the fire risks in the comparison group of other eligible lands. The data on fire risks comes from the USDA Forest Service.

We could not find any statistically significant adverse selection as of now. However, we did not fully finish the analysis yet, and the project is still in progress. Possibly, we did not detect adverse selection due to a small sample of the projects. We still need to estimate the locations of the remaining 68 projects. The total acreage of all 147 projects should be around 10 million acres. The second factor might be a comparison group of eligible lands. We are yet to refine the comparison group and exclude all potential lands that have a low chance of participation regardless of the fire risk of the project.

Specifics of the Project:
This project aimed to better understand consumer demand for plant-based meat. In contrast to other meat substitutes (tofu, veggie burgers), plant-based meat products are being marketed as indistinguishable in taste and appearance to meat. Producers of plant-based meats (PBM) aim to compete more directly with, and eventually replace, meat.

We received a panel of purchases from well over 200,000 grocery store shoppers who have bought plant based meat at least once, as well as a similarly sized control panel in a proprietary dataset from a nationwide grocery chain. We found that buying and more importantly, rebuying PBM is associated with having previously bought less meat and more meat substitutes. In addition, the people entering the PBM market are no more likely to have bought meat than those who first started buying it, suggesting PBM is struggling to expand its reach to those who could most easily switch away from real meat. In addition, because of how promotional pricing is determined at this nationwide chain, we were able to run event study regressions to test the theory that PBM has is a robust substitute for beef in grocery stores. In these regressions, we find little evidence for switching between meat and PBM.

Summary of Results:
After receiving the large sample of purchases, we worked to create covariates from the purchases. While PBM was only first introduced in limited stores in mid 2017, all purchases since the beginning of 2016 are included in the data. This allowed us to characterize households by their purchases in prior periods, as well as flexibly examine what types of purchases could ultimately predict who will buy PBM and who will rebuy it, with the ultimate aim of trying to better understand if PBM are attracting the types of customers who are likely to substitute it for real meat.

Covariates created from this pre-period, as well as those matched from a credit company who provides estimates of the head of household’s age and race, and the size and income of the household, were used to see how well we could predict three things: buying plant based meat, rebuying plant based meat, and the date of first buying plant based meat.

Those who bought PBM were more likely to be younger, to have a higher income, and to be shopping in more liberal voting areas. While households who bought PBM were more likely to be "low meat" households that spent less than 5% of their pre-period budget on meat, the difference was only a half percentage point. This supports the often repeated fact that PBM customers also buy meat. However, they still bought less meat, with a larger gap in their average market basket raw percent devoted to meat in the pre period. In addition, buyers of PBM bought about 3 times as much veggie burgers and tofu.

Rebuyers were more likely to be "low meat" purchasers and had spent more on traditional replacements and a little less on meat in the pre-period. That seems to infer that PBM was more likely to catch on with customers who were already interested in replacing meat in their diets. Early buyers were younger and less wealthy (although still wealthier than those who never bought at all). Earlier buyers were more likely to be ’low meat’, bought less meat and ground beef, and bought more meat replacements.

In order to examine whether the large number of covariates coming from pre period purchases could more accurately predict buying PBM, rebuying it, and buying it later or earlier, we implemented regressions using the least absolute shrinkage and selection operator (Lasso), a common machine learning algorithm that automatically selects a model so that many coefficients go to zero. While not causal, these models are used to better understand who the buyers were, and who was most likely to rebuy or buy later.

The OLS and Lasso results have less predictive power than hoped. The highest R^2s are for the buy/don’t buy decision. There, the OLS models (with nearly 500 variables) have an R^2 of just 0.16. The Lassos have slightly lower R^2s. Unfortunately, the Lassos only collapse the coefficients to zero of about half the categorical variables. This means we are not able to single out a handful of traits that are much more important in predicting the buy/don’t buy decision.

The models on early and late adoption and rebuying within 3 months have even lower R^2s, but are at least able to reduce the number of non-zero coefficients a bit better. When the sample is reduced even further to just California based stores, the Lasso was able to collapse many more coefficients to zero. Table 1.8 and 1.9 from our paper are reproduced below, showing the covariates related to rebuying and related to the week of first buy.

From a climate and animal welfare perspective, we would have liked to see rebuying popular with meat eaters, but instead, our Lasso regressions found the opposite (Table 1.8). Milk, bacon, beef, meat, pork, and ground beef are all predictive of not rebuying PBM, while meat and dairy alternatives, seafood, tofu, kombucha and expensive produce are related to rebuying. This table tells a clear story that customers who are more likely to incorporate plant based meat into their diets are much less likely to have had a diet high in the ground beef that PBM cheerleaders hope to be replacing.

