Citation
The Determinants of Energy Intensity

Material Information

Title:
The Determinants of Energy Intensity
Creator:
Thaler, Dustin
Publication Date:
Language:
English

Subjects

Subjects / Keywords:
Energy ( jstor )
Energy efficiency ( jstor )
Gross domestic product ( jstor )
Heating ( jstor )
Imports ( jstor )
Manufacturing industries ( jstor )
Manufacturing processes ( jstor )
Renewable energy ( jstor )
Standard deviation ( jstor )
Standard of living ( jstor )
Energy consumption
Genre:
Undergraduate Honors Thesis

Notes

Abstract:
The objective of this paper is to answer the question of which country-specific factors have the greatest impact on a country’s energy intensity level. The analysis is not built on any known research on this subject, and the data has been taken from eclectic sources and independently brought together for the purposes of this paper. The paper discusses the dependent and independent factors, then provides hypotheses for the expected results of each factor backed by economic rationale. The paper then utilizes linear regression techniques with the gathered dependent and independent variables. The results, including hypotheses confirmation, levels of significance, and impact are disseminated and discussed. The paper then concludes with a discussion of research, results, and future prospects for energy intensity. ( en )
General Note:
Awarded Bachelor of Science in Business Administration; Graduated May 3, 2011 magna cum laude. Major: Economics
General Note:
Advisor(s): Lawrence Kenny
General Note:
College/School: Warrington College of Business Administration

Record Information

Source Institution:
University of Florida
Rights Management:
Copyright Dustin Thaler. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

1 UNIVERSITY OF FLORIDAGainesville, Florida THE DETERMINANTS OF ENERGY INTENSITYHonors Thesis submitted by Dustin Ryan Thaler 2011 Advised by Dr. Larry Kenny

PAGE 2

2 I. ABSTRACT The objective of this paper is to answer the question of which country-specific factors have the greatest impact on a countrys energy intensity level. The analysis is not built on any known research on this subject, and the data has been taken from eclectic sources and independently brought together for the purposes of this paper. The paper discusses the dependent and independent factors, then provides hypotheses for the expected results of each factor backed by economic rationale. The paper then utilizes linear regression techniques with the gathered dependent and independent variables. The results, including hypotheses confirmation, levels of significance, and impact are disseminated and discussed. The paper then concludes with a discussion of research, results, and future prospects for energy intensity. II. INTRODUCTION It is impossible to ignore buzzwords such as energy, efficiency, and sustainability as we enter the 2 nd decade of the 21 st century, and one of the most important measures of success in this looming energy era is that of energy intensity In the languageof business terminology, business productivity is to business efficiency as energy intensity is to energy efficiency. Thatis, energy intensity explains how much output a country can create (measured by gross domestic product) for the energy that it puts in (measured by quadrillion BTUs).The ratio is Total Energy Consumption/GDP in US dollars, which brings us to a number that describesthe amount of energy consumedper US dollar.The measure of energy intensity does not tell the whole efficiencystory,but one can certainly assume that a smaller ratio impliesthat a country is better

PAGE 3

3 at translating itsenergy into production. Thus,intensity becomes a proxy for oftdifficult-tomeasure efficiency. An article in Forbes Magazine that assessed the 10 most energy-efficient countries based on a countrys specific energy intensity number galvanized interest into this subject. Forbes declared that, Not surprisingly, the countries with the most energy-efficient economies are those who import their energy supplies, and [countries] with little domestic energy production are forced to import most of their fuel supply creating a powerful economic incentive to use those expensive imports efficiently. 1 However, energy imports alone certainly cannot be the end all be all of energy intensity determinants. So what else causes fluctuations in a countrys energy intensity? In this paper many factors are exploredwhether they be geographical, industrial, demographical or otherwisein order to discover the answer to that question. III. SAMPLE This is a cross-sectional study ofthe year 2006, which is the most recent year for which data areavailable. Thesample includes all countries for which data areavailable, which amounts to 160 countries. Notable exceptions include Iraq, Afghanistan, Taiwan, Georgia, and Puerto Rico. The special administrative region of Hong Kong was included as a separate entity from China. IV. DEPENDENT VARIABLE Energy Intensity 1 Forbes.com The Most Energy-Efficient Countries

