Proposed Objectives of the Project:
The first objective of this project was to investigate empirically the link between the media coverage of wildfires and the suppression efforts by U.S. agencies in charge of wildfire management. The second objective was to see if there is potential political influence on these efforts. We measure the suppression effort by the number of firefighting employees involved, the aircraft use, and the strategy choice. The media coverage is a function of the number of citations of specific wildfires in the news and other forms of media. We measure the potential political influence by the relative power of political leaders at different levels and their incentives to affect the results of upcoming elections. If at least one of the effects is statistically different from zero, the third objective will be to estimate the welfare effects of arguably suboptimal suppression efforts.

Summary of Results:
Over the past year, we have focused on the first objective. We have conducted an Instrumental Variables analysis of the effect of the News Publication about wildfires on the subsequent resources used in their suppression. Using U.S. data for the period 1999 to 2014, we find a positive and statistically significant effect of both mean and median numbers of daily news articles published about wildfires on resources used to combat them. We measure resources using two outcome variables. One is the total number of wildfire personnel involved in suppression. Another one is the total number of aerial resources used. These data come from the ICS209+ dataset, a repository of fire incident reports maintained by the Earth Lab at the University of Colorado, Boulder. Newspaper coverage is constructed using the publicly available database from the Newslibrary maintained by the Newsbank. We identify relevant newspaper articles by conducting a search by relevant keywords for each fire.

Our main instrumental variable is the presence of a “big event” (e.g., the War in Iraq) before a wildfire. We estimate it by measuring the information entropy of the New York Times front page articles. The idea is that wildfires occurring at the same time of a “big event” will receive less media coverage, everything else equal, than wildfires occurring during times where other newsworthy events are scarce. To the extent that big events are unrelated to firefighting resources, this strategy would deliver unbiased estimates of the impact of news coverage on fire suppression efforts. Our results are still preliminary as we refine our data, reduce the measurement error, and add more controls. Our latest set of controls includes weather variables (wind speed, temperature, precipitation, water vapor), state by year and season fixed effects, as well as an indicator for wildfires larger than 115,000 acres. According to the latest estimates, increasing the median daily number of articles about a wildfire by one leads to an increase in the number of total personnel by around 1,887 and an increase in the total number of aerial resources used by 20.4. The mean and standard deviation of total personnel and aerial resources are (1110.95,6689.16) and (24.31,85.12), respectively. These effects are admittedly large, and we expect their size to decrease as we refine the models and the data.

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:
The objectives of this research are to (1) quantify past adaptation to climate change in United States field crop production, (2) estimate the long-run consequences of climate change on this type of agriculture, and (3) assess the potential of one specific adaptation, double-cropping, to attenuate the potential losses resulting from climate change.

Project Report/Summary of Results:
We have developed a theoretical framework whereby agricultural production is related to both contemporaneous weather and long-run climate through simple quadratic functions. The framework implies a long-run response function to climate that is the outer envelope of short-run responses to weather conditional on climate. Applying this framework to a 66-year panel of yields and weather, we simultaneously estimate short- and long-run responses to climate for the United States (US) corn sector. We find evidence of significant climatic adaptation. In the case of temperature exposure (captured through growing degree days), traditional “myopic” panel methods that do not explicitly model climate adaptation nonetheless deliver estimates similar to those of our estimated long-run response. We formally show that this is due to the large cross-sectional variation in climate temperature relative to locational weather fluctuations. In contrast, for precipitation, which exhibits larger year-to-year variation relative to cross-sectional variation, models that do not account for climatic adaptation deliver estimates that are biased, but can be viewed as an average of the long- and short-run responses.

In complementary work, we focus on one specific adaptive action: double cropping. Although prior agronomic research suggests that climate change may expand the area suitable for double cropping in the US, there is relatively little empirical evidence of the relationship between climate and farmers’ decisions to double crop. We link high-resolution land cover data with detailed soil and climate data to explain farmers' propensity to double-crop soybeans with winter wheat in the Eastern United States. We find small and slightly negative effects of warming on double-cropped acreage. A fixed-effects panel model of county yields further indicates that yields of double-cropped soybeans are about 9.9% lower than those of single-cropped soybeans. Accounting for both of these effects, we conclude that double cropping is unlikely to offset negative impacts of climate change on US crop production.