While those households who bought more meat in the pre period were less likely to buy PBM, we were hoping that those types of households would be more likely to have bought PBM in the later period (as information about PBM started reaching more households due to increased media stories), but in table 1.9, we did not find that to be the case. Early adopters were more likely to have bought meat and dairy alternatives, tofu, and organic vegetables. This fits the story that early adopters of PBM were already interested in meat alternatives. However, none of the coefficients positively related to week of first purchase (meaning they bought later) were related to meat purchases, which suggests PBM was not spreading to meat eaters during the end of 2019. From these data, I conclude that PBM has a lot longer to go on making itself more attractive to the kinds of customers who are most likely to be substituting away from beef.

To study if PBM was crowding out beef, we used the as good as random variation of when PBM went on promotion (sale price). After interviewing officials with the corporate office of the grocery chain and running our own tests, we found that the temporary price cuts of PBM were imposed at the region level, and were not coordinated with each other, nor were they correlated with prior sales. We then used these events in event study difference in difference regressions.

While these regressions did show that PBM promotions greatly increase sales of PBM (figure 1.3 from our paper reproduced below), they have little to no measurable effect on beef (figure 1.4). There are a number of reasons to think these regression results are not perfect. First, the number of clusters, 12, was quite low, but that was the level at which prices were set. Second, the predicted demand change for PBM in the first week of the promotion (0.003 lbs) was probably not large enough to overcome the normal noise in the much, much larger ground beef markets. Yet, the main finding remains that we are unable to detect any effect of PBM on the demand for meat.

Specific Objectives of the Project:
1- Combine detailed data on nutrients along with scanner data on food expenditure (Nielsen, FCD-USDA and Label insight).
2- Model and
3- Estimate the functional form of demand with price aggregators.
4- Examine implications for price elasticities and substitution effects.
5- Publish the final article in a top refereed journal, and a summary in the ARE Update.

Summary of Research to Date:
We have cleaned and merged the three main databases (step 1); we have our main specification equation (step 2); and we are working on estimation (step 3).

Specific Objectives of the Project:
The overall objective of this research project is to analytically and empirically address nonpoint salinity pollution from agricultural activities by simulating a market framework for pollution permits and other regional cooperative arrangements between agricultural growers (polluters) and regulatory agencies aimed at controlling pollution to a water body (e.g., river, aquifer). That framework will then be applied to the case of salinity management in the San Joaquin Valley, California.

Specific objectives include:

1. Develop an analytical framework for quantifying, comparing and prioritizing policy interventions aimed at regulating nonpoint salinity pollution from irrigated agriculture.
2. Apply the analytical framework, developed under objective (2) to the case of the San Joaquin River Basin, California where agricultural producers are constrained by regulations that limit their ability to farm, leading occasionally to non-compliance with these regulations.
2.1 Compare several regional settings: including a status quo without regulation; a cap and trade of salinity pollution; and several cooperative arrangements among agricultural producers and between the agricultural producers and the regulating agency.
2.2 Introducing information and data sharing equipment to allow calculation of the value of information to the polluters and to the regulatory agency.

Summary of Results:
This project presents an alternative approach to salt regulation and control that follows first attempts to implement the 2002 TMDL, when it was realized that TMDL policy objectives could not be achieved without potential annual costs to stakeholders in the millions of dollars annually, using typical penalty schedules for daily exceedance of a 30-day running average EC objective at a single downstream compliance site. These costs would have potentially risen with the inclusion of two additional upstream compliance monitoring sites adopted to protect agricultural riparian diverters from high salt concentrations in irrigation applied water.

The novel concept of “Real-Time Water Quality management” relies on a continually updated forecasting model to provide daily estimates of salt load assimilative capacity in the San Joaquin River and assessments of compliance with salinity concentration objectives at three monitoring sites on the river, based on the 30-day running average EC. A water quality forecasting model WARMF was developed as part of this alternative regulatory schema, which served both as a compliance forecasting tool and the means by which salt load allocations and salt exports from each of the seven contributing subareas could be estimated and compared. The trading of salt loads between subareas is now feasible as both the regulatory salt load allocation and actual salt load discharge to the river can be quantified.