PAGE 4

4 The dependent variable is energy_intensity which is defined as total primary energy consumption per dollar of gross domestic product(GDP). Total primary energy consumption is measured in British Thermal Units (BTUs) and GDP is measured using 2006 market exchange rates with year 2000 level US dollars as an index.A British Thermal Unit is defined as the amount of energy required to heat 1 pound of water 1 F. Data aretaken from the U.S. Energy Information Administration. 2 To avoid confusion in verbiage for the rest of this paper, an increase in energy intensity denotes a decrease in energy efficiency. For example, Denmarks energy intensity is 4,971 BTUs per US dollar of GDP, and Belgiumsenergy intensity is 10,563 BTUs per US dollar of GDP, which means that Denmark ismore efficient than Belgium.V. INDEPENDENT VARIABLES Standard of Living A high standard of living is indicative of increased education, increased consumption, increased leisure,increased technology, and across the board enhancementsin all aspects of life within a country. In economics, a better and more efficient environment can be considered a normal good. Thus, as standard of living increases (and income along with it), the demand for a better environment and a greater efficiency will increase as well. The independent variable standard_living in this case is defined asreal GDP perworker (GDP divided by the labor force). Data for the countries aretaken from the Penn World Tablesfor the year 2006.3 The predicted sign of this measure is negative, where an increase in standard of living will reduce energy intensity. 2 U.S. Energy Information Administration: Energy Intensity Data 3 Penn World Tables

PAGE 5

5 Education The education level of a country should have a profound effect on the level of energy intensity. With more education comes more environmental, energy, and economicawareness. Thus, citizens of a country with a high level of education should have greater knowledge in all aspects of academia, andperhaps more germane tothis study, they should have an enhanced knowledge ofenergy conservation and efficiency efforts. In this study, the independent variable education is defined using the 2006 Education Index, whichgives a two-thirds weight toadult literacy rates anda one-third weight to the combinedprimary, secondary, and tertiary enrollment rates. 4 The literacy rates denote an adults ability to read and write, while the enrollment rates give a strong indicator of the average level of educational attainment within a country. These two measures combined give us a good sense of how educated a given country is. The values of the education index range from a low .286 in Niger and Burkina Faso toa high of.993 in Denmark. It is hypothesized that an increase in education level will reduce energy intensity (increaseenergy efficiency) because the countrys inhabitants will have a greater understandingand awareness of energy usage.Energy Imports Many countries such as Japan and Italylack domestic resources like coal and oil required to produce their own energy.Without sufficientenergy an economy cannot operate atfull potential, so in order to bridge the gap between available domestic resources and full potential requirements, alack of energy must be compensated for byforeign imports. Importing energy is 4 Human Development Report: Indicators

PAGE 6

6 more expensivethan producing it domestically, so nationswho rely heavily on importsmust take special notice of how efficiently they use their energyto avoidbeing struckby high energy costs. The independent variable energy_imports is defined as the natural log of a nations production of energy in quadrillion BTUs divided by its consumption of energy in quadrillion BTUs or ln( Production/Consumption) Energy Production for this data set is defined as the production of petroleum, dry natural gas, and coal, and the net generation of hydroelectric, nuclear, geothermal, solar, wind, wood, and waste electric power. 5 Energy Consumption for this data set is defined asthe use of energy as a source of heat or power or as a raw material input to a manufacturing process, sources of which include petroleum, dry natural gas, coal, and net hydroelectric, nuclear, and geothermal, solar, wind, wood, and wasteelectric power.Information for these measures isobtained from the U.S.Energy Information Administrationfor the year 2006. 6 A ratio under 1,such as the United States .704(natural log = -.3508),implies that a country must importenergy in order to meet demand(in this case, 29.6% of its energy). These nations must take heed of the efficiency in which they use the importselse they get too costly. A ratio over 1,such as Qatars 4.798(natural log = 1.5683),implies that a country can meet its own demand and export the leftovers to foreign nations. Theseenergy-fruitfulnations have less incentive to be thrifty with the way they operate with their energy. Therefore, it is predicted that an increase in the Production/Consumption ratio will lead to an increase inenergy intensity, because as a nation nears its ability to cover its energy needs domesticallyit has less incentive to be prudent. 5 For the U.S., energy production also includes biomass and geothermal energy. 6 U.S. Energy Information Administration: International Energy Data

