Energy Heating Methods And Associated Risk of Poverty

By Aaron Murphy

Introduction

The detrimental impacts from the burning of fossil fuels has filled the news lately.

Fossil Fuels Increase CO2 emissions – leading indicator of climate change.

• Local, State, and Federal governments have implemented new policies to advocate newer, cleaner forms of energy.

This Capstone will look at the heating methods households choose in the state of Virginia and the associated risk this has on poverty.

Income inequality is of growing concern leading to energy “poverty”.

Virginia, a Mid-Atlantic state has an increasing population, requiring increasing energy usage.

Literature Review

• The United States must be at the forefront of energy development around the World; the US consumes 20% of all energy produced worldwide.

• Aslani, Alireza, and Kau-Fui V. Wong indicate 600,000 GWh of energy generated by 2030 will be from renewable resources.

• Daim, Tugrul, et al. growth curves show 14.7% of total energy production by 2023 will be from renewable energies.

• Pereira, Diogo Santos, et al. indicate that certain energies increase the risk of poverty; hydro and solar had little impact, however increased natural gas use correlated to higher rates of poverty.

• Xu and Chen show that low income households have much lower participation rates in energy efficiency programs / assistance with bill payment. Only 5.5% of low income households received assistance.

Questions & Why It Is Important

• Do specific heating methods have a higher connection to poverty than another?

• What does the rate of poverty look like in the state?

• What forms of energy are more widely utilized in the state of Virginia?

• This leads to further questions such as:

These questions need answered because energy and poverty have significant impacts on our quality of life. The targeting of state financial assistance and the promotion of energy development makes this the fundamental foundation of our understanding.

Data

Table 1: §1603 ARRTA program statistics March 2017. Data provided by US Dept. of Treasury.

Table 2: Number of projects funded by the ARRTA Program March 2017. Data provided by US Dept of Treasury.

The American Recovery and Reinvestment Tax Act (2009)

Table 3: The US State of Virginia statistics for the ARRTA Program March 2017. Data provided by US Dept of Treasury.

The program’s intention is to invest into renewable energy development projects across the United States.

Data

Table 4: The US State of Virginia 2017 Household Energy Method. Data provided by the US Census.

Energy Methods in the US State of Virginia

Methods

• Total households in Virginia will be statistically analyzed to see trends of heating methods.

• The state data will be compared to local data (independent city/county) within ArcGIS.

• Finding where high population exists in combination with higher poverty rates will show where investments need to go.

• This will be done by gathering data from county population, income, energy use data from the US Census Bureau.

• The data will be run through a series of OLS/Geographically Weighted Regressions (GWR).

• The dependent variable will be poverty rate/income.

• The explanatory variables will be the type of energy used.

Expected Results:

The anticipated results upon completion of this capstone are that households which utilize natural gas and wood heating sources are likely to have higher rates of poverty given the findings by Pereira, Diogo Santos, et al. (2019).

​

Data

The number of households in Virginia show clustering in the urbanized areas of Virginia, particularly in the southeast Tidewater, around the state capital of Richmond, and in the Washington DC region of northern Virginia.

The highest income found in Virginia are located primarily in the northern sections of the state (Fig 3) with lower incomes found in the west and south.

​

Significant spatial clustering exists in the western and southern portion of the state; focused especially on the Shenandoah Valley, the Appalachian plateau, the southern Piedmont

Energy Use By Type

Regression Models:

The OLS results show that a negative relationship exists between poverty and Bottled Tank/LP and wood heating sources. Whereas a more extreme positive relationship exists with the use of Fuel Oil/Kerosene heating. This means that households which use Fuel Oil/Kerosene as a fuel source tend to have higher amounts of poverty.

A GWR on these variables yields a slightly better R² & Adjusted R² value over the OLS, from 0.86 to 0.90 slightly lower AICc score

Results

The predicted poverty rates (from GWR) are highest across the western part of the state toward the Piedmont and Tidewater region; the northern part of the state nearer to Washington DC has the lowest predicted rates.

The warmer red colors indicate areas where the predicted poverty rate is underestimated, whereas the colder, blue colors indicate overestimations. This is most notable in the Virginia tidewater.

​

Model Performance

Summary of Results

• The most impoverished counties tend to have higher usage of Fuel Oil/Kerosene heating types – making up nearly a third of heating in counties across southern Virginia.

​

• The model’s performance had an adjusted R² value of .90

• The model explains 90% of the poverty rate; a statistically significant percentage.

• The Standard Residual Map shows where discrepancies are observed between the type of energy used and the poverty rate. The tidewater showed the most issue with overestimations.

• Using regression models has shown that correlation exists between the type of energy used and the rate of poverty.

• It is strongest with Fuel Oil (Coefficient 3.561136) and Electricity (Coefficient 0.231686) vs. other heating sources.

• Median incomes in the northeastern portion of the state are highest whereas incomes go down as you go south and west in Virginia.

• Looking spatially at the state the main energy used in the northeastern corner is gas provided by a utility. Whereas the rest of the state has a combination of heating types

Conclusions / Future Research

• The results from the regression models are not what was expected; originally, I anticipated wood heating and natural gas to be an indicator of poverty, as indicated in the by study by Pereira, Diogo Santos, et al. however, the OLS regression showed that in the state of Virginia there was not a significant correlation between natural gas and wood heating to poverty.

• A different approach that I would have taken is to utilize population densities and energy type. In addition, seeing how other demographics such as gender, age, and ethnicity correlate to energy use would be an area that would be intriguing and beneficial for those in government and nonprofits to know.

​