Proposed Objectives of the Project:
The objective of this study (Devin Serfas’ dissertation project) is to exploit a unique panel dataset from Saskatchewan (1) to model and measure the implications of alternative cropping choices—including tillage systems, crop rotations, and fallow—for quantities of atmospheric CO2 sequestered as soil organic carbon (SOC), (2) to measure the benefits and costs to farmers from adopting cropping choices that increase SOC over time, and (3) to explore potential policies for encouraging farmers to adopt cropping choices that sequester atmospheric CO2.

Summary of Research to Date:
The project has gone largely according to plan. Devin has completed good drafts of the first two chapters of his dissertation, corresponding to the first two objectives enumerated above, and has made a beginning on the third. The plan is to complete the first two chapters in the current quarter (Spring 2023), to make substantive progress on the third chapter over the Summer while in residence at the University of Saskatchewan, and to wrap up the dissertation during the Fall quarter, 2023. The Giannini Foundation mini-grant was used to leverage funding support from the Saskatchewan Wheat Development Commission which will continue to support Devin through the Summer and Fall of 2023.

Specific Objectives of the Project:

1. Estimate the short- and long-term impacts of agritourism on farms’ revenue, use of farmland, agricultural production, and agriculturists’ welfare.
2. Assess the differential impacts between small and large farms

Project Report/Summary of Results:

This initiative aimed to evaluate the short- and long-term impacts of agritourism on farms' revenue, use of farmland, agricultural production, and agriculturists' welfare in California. We managed to collect daily data on Airbnb rural listings and transaction records in California between 2015 and 2022. We also successfully developed metrics to assess agritourism development, incorporating the number of farm stays openings and occupancies on Airbnb near primary cities such as Los Angeles, San Francisco, and San Diego. In our attempt to establish causality, we employed non-farm stays as an instrumental variable for agritourism operation. We then applied a 2SLS regression model to examine the effects of agritourism development on agricultural outcomes.

Our analysis successfully assessed the short- and long-term impacts of agritourism on California agriculture, identifying key trade-offs between short-term economic benefits and long-term agricultural productivity. The primary findings include:

1. Over the short run, agritourism was found to have positive effects on farms' economic performance and sustainability, with small farms benefiting more than large farms.
2. In the long-term, agritourism led to the diversion of farmland resources away from agricultural production, with more significant impacts on small farms. The results suggested a trade-off between short-term economic benefits and long-term agricultural productivity due to agritourism activities.
3. The spatial distribution of agritourism activities varied across California, with certain regions experiencing more significant impacts on agricultural production and land use.

Our findings have substantial policy implications, providing insights into the efficient allocation and use of farm resources, and aiding in the formulation of specific strategies for small farms. Policymakers are urged to consider the trade-offs between immediate economic benefits and long-term agricultural productivity when crafting policies to support agritourism. Furthermore, our findings can guide the development of targeted strategies for small farms to balance agritourism activities and agricultural production.

In conclusion, our research offers valuable insights for cost-benefit analysis, allowing policymakers to compare the net benefits of agritourism with alternative strategies such as agricultural diversification and off-farm works.

Specific Objectives of the Project: The objectives of this project are to

1. Create an evidence based and detailed primer that characterizes supply and demand in the underlying market reflected by California’s Veles spot market water index (NQH20).
2. Use time series econometric methods to identify the determinants of spot water prices in the Veles water market.

Summary of Results:
Preliminary results provide the following facts about the market for water in California:

1. 82% of water trades are one-year leases, meaning that they are a temporary transfer of the right to extract either surface or groundwater that last for one year only. All the following stylized facts only apply in the market for one-year leases.
2. Most short-term water trades happen in the groundwater market, but most of the volume is traded in the surface water market. 75% of one-year leases are for groundwater in adjudicated basins, but these only represent 25% of the total volume traded.
3. The market for surface water responds to drought conditions, but the market for groundwater does not. Prices, price volatility, total volume traded, and number of trades increase in response to drought in the market for surface water but remains consistent in the market for groundwater.
4. In the groundwater market, virtually all transactions happen in May, and the price is constant over time at around $200. The limited number of transactions that take place during other times of the year generally have a higher price, and the price trends upward over time. Seasonality is also present in the surface water market, but it is less pronounced and less important.
5. Beyond seasonality, most of the price variation in the groundwater market comes from differences in location, specifically subbasin. Location and month fixed effects control for 92% of the variation in price for groundwater.
6. In contrast, most of the price variation comes over time in the market for surface water. During times of drought surface water prices can be five times higher than when there is no drought or less severe drought. In fact, during some months of the year when there is no drought the price of surface water can be close to zero. This wide range of prices reflects the dependence of the surface water market on drought conditions.

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.

Proposed Objectives of the Project:
Use artificial intelligence algorithms to inform Clean Water Act regulation; measure costs of Clean Water Act regulation of land use; study how water quality policy affects human health; study efficiency of groundwater extraction; investigate efficiency of spending on water infrastructure projects.

Summary of Results:

1. A substantially advanced paper builds a machine learning model, WOTUS-ML, to describe what the Clean Water Act (CWA) regulates. The paper has maps of 4 million points across the US and describes CWA regulation at each point under three recent Clean Water Act rules (Obama’s Clean Water Rule, Trump’s Navigable Waters Protection Rule, and the Supreme Court’s Rapanos rule), in addition to describing whether these points are adjacent, abutting, or isolated wetlands; ephemeral, intermittent, or perennial streams; or other water resource types. I hope the paper will be submitted and published in 2023.
2. This paper is in progress, the analysis has preliminary results using Zillow land transaction data (ztrax) that find large costs of CWA regulation (on a parcel that is 10 percent wetlands, CWA regulation decreases the parcel’s value by several percentage points). The analysis has now incorporated both results from NWPR in 2019 and SWANCC in 2000. The analysis needs some extra sensitivity tests, e.g., to account for differences in land use classes over time and space and wetland categories, but hopefully will be ready to start moving from tables and figures to a draft paper later this calendar year.
3. This water quality research is developing rapidly. It received an NIH R01 grant focused on drinking water pollution, and with the goal of preparing the drinking water pollution data to be available to other researchers. This paper has a complete draft that is being polished but has been presented a few times and should be ready for submission to a journal later this year.
4. The groundwater sustainability paper now has a working dynamic programming model to measure the sustainability of groundwater extraction in all major aquifers around the world. The analysis is seeking to incorporate some important hydrological features (e.g., porosity of underground aquifers and pressure).
5. I obtained initial results describing how discontinuities in local ballot initiative vote shares affect water investment. The number of usable initiatives in readily available data is fairly small, so I have put this project on hold while I consider additional ways to approach it.