Column 3a reports estimates of the previous coefficient α on the interacted variable for Equation 3. This equation also includes a triple interaction coefficient γ with female to determine the impact of drought on females. This coefficient γ is listed in column 3b. When we look at the female interaction that is added in column 3a, the reduction in income rises slightly to 13%. However, when looking at the triple interaction coefficient in column 3b representing the female minus male difference in drought impact, there was no significant or large additional impact on females either in employment or income. This signals that the drought did not disproportionately impact females in agriculture.We observe very different results when using a Hispanic interaction. Column 4a reports estimates of the coefficient α on the interacted variable for Equation 4. This equation also includes a triple interaction coefficient θ with Hispanic to determine the impact of drought on Hispanic individuals in agriculture. This coefficient θ is listed in column 3b. When observing the change in employment rate in column 4a, the regression observes no significant changes to agricultural employment as result of the drought. Column 4b reveals that virtually all ofthe reduction in agricultural employment due to the drought was for Hispanic individuals. This aligns with literature observing high rates of food insecurity among Hispanic agricultural families during the drought . Table 4 attempts to report estimated spillover effects.

This table uses Equation 2 as the only source for coefficients as it included all control variables. The addition of female interaction in Equation 3 tended not to significantly change coefficients,25 liter plant pot so it was not displayed for ease of viewing. Table 4 displays the most important outcomes for determining changes in each industry. This is repeated for each of my selected industries: agriculture, food manufacturing, food wholesale, food industry workers, transportation, and construction. These are arranged by ascending economic connection from agriculture. I observe significant reductions in agricultural income and employment, however, when looking at closely related industries such as food manufacturing and wholesale there are no significant reductions in either employment or income. There is a significant drop in food industry incomes, but when this is logarithm adjusted to account for trends in incomes the effect is absent. This suggests that negative spillovers from agriculture due to drought did not occur. I also observe large significant increases of 12% in construction employment accompanied by small increases of hours worked. This suggests that agricultural workers who lost jobs due to drought possibly began working in construction, which experienced relative expansion. This indicates that construction and agricultural jobs were treated as substitutes as the percentage of jobs lost in agriculture was fully recovered in construction. Table 5 scales the results from Table 4 to show the true impact of the drought on employment, hours worked, and income. I do this by using the IPUMS provided individual person weights to account for their true representation in the population, and scale these up by the percent of individuals in each industry. These are standardized to the change from 2012 pre-drought levels of each outcome. Hours worked weekly is scaled to show the aggregate change in hours worked per week in each industry due to the drought.

Since these changes are often marginal, this scale helps to better observe the way these incremental changes impacted California’s Central Valley overall.We can see that almost all of the spillover impact is limited to Hispanic workers.It also seems that upon further breakdown, spillovers that better fit the typical framework from input-output models can be seen in the Hispanic interaction, however these are not always significant differences. What is striking is that we can see clear evidence suggesting that Hispanic construction work increased proportionately with agricultural decreases. I attempted to measure the spillover impact of California’s drought from 2012 to 2016 from agriculture into closely related sectors. In prior difference in difference estimation of agricultural spillovers, short-term expansion was found in other sectors . We have seen that volatility in agricultural inputs and markets as a whole creates volatility in prices of outputs, impacting businesses that purchase agricultural products . There is also evidence suggesting that closely related industries experience the greatest spillover effects in the presence of positive or negative change . I estimate a difference in difference regression that tests the interaction between counties severely impacted by the drought and the time period of the event to find the causal impact of drought for individuals in each industry. I define severely impacted counties as San Joaquin and Tulare as they collectively contained 90% of fallowed land and experienced the highest groundwater pumping costs. The severity of the 2012 to 2016 California drought and its unique status as the most severe occurrence in recent state history would suggest that this event would have a severe impact on both agricultural incomes and employment. I find an estimated 9% reduction in agricultural employment and 11% reduction in agricultural wages.