PAGE 7

7 Extreme Climate As is common practice, we turn down the dial on the air-conditioning to cooloffwhen it is hot outside, and we blast the heat to warm upwhen it is cold outside. These heating and cooling processes keep our homes and buildings at a comfortable temperature, but also require extra energy in order to do so. Countries that face consistently hot or cold temperaturesshould lose out on some efficiencybecause they are constantly pumping more energy into their buildingsand homes(non-production). The independent variable extreme_climate isdefined using atool called energy degree days,which measures variations from a reference temperature.A heating degree day (HDD) is a measure that occurs when it is cold, and a cooling degree day (CDD) is a measure that occurs when it is warm. An HDD is calculated by taking the average temperature for a day, which is the high plus the low divided by two, and subtracting it from 65 (thereference number). So, for example, if for a given day the average temperature is 35, it would be equal to 30 heating degree days (65 35= 30). A CDD is taken if there is a temperature above the reference of 65, in which case you would subtract 65 from the days average temperature. So, for example, if for a given day the average temperature is 75, it would be equal to 10 cooling degree days (75 65= 10). For this study, a years worth of monthly highs and lows were averaged and multiplied by the number of days within that month (31 for January, 29 for February, 31 for March, etc.) to arrive at the CDD and HDD for the specific country over the course of a year. The final HDD numbers are added to CDD numbers to arrive at the number known as energy degree days. Data for

PAGE 8

8 these numbers are taken from climatetemp.info. 7 The hypothesis is that as the amount of energy degree days increase, so will too the energy intensity, denoting a decreased energy efficiency. Manufacturing As a country increases its output, it needs more energy to support such activities. Countries have very different mixes of industry to reach their GDP levels, whether it isprimarily through manufacturing, construction, mining, services, research & development, etc.The most energyintensive process of all these industries, however, is manufacturing. 8 Thus, in this study an independent variable for manufacturing is included, which is calculated by the GDP due to manufacturingwithin a countrydivided by the countrys total GDP.Manufacturing industries are classified with an ISIC-D ( International Standard Industrial Classification) code and data are taken for the year 2006 from the United Nations National Accounts Main Aggregates Database.9 It is of note that other industries are also energy-heavy, but for the sake of simplicity and data attainment only manufacturing will be used.It is hypothesized that countries that rely more on the energy-heavy processes of manufacturing for their output willhave higherenergy intensity, and thus the sign of the coefficientfor manufacturing will be positive. Renewable Energy In an effort to become more sustainable now and forfuturegenerations, many countries have turned to adopting renewable energy sourcesas opposed to exhaustible sources. In this study the independent variable of renewable_energy will be defined as the amount of energy a country 7 www.climatetemp.info: Climate and Temperature 8 U.S. Department of Energy: Energy Intensive Industries 9 United Nations Statistics Division

PAGE 9

9 consumes from renewable sources (wind, solar, hydro, geothermal, etc.) divided by the countrys total level of energy consumption(Renewable energy consumption/Total consumption ). Data are taken for each country for the year 2006, and areobtained from the U.S. Energy Information Administration. 10 In terms of pure economic efficiency and how it factors into energy intensity, these renewable energy sources are currently costly and uneconomical when compared totraditional sources. This effectmeans that as the ratio of renewable consumption to total consumption increases, so should too the level of energy intensity,indicating reduced efficiency. On the other hand, a high share of renewable energy sources suggests a countrysforward thinkingand progressive nature. Countrieswith more renewable energy should also be the ones thatpay the most attention to matters such as energy efficiency. This effect works in the opposite direction of the inefficient renewables effect, and it is unclear whicheffect will dominate more. Therefore,a two-tailed test will be run for this variable.Fuel Prices High fuelprices force people to find alternative methods of transportation, whether it isby walking, bicycling, taking public transportation, or in various other manners. These methods of transportation are more energy efficient, as they reduce the amount of energy that goes towards driving, and thus non-production (with the exception of transportationof goods and serviceinvolved transportation as these are energy processes that go towards production). The independent variable fuel_price is defined as the weighted average of gasoline prices and diesel prices in each country, both measured in United States cents. A .60 weight is assigned to 10 US Energy Information Administration: International Energy Statistics

