Recipient Organization
WEST VIRGINIA UNIVERSITY
886 CHESTNUT RIDGE RD RM 202
MORGANTOWN,WV 26505-2742
Performing Department
Agri Economics
Non Technical Summary
This project aims to improve our understanding of 1) the livelihood impacts of power plant developments and operations, 2) potential changes in people's economic decisions on employment and occupations as a mechanism behind any effects from electricity generation developments, 3) comparison of economic development implications between the use of non-renewable and renewable energy sources for the electricity generation, and 4) the effectiveness of an economic revitalization program called Partnerships for Opportunity and Workforce and Economic Revitalization (POWER), funded by the federal government to revitalize regions affected heavily by the decrease in the coal mining industry in the Appalachia. Our unique contribution lies in assessing the impacts of energy developments on local prices, which has direct implications for local livelihoods but has been largely ignored in the academic literature. We combine a very unique geocoded household survey dataset that has information on local prices and quantity of consumer goods purchased by households with locations all power plants by energy source type to measure net impacts on households' welfare. Our impact assessment of the POWER project will provide further insights into the identified impacts of power plants and their mechanisms by quantitatively estimating how this kind of economic development initiative might or might not work.It is increasingly becoming essential to understand the direct and indirect impacts of past energy developments that might vary by energy source type in the face of the increasing pressure to mitigate climate change. The knowledge gained from this research will inform how we would be able to mitigate the negative impacts of transitions to renewable energy sources, therefore, determine future implications of energy policies on livelihoods. Moreover, the proposed research provides an excellent opportunity for a more fundamental understanding of changes in household behaviors under increased economic opportunities from new energy developments and how associated policies and programs for economic development might have implications for them.
Animal Health Component
40%
Research Effort Categories
Basic
40%
Applied
40%
Developmental
20%
Goals / Objectives
The overarching goal of this project is to understand the impacts of energy development, particularly those of power plants, on changes in local livelihoods and their mechanisms. We focus this investigation on power plants' impact on various livelihood outcome indicators, including income, employment, and local prices. This project also compares the effects of power plant developments and operations by the energy source type: non-renewable and renewable. We address the following two knowledge gaps building on theoretical predictions, empirical models, and findings in the literature by combining unique datasets of energy developments and household surveys.Objective 1: systematic analysis of the impacts of non-renewable and renewable power plant development and operations in the U.S. and the investigation of causal mechanisms of impacts by energy sources for electricity generationDespite the increasing number of studies that examine the economic and welfare impacts of energy development in various forms, few studies have considered general equilibrium impacts on local prices that can significantly offset increased income effects. This project aims to fill this gap by studying the impact of opening and closure of power plants on local prices of goods consumed by households along with other welfare indicators. By combining spatially-explicit power plant data of all types, nationwide household survey data, and other biophysical datasets in the U.S., this project enables the testing of impacts of opening and closing of power plants on local livelihoods and their mechanisms. Our investigation of the causal mechanism behind any effects involves the examination of changes in employment, occupation, and consumption patterns from the available household survey data. It contributes to providing general insights on household behaviors under changes in incentives and the economic environment, which will lead to the better establishment of policies that aim to achieve economic developments by changing household's behaviors in their economic decisions.More specifically, this project will compare the impact of electricity generation and their causal mechanisms by the source of energy: renewables and non-renewables. This comparison is a critical assessment reflecting the current changes in trends and policies in the face of efforts to mitigate climate change. The overall goals of U.S. energy policy are 1) to assure a secure supply of energy, 2) to keep energy costs low enough to meet the needs of a growing economy, and 3) to protect the environment while producing and consuming that energy from the 1970s (Yacobucci, 2016). From the late 20th century, many state governments have implemented policies to diversify energy sources for electric generation by incentivizing renewable energy development and emission reductions (Murray and Niver, 2020). Examples of state government policies include renewable portfolio standards (RPS) and the regional greenhouse gas initiative (RGGI). The RPS aims to increase the production of electricity from renewable sources to reduce the use of fossil fuels and nuclear for electric generation (NREL, 2020), and the RGGI regulates carbon emissions from the electric generation sector through a regional emissions