The results of the study have shown that the policy combination of well-crafted river salinity objectives by the regulator and the application of an easy-to use and maintain decision support tool by stakeholders have succeeded in minimizing water quality (salinity) exceedances over a 20-year study period. The WARMF model improvements, and consequent increase in stakeholder and agency confidence in this decision support tool, suggest its potential application in other river basins facing similar challenges. Our framework allows farmers and regulators to jointly understand and evaluate the meaning of various regulatory policy interventions on the emission of salinity and on the cost to be incurred by farmers at various locations along the river. The results support the development of close collaboration between farmers and regulators in the application of non-point source pollution policy. The results also suggest significant benefit from better cooperation and coordination among and between farmers and other dischargers of salt load who rely on the river for drainage disposal and who are already organized into sensible subareas for salt management. This can provide a cost-effective pathway for agricultural sustainability.

Specific Objectives of the Project:
In 2018 the U.S. government started a trade war with China and other trading partners. In response to import tariffs imposed by the U.S., China and other countries responded with retaliatory import tariffs, explicitly targeting U.S. agricultural exports. This study investigated the consequences of retaliatory tariffs on the agricultural and food industry in California. We had two specific objectives. First, we measured the ‘net’ impact of the retaliatory import tariffs on California’s agricultural and food exports. Second, we measured the extent to which U.S. federal trade-war subsidies offset producer losses arising from the retaliatory tariffs.

Summary of Results:The U.S. substantially increased a number of import tariffs in 2018, precipitating a trade war that was very costly to U.S. agriculture, given its dependence on international trade. The losses arose because a number of trading partners retaliated with import tariffs targeted at U.S. agricultural exports. The U.S. government then created a Market Facilitation Program (MFP) to compensate U.S. farmers for trade war losses. We have found that except for cotton and rice, California farmers were not made whole by the MFP payments. California’s producers of tree nuts, dairy, and processed fruits and vegetables were the biggest losers. Our research concludes that the U.S. government under-estimated the overall economic losses incurred by California’s agricultural and food producers due to the 2018 trade war. While the MFP program overcompensated some farmers, others faced substantial tradewar losses that outweighed MFP payments. Particularly export-oriented food processors were heavily affected by retaliatory tariffs but received no compensation from the U.S. government. The unequal treatment of agricultural and food producers impacted by the 2018 trade war is a pattern also observed in other states. However, these inequalities are more pronounced in California than in any other state. California’s producers focus on high value-added products and have a significant stake in reducing trade barriers everywhere and in particular in China. We believe the MFP payments may have jeopardized international trade arrangements because the excessive payments violated U.S. farm subsidy commitments to the World Trade Organization (WTO), and this could be challenged at the WTO. This means the effects of the trade war may drag on.

Specific Objectives of the Project:
The goal of this project was to investigate the implications of measurement error in remotely sensed land-use data for accurately pinning down parameters in economic analyses relevant to California agriculture. Data products that identify land use and crop categories from satellite imagery are commonly used in research and planning models, but they inevitably contain classification errors. To understand the implications of remote-sensing land use products for accurately pinning down economic parameters, I systematically compared different land use data products for California and evaluated the magnitude of the measurement error relative to detailed field surveys.

Summary of Results:
Measurement error in derived data products is known to pose challenges for inference. When the error is non-classical, which is often the case with satellite data, then it can introduce bias in parameters that are otherwise well-identified. Knowledge about the nature of the error becomes important for credible inference. To better understand the implications of remote-sensing land-use products for accurately pinning down economic parameters, this paper first assesses the accuracy of two different remotely sensed land-use data products with California coverage by comparing them against field surveys from an agricultural region on the California coast. I then systematically compare the two across the state, a geography that features a diversity of crops unlike the rest of the United States. I find that overall accuracy varies widely between the two datasets - 4.7% to 5.9% in one and 77.8% to 87.25% in the other, depending on the year. This validation exercise illustrates the nature and magnitude of the measurement error in California, with important implications for policy.

Specific Objectives of the Project:

• Distribute Fitbit activity trackers to workers on 3 strawberry farms in California.
• Construct measures of temperature and pollution from gridded temperature and satellite data.
• Merge the activity tracking and pollution data with hourly productivity of strawberry harvesters to examine the immediate, lagged, and cumulative effects of pollution on both activity and productivity.
• Obtain measures of activity and caloric exertion for agricultural laborers in California.

Summary of Results:
This project was significantly impacted by the COVID-19 pandemic. We were unable to access our field sites during in 2020 and 2021, which shifted the deployment of our data collection exercise until July 2022. During the grant period we acquired and developed all materials necessary to go to the field.
Accomplishments to date:

• IRB approval and all consent materials and survey instruments in both English and Spanish.
• Fitbit movement trackers were acquired and Fitibase software to manage real time data collection as well as sending reminders to participants was piloted.
• Code to create pollution and temperature measures during the planned intervention was obtained.