PAGE 10

10 gasoline prices, and a .40 weight to diesel prices. These weights are based on world consumption numbers, in which roughly 60% of all fuel consumed is gasoline. Data are taken from the German Society for Technical Cooperation (GTZ). 11 The data are from the year 2007, as the GTZ only publishes their study on fuel prices once every two years. It ishypothesized that a higher gas price will lead to higherenergy efficiency,and thus lower energy intensitybecause countries with high gas prices will be less inclined to use their cars and more disposed to finding alternative methods of transportation. VI. SUMMARY STATISTICS 11 International Fuel Prices 2007 provided by GTZ Variable Count Mean Std Dev Minimum Maximum energy_intensity 160 21414.709 24769.502 1089.960 (Chad) 190694.109 (Tajikistan) standard_of_living 160 27007.777 26113.154 996.59 (Congo DR) 126709.09 (Qatar) education 160 0.794 0.180 0.286 (Niger) 0.993 (Denmark) energy_imports 160 -1.092 2.163 -11.634 (Hong Kong) 4.536 (Chad) extreme_climate 160 5113.798 1970.189 1250.5 (Eritrea) 14243.5 (Mongolia) manufacturing 160 0.131 0.071 0.020 (Ecuador) 0.412 (China) renewable_energy 160 0.148 0.215 0 (Several countries) 1.243 (Paraguay) fuel_price 160 93.248 38.147 1.598 (Turkmenistan) 182.757 (Iceland)

PAGE 11

11 VII. REGRESSION RESULTS One multivariate linear regression was run, taking the form of the equation: 0 1 (standard_of_living)+ 2 (education)+ 3 (energy_imports)+ 4 (extreme_climate)+ 5 6 7 (fuel_price) A discussion of the effect and significance of each independent variable can be found below. The outputtable of the regression may be found on page 13.It is also worth noting that no significant correlation was found amongst variables, avoiding the problem of multicollinearity. standard_of_living This independent variable was found to have the predicted negative sign, indicating that an increase in standard of living causes a decrease in energy intensity and thus an increase in energy efficiency. The variable was found to be significant at the 1% level. A one standard deviation increase in the standard of living causes a decrease of 7625.56 in energy intensity, making it the third most impactful variable in this study.One opposing effect not factored in the hypothesis is that of household and commercial appliances. A country with higher standard of living will have more household and commercial appliances such as refrigerators, dishwashers, and washer/dryers, all of which require more energy to operate, increasing the energy intensity of the country and ultimately decreasing the impact of standard of living. education

PAGE 12

12 Education was the only variable that opposed the hypothesis. The predicted sign was negative, but the regression output resulted in a positive coefficient. The variable was also found to be significant at the 1% level. The impact of this variable is as follows: aone standard deviation increase in education causes a 7279.86 increase in energy intensity. One might suspect a high level of collinearity among education and standard of living. However, the two variables only have a moderate correlation of .5925. It is difficult to generate a logical economic rationale behind the coefficient of this variable, and so, the numbers have been examined more in detail. One possible explanation is that many of the low-education countries, such as Burkina Faso (.286), Chad (.334), and Niger (.286) also have extremely low energy intensities that could be due to other independent factors aside fromeducation. One such factor, for example, is that Chad relies on low-technology subsistence farming for much of its output. Thisspecific type of output does not require much advanced knowledge oran intensive amount of energy to produce. Countries such as Chad that rely on low-tech, low-energy, low-intelligence production may have resulted in a noteworthy skewing of the data. energy_imports The independent variable for the level of energy imports in a country was found to correspond with the hypothesis with a negative sign, and was also found to be marginally significant at the 10% level. A one standard deviation increase in thevariable was found to decrease the dependent variable by 2633.24. extreme_climate