trading program (Murray and Maniloff, 2015). As a result of the prior policies, the energy sources for an electric generation have been diversified. Over 70% of total electric generation was from fossil fuels, with less than 10% from renewable sources in the 2000s (EIA, 2020). However, fossil fuel use for electricity generation has declined, while the use of renewable sources has increased during the 2000s and 2010s. For example, the proportion of fossil fuels and renewable sources for electric generation was 62% and 17%, respectively, in 2017 (EIA, 2020). At the same time, the mix of fossil fuels has changed with the use of natural gas increasing due to its low price (Murray and Niver, 2020). Meanwhile, the use of nuclear energy for electric generation has remained constant, at around 20% of electricity generation during the 2000s and 2010s. Since the nuclear disaster in Japan occurred in 2011, some funding projects for nuclear plants have been canceled (Murray and Niver, 2020).Our comparison of the impacts of power plants by energy sources will contribute to the understanding of the implications of current and future energy policy. It will provide us with welfare implications for households living around power plants that use different types of energy sources and help us understand changes in their behaviors and decisions as we transition from non-renewables to renewables as the source of electricity generation.Objective 2: evaluation of the impacts of economic development programs to revitalize the economy within rural communitiesThis project tests the effectiveness of a federal government program that intends to promote the economic development of rural communities that heavily depend on the energy sector. We test the effectiveness of a program called POWER (Partnerships for Opportunity and Workforce and Economic Revitalization), which aims to help communities that have been affected by the decline of coal mining and coal power plant operations. It has been administered by the Appalachian Regional Commission (ARC) since 2015, where it has invested over $195 million in 242 projects spread over 350 counties in the Appalachian region. It has been helping communities with creating or retaining job opportunities by subsidizing such activities as diversifying sources of income, conducting job training, and attracting new investments. We plan to combine publicly available specific funding and locational distribution information from such organizations as ARC that manages POWER with household survey data. By doing so, we tease out the causal impacts of an economic revitalization program on economic indicators that are outlined in Objective 1. This effectiveness test will bring in more information, in addition to Objective 1, on what kind of interventions might or might not work in improving local welfare faced with a transition from coal to other types of energy sources.These tests outlined by objectives 1 and 2 will provide an unprecedented opportunity to understand the fundamental causes of changes in human behaviors in the use of human resources and the consumption of goods and services. This understanding can further advance economic theory and behavioral sciences by showing how changes in incentive structures caused by power plant developments and implementation of an economic development project can change households' economic decisions and welfare and what the mediators for those changes are. Practically, this understanding will be useful for policymakers and any agencies and funders that aim to improve the livelihoods of communities that heavily depend their livelihoods on coal extraction and power plants in the face of a transition of energy sources from coal to other types of non-renewables and renewables.
Project Methods
DataThis project combines spatially explicit information on locations and operations of all power plants in the U.S. with Consumer Panel Data collected by Nielsen for commonly available years between 2004 and 2018 to estimate the impacts of power plants on households' changes in welfare and investigate mechanisms. Historical records of all power plants in the U.S., including non-renewables and renewables, comes from the Emissions & Generation Resource Integrated Database (eGRID) issued by the U.S. Environmental Protection Agency (EPA) and the survey Form EIA-860 from the Energy Information Administration. It has information on locations of power plants, month and year of initial operation, operational status, primary source of fuel/energy, types and the amount of emissions, and total electricity generation amount. We use this information to determine our empirical strategy for the causal identification of impacts of power plants. We focus on non-cogeneration power plants that opened or increased the number of generators. We exclude cogeneration power plants because they are often constructed with other industrial plants or commercial buildings, which could confound the estimation of power plants' impacts (Davis, 2011). The Consumer Panel Data has approximately 40,000 households between 2004 and 2006 and 60,000 households since 2006. It will be made available through PI's collaboration with Prof. Katare at Purdue University. The project will use demographic and socioeconomic information from the data, which include economic outcome indicators such as household income, and the possession of assets, e.g., microwave, dishwasher, garbage disposal, tv, cable, internet connection. The unique data that we focus on using from the Consumer Panel Data are consumption patterns and price data. Consumption patterns include changes in the purchase of organic products, retail channels (e.g., dollar