There were reductions in Hispanic agricultural employment of 12% and 13% income decreases. I find no evidence of negative spillover in employment rate or income in closely related industries. While there was evidence of limited spillovers when analyzing Hispanic interactions, these were not always significant. This is particularly striking as it contrasts with the intuition from prior literature, and by empirical design the spillover impacts would be overstated . While the closely related industry spillover did not appear as in prior literature, there are many possible reasons to consider when questioning the lack of spillover. In the height of the drought, the state of California pushed an agenda to limit household water usage and provide aid to individuals living in rural areas. The Emergency Community Water Assistance Grant provided water to communities experiencing significant declines in availability or quality of water available. The Drought Housing Relocation Assistance Program was funded in 2015 to provide assistance and com- pensation to families forced to relocate due to a lack of access to water . There are also programs serving industrial interests. A 2014 Federal Farm bill requires farmers to rely on crop insurance as part of their contingency plan in case of drought to recover losses. In another novel way to mitigate losses, farmers began to trade their rights to water permits . Allowances are given to each plot of land, based roughly on seniority and size. In the face of the drought, some farms sold their water rights to other farms to recover losses from reduced crop sales. In some cases water was trading at $1.7/m3 . Even at averages of $0.8/m3 for the duration of the drought, prices were approximately three times regular non-drought water prices . There have been documented instances of farmer led water innovation in the face of mounting financial pressure. Kousky shows that when financially incentivized correctly through policy,black plastic plant pots farmers voluntarily implemented water management that created economic and environmental gains for all parties. This new form of drought mitigation likely reduced the decline in farm incomes and prevented a sharper spillover by stabilizing output prices. There is also evidence that the California agricultural industry was forced to invest in new water-saving technology and more efficient methods as a result of the drought . When analyzing the interactions with female and Hispanic individuals to see the difference in impact for these groups we see varied results. There was no additional causal impact from drought on either employment or wages for females, however there were substantial contractions in Hispanic employment and wages within agriculture. This aligns with prior literature that finds that rural and low-income households experience the greatest detrimental impact during crisis and natural disaster.

As Hispanic individuals make up the largest proportion of agricultural workers, this is a disparity that must be addressed by further research. The estimated impact to Hispanic individuals is likely to be understated in my results due to an uncertain amount of undocumented Hispanic agricultural workers that undoubtedly experienced labor market tightness during the drought. There are limitations to this data set because of issues with reporting undocumented labor in census data, but some estimates have attempted to solve this problem . There is also evidence that some agricultural labor switched to the construction sector as the increase in construction labor more than matched the agricultural decrease. When we look at spillovers for Hispanic individuals [Appendix: Table 8] we can see some significant declines in food wholesale. While these results are more closely aligned to traditional assumptions about spillovers, the lack of consistent significant results prevents conclusions about His- panic spillovers. Similar to previous regressions, when limited to Hispanic interactions we also observe significant increases in construction employment. This gives further credibility to the idea that construction was a substitute occupation for Hispanic individuals who lost or left jobs in the agricultural sector. While my study finds interesting evidence of policy and innovation working to limit negative spillovers from this shock, there are limitations to the data available. Without ex post data, there is a possibility that a more refined control group could be created to match pre-drought trends more accurately. This also prevents long-run impact analysis. Additionally, a wider range of data with more characteristics and a larger sample would improve the accuracy and validity of my result. For example, more granular data on the level of farm employment that distinguishes between farm workers and managers would allow me analyze welfare impact at a deeper level. This paper suggests that contractions due to drought have the potential to display limited local spillovers, despite having substantial impact on the local agricultural sector. As the frequency and extent of droughts increase, there is potential for adequate policy responses and a better framework for resource trading to continue to mitigate agricultural and local losses. These policy responses must account for the disproportionate contractions to employment and income of Hispanic agricultural workers to best serve the welfare of the agricultural and Hispanic communities. The beginnings of a network created by farmers to trade water rights and the ability of farm labor to switch to substitute industries are interesting topics for further investigation into the lack of spillovers as a result of the 2012 to 2016 California drought. The soil carbon cycle consists of inputs such as plant residues and exudates and outputs such as harvest, carbon dioxide released by microbial decomposition and soil carbon leaching . The carbon cycle equilibrium is often disturbed by the conversion of natural ecosystems to agriculture, and particularly so in tropical regions. This is mainly due to soil structure disturbances by cultivation and decreased carbon input as a result of typical practices such as burning or crop residue removal. These practices elevate soil organic carbon loss by increased rates of decomposition, organic matter oxidation, leaching, and erosion . Such management practices typically result in rapid SOC depletion followed by a slower rate of decrease for several decades, before a new SOC steady-state is reached . These losses can range between 20% and 70% of the original SOC content , but can be remediated with the use of cover crops and minimum tillage, when the residue is not removed . Reduced tillage increases SOC retention through macro-aggregate preservation , and has been proposed as a primary method for optimizing SOC in fine-textured soils . In recent years, agricultural practices common in tropical regions have been re-evaluated with the goal of gaining benefits from developing a winter fallow crop that would provide food, sequester more soil carbon and improve soil fertility and structure. Since deforestation rates are greater in the tropics than rates of current or historical land use changes in any other region , management practices that preserve soil carbon are expected to have a greater impact on reducing atmospheric CO2, and in doing so, potentially provide monetary incentives for farmers from the developing carbon markets . While much study has focused on the chemical and physical changes to soil due to different cropping systems, the associated shifts in soil microbial community structure and function remain largely unknown. Soil microbial communities are responsible for organic matter mineralization and are thus an integral component of soil productivity and the global carbon and nitrogen cycles. Here, we investigated microbial community shifts using plots under different land managements where the treatment effects on carbon mineralization rates , crop yields and soil carbon and on microbial biomass are documented.