Specific Objectives of the Project
The objectives of this research are to (1) quantify past adaptation to climate change in United States field crop production, (2) estimate the long-run consequences of climate change on this type of agriculture, and (3) assess the potential of one specific adaptation, double-cropping, to attenuate the potential losses resulting from climate change.

Project Report/Summary of Results
We have developed a theoretical framework whereby agricultural production is related to both contemporaneous weather and long-run climate through simple quadratic functions. The framework implies a long-run response function to climate that is the outer envelope of short-run responses to weather conditional on climate. Applying this framework to a 66-year panel of yields and weather, we simultaneously estimate short- and long-run responses to climate for the United States (US) corn sector. We find evidence of significant climatic adaptation. In the case of temperature exposure (captured through growing degree days), traditional “myopic” panel methods that do not explicitly model climate adaptation nonetheless deliver estimates similar to those of our estimated long-run response. We formally show that this is due to the large cross-sectional variation in climate temperature relative to locational weather fluctuations. In contrast, for precipitation, which exhibits larger year-to-year variation relative to cross-sectional variation, models that do not account for climatic adaptation deliver estimates that are biased, but can be viewed as an average of the long- and short-run responses.

In complementary work, we focus on one specific adaptive action: double cropping. Although prior agronomic research suggests that climate change may expand the area suitable for double cropping in the US, there is relatively little empirical evidence of the relationship between climate and farmers’ decisions to double crop. We link high-resolution land cover data with detailed soil and climate data to explain farmers' propensity to double-crop soybeans with winter wheat in the Eastern United States. We find small and slightly negative effects of warming on double-cropped acreage. A fixed-effects panel model of county yields further indicates that yields of double-cropped soybeans are about 9.9% lower than those of single-cropped soybeans. Accounting for both of these effects, we conclude that double cropping is unlikely to offset negative impacts of climate change on US crop production.

Specific Objectives of the Project

The objective of this project is to empirically investigate the role of public certification schemes in resolving information asymmetries in vertically quality-differentiated markets, using the French wine market as a case study.

When quality-differentiated products are sold by informed sellers to uninformed buyers, bad products may drive out good products from the market, resulting in an inefficient equilibrium where only low-quality products are sold. Although a potential for mutually beneficial trade in high-quality varieties exists, such potential is not realized because discriminative buyers cannot be certain of the quality of the goods they are purchasing. Certification schemes aimed at reducing information asymmetries between sellers and buyers can theoretically remediate the lemons problem.

We wish to test this hypothesis in the context of wine production in France over the course of the twentieth century. More specifically, we will investigate the extent to which a series of laws enacted throughout the first half of the twentieth century to reduce information asymmetries regarding the origin and quality of wines and spirits (in 1905, 1919, 1927, but most importantly in 1935) can be related to the emergence of a strong high-quality wine market. Our hypothesis is that the enactment of the 1935 law, which essentially codified the use of wine geographical denominations, allowed a segmented high-quality market for fine wines to emerge, resulting in higher welfare and higher market value for wine products in general.

We will use department-level data on wine acreage, production, and value, from 1925 to 1955 to test the hypothesis that the 1935 law fundamentally altered market structure and led to an increase in market value through product differentiation (leading to a separating equilibrium).

Project Report/Summary of Results

We have encountered delays in obtaining the data and are finalizing our data entry process (from paper yearbooks). We have managed to get out data without spending the funds. Some additional yearbooks have been released by the French statistical office (INSEE) during the past year, while others were obtained from the Cornell library and the UC Berkeley library.

This study assesses the economic viability and implications for water and cropland allocations of adopting switchgrass in the Central Valley of California. It adapts the approach of Positive Mathematical Programming (PMP) to construct amulti-region, multi-input and multi-crop model of agricultural supply. The model is calibrated against both observed input allocations and the exogenous supply price elasticities. For each of the regions, the model derives the supply curve of switchgrass thereby indicating the extent and location of potential switchgrass production in California. The study finds that the regions in or near the Southern San Joaquin Valley corresponding to counties of Fresno, Kings, Tulare and the cluster of regions at the frontier of the Sacrament and San Joaquin Valley appear to be the early and dominant adopters. Sensitivity analyses show that the findings are quite robust to a range of reasonable changes in the underlying technological assumptions regarding both switchgrass and existing crops.