store, convenience store, grocery, etc.), product groups (e.g., frozen foods, packaged meat, dairy, deli, etc.), and purchased brands. The survey data also has information on prices of local goods that households have purchased, either reported or extracted from the point of sale data. We use information from the POWER program provided by the Appalachian Regional Commission (ARC) to test the impacts of an economic development project on local livelihoods and mechanisms. It has invested over $195 million in 350 counties across Appalachia since 2015. The program aims to revitalize the economy of communities that have been heavily affected by the decrease in coal mining and operations by supporting job training and re-employment opportunities and attracting new investments, for example. We use the information regarding counties that have been supported by the POWER program, funding amount, and different types of activities that are available through ARC.Empirical ModelThe investigation of welfare impacts and their mechanisms as a result of power plant developments in this project is based on the literature on local economic impacts of shocks in labor supply and demand (Roback, 1988, 1982; Rosen, 1979). We first investigate the effects of power plants on households' welfare by using household income, employment, consumption patterns as outcome indicators. The main mechanisms of welfare impacts that we investigate include changes in employment, occupations, and general equilibrium impacts on changes in prices of and expenditures on grocery and other durable goods. Given the non-random nature of power plant developments, we use event-study specifications to check any pre-existing differences in outcome variables and variations in impacts on outcomes after the opening of power plants. We employ different empirical model specifications using time and household fixed effects to check the consistency of results. We estimate the following equation for each household i for each fuel source type k in time period t:In this equation, is an outcome variable for household i in time period t; is a fixed effect for each household that controls for any time-invariant unobservable effects; is a time fixed effects that control for any temporal changes in time period t affecting all households; includes a constant and other control variables, including demographic, socioeconomic, and biophysical information from the Consumer Panel Dataset and other data sources. We use a binary indicator of equals to 1 if a household i is located within the same zip code of power plant k's location. Event year dummies are equal to 1 when the observation year is y= -14,..., 0,..., 14 years from the year , which is the time period that a household becomes located within the same zip code of a power plant k (y=-1 is omitted as the baseline). The point estimates of shows the difference between households who eventually becomes within the same zip code of power plant k compared to those who are not after controlling for all other covariates. The estimates of measure the effects of power plants on households within the same zip code in year y relative to the year before the power plant started its operations. We also consider the impacts of closure or opening of additional power plants and generators by multiplying the binary indicators of with the nameplate capacity of all power plants within a zip code in a given year. Under this specification, measures the effects of a unit increase/decrease in power plants' nameplate capacity on outcomes. We first use economic outcome indicators, including income and asset wealth, and consumption patterns that are alternative indicators for changes in wealth (Dubé et al., 2018; Kaplan et al., 2020). Next, we estimate changes in potential mechanisms such as employment and occupations as a result of the development of power plants using the same set-up with occupation categories as a binary dependent variable. Our investigation also involves changes in prices of goods purchased by households that are indirect general equilibrium impacts from the development of power plants but that directly affect the household's indirect utility. Any increases in prices of goods as a result of increased economic activities will offset any income gain from potentially increased employment opportunities caused by power plant development and operations (Jung et al., 2019; Moretti, 2011). We compare these impacts of power plants on the outcome and mechanism indicators by the fuel source: non-renewables and renewables as well as for each fuel source type k using the information about the plant primary fuel from the eGRID data. We examine the impacts of the POWER program on outcome indicators by using a similar set-up of the above equation with the replacement of the power plant's "intervention" with participation in projects funded by the POWER. We use the same economic outcome indicators, i.e., income, asset wealth, employment, and occupations, which are the expected outcomes that the program is aiming to improve. More specifically, we compare counties that have had POWER awards in earlier years starting from 2015 to 2017 to those that have had the awards between 2018 and 2019. This set-up reduces the selection bias by comparing those that participated early and those late. Our identifying assumption is that counties that participated early would have had the same trend in outcome variables as those that participated late without the participation in the POWER program. We test this assumption using our outcome indicators as well as other covariates by comparing them between the two groups before 2015, i.e., pre-treatment trends.