Specific Objectives of the Project:
Develop appropriate datasets and modeling framework to be used to estimate SGMA impacts.
Develop the modeling capability to simulate changes in crop mix in the San Joaquin Valley and to carry out the simulations.

Project Report/Summary of Results:
We developed a new approach for estimating supply elasticities for California tree crops. The purpose was to obtain reliable elasticities to populate PMP models of water basins in the San Joaquin Valley. Supply elasticities are estimated by regressing observed lagged acreage changes on estimated demand shifts and then converting the resulting regression coefficients into a price elasticity using the assumption of rational expectations. Estimated demand shifts can be estimated from observed prices and quantities if demand elasticities are known. We estimated possible ranges in demand elasticities from observed, unique, events that shifted demand. The supportable range of annual demand elasticities and the most likely elasticities are shown in Table 1 for important California tree crops.

Table 1. Estimated Elasticities of Demand for Citrus and Tree Nut Crops

California acreage and supply elasticity estimates for almonds, mandarins, pistachios, and walnuts are shown in Table 2. Estimated U.S. corn and soybean elasticities are also shown to facilitate a comparison of estimates made using our new approach and estimates made using more traditional approaches.

Table 2. Calculation of Supply Elasticity from Acreage Response and Demand Elasticity (1980–2019)

Specific Objectives of the Project:
(1) Estimate whether eucalyptus trees increase wildfire spread relative to other land use types, including other forest, cropland, or rangeland.
(2) Assess wildfire threat to California agriculturalists from eucalyptus trees

Project Report/Summary of Results: We exploit exogenous variation in wind direction to estimate whether eucalyptus trees increase fire spread. Contrary to popular belief, we find no evidence that eucalyptus trees increase total burn area relative to other forest types. Using the full sample of 188,926 fires in California between 1992 and 2015, we estimate that eucalyptus trees decrease total burn area by 8.8 percentage points. When we restrict the sample to the 7% of fires that burn more than 10 acres (0.04 km2), the effect of eucalyptus is indistinguishable from zero (0.8 percentage point decrease with a standard error of 0.6 percentage points). We conclude that California agriculturalists face little fire threat from eucalyptus trees. Removing eucalyptus trees does not appear to be an effective strategy for mitigating the threat of wildfires in California.

Specific Objectives of the Project:

The project’s objectives were to answer the following research questions:

1. Under a wide range of climate models and GHG policy futures, what will be the geographically distinct impacts of climate change on yields of key agricultural products (maize, soy, rice, wheat, sorghum, and cassava)?

2. What is the role of farmer adaptation to a changing climate, as well as changing incomes, in mitigating the effects of climate change on yields?

3. How much do the costs of farmer adaptation offset any adaptation gains they might realize?

Summary of Results:

Over the time period of the GF grant we estimated the responsiveness of spring and winter wheat to temperature and precipitation shocks, accounting for heterogeneity over local climate (long-run temperature and long-run precipitation), local income, and local access to irrigation. We additionally completed a Monte Carlo uncertainty analysis of the impacts of climate change on yields of six staple crops at global scale, accounting for adaptation: corn, soybean, rice, wheat, cassava, and sorghum. We found that adaptation is at least partially protective for all crops, but global productivity losses ranging from 25-45% percent (point estimates) persist for all crops except rice, for which adaptation is fully protective.

Specific Objectives of the Project:

Create and validate a structural econometric methodology to jointly estimate crop choice and yield response functions using satellite measurements of field level outcomes. The method will account for unobserved heterogeneity in field quality and the dependence on lagged dependent variables---two factors which characterize California agriculture. The method will allow us to separate the intensive and extensive margin responses to prices.

Summary of Results:
Estimation of a new structural model is a time-consuming process, so we have yet to complete the project. However, we have completed an intensive data-gathering and validation process and are wrapping up descriptive results that motivate and inform the structural model. We have begun simulation and validation exercises for the econometric model. The ultimate objective is to work on California crops, but since the set of crops in California is so complex, we focus our initial work on the simpler setting of corn and soy in the Midwest.

We have a number of findings that are consistent with the structural model of crop choice and yield that we plan to estimate. These include:

• Unobserved, persistent heterogeneity in corn and soy yields seems to be present in the satellite data. We are finding that the unobserved heterogeneity in corn and soy yields is correlated, but not perfectly.
• Fields seem to be specializing and making rotation decisions based on this unobserved heterogeneity.

Once we estimate a structural model of the production and profit functions, we’ll be able to simulate how rotation policies affect supply, crop choice, and nitrogen use.