PAGE 13

13 The regression output for the variable extreme_climate was in agreement with the hypothesis that the coefficient would be positive. This variable was found to have the highest level of significance, being considerably below the 1% level. Aone standard deviation increase in extreme_climate was found to increase energy intensity by 9816.92, which is the highest impact of any variable. The level of climate in a country is themost important variable among the tested variables in determining the energy intensity of a country.manufacturing The independent variable manufacturing (for the level of manufacturing as a percentage of total GDP) matched with the hypothesis that the effect on the dependent variable would be positive.The variable was found to have just barely missed the 10% cutoff for marginal significance, with a p-value of .10075.The impact of this variable is that a one standard deviation increase will cause an increase of 2229.93 in energy intensity, otherwise stated as a decrease in energy efficiency. renewable_energy The hypothesis was one of ambiguity, so a two-tailed test was run for this variable.The result of the regression shows that the coefficient has a positive sign, which can be interpreted withthe inefficient renewables effect that was previously discussed in the paper.This variable is significant atthe 1% level. The impact ofthe variablerenewable_energy isthat a one standard deviation increase will cause an increase of4499.12 in energy intensity.fuel_price

PAGE 14

14 The hypothesis for this variable, that as the fuel price increases a country's energy intensity will decrease, was supported by the results of the regression. This variable was found to have a significance that was markedly below the 1% level.The impact of this variable is that one standard deviation increase in fuel_price will decrease energyintensity by 8817.05, making it the second most important variable among ones tested.

PAGE 15

15 VII. REGRESSION OUTPUT *It should be noted thatall regressions were run using a two-tailed testby default. For all variables aside from renewable_energy the reported p-value should be divided by twoin order to get the actual p-value.SUMMARY OUTPUTRegression Statistics Multiple R 0.59978572 R Square 0.359742909 Adjusted R Square 0.330257386 Standard Error 20270.8165 Observations 160 ANOVA df SS MS F Significance F Regression 7 35093276518 5013325217 12.20066194 2.53086E-12 Residual 152 62457712257 410906001.7 Total 159 97550988775 Coefficients Standard Error t Stat P-value Lower 95% Upper 95% Intercept -15269.05456 9353.69041 -1.632409658 0.104663155 -33749.08359 3210.97447 extreme_climate 4.982733212 0.856293414 5.818955429 3.38948E-08 3.290959542 6.674506882 education 40443.66364 11927.59413 3.390764575 0.000888292 16878.38875 64008.93853 manufacturing 31407.52074 24482.95513 1.2828321 0.201503589 -16963.30421 79778.34569 standard_of_living -0.292015972 0.079477676 -3.674188633 0.000330113 -0.44903953 -0.134992414 fuel_price -231.1336466 47.50584639 -4.865372669 2.82972E-06 -324.9906573 -137.276636 log of energy imports -1217.402736 872.694282 -1.394993368 0.165052903 -2941.5795 506.7740285 renewable_energy 20926.13218 8344.687376 2.507719132 0.013201701 4439.584288 37412.68008

PAGE 16

16 VIII. DISCUSSION AND CONCLUSIONGetting back to what sparked interest into this paper, analysis of the data indicates that, while the level of a countrys energy imports does playa role in its energy efficiency, it is not nearly the most important variable. In fact, the most important variable, extreme climate, is the only variable studiedthat is beyond the control of a countrys reigning regime. The other factors may be indeed difficult to change, but they are not immutable. A couple caveats due to the climate variable. One, heating and cooling technologies differ. Heating requires more energy than does cooling, which, if weighted properly, would ultimately change the final numbers for that variable. However, an ample weighting mechanism was not discovered despite contact with a few experts. This will be something improved upon in future editions of this paper. Two, the climate variable does not take into account the amount of construction that requires heating or cooling within each country. For example, Mongolia may have the highest extreme_climate number, but its overall population of 2,670,966 suggests that it does not have nearly the amount of buildings to heat as Russia with its population of 141,850,000.This will also be addressed in future editions of the paper. Also,a quick noteregarding the renewable energy variable. Lately, so much attention has beenput towards developing efficient renewable energytechnologies. As the necessity for renewable energy (or some alternative form) becomes more apparent, there could be a day somewhere in the near future where the inefficient renewables effect described in this paper becomes moot.

PAGE 17

17 Another interesting pointdiscussed in an article in The Economist12 is that over a period of years, energy intensity usually peaks when the activity level of heavy, energy-intensive smokestack industries within a country is at its highest. For the United States, this peak came around the 1920s, and for China it came around the 1960s. As countries move towards more modern lighter industry and service-driven economies, that peak has fallen, and it is predicted that it will continue to fall. The idea that energy intensity is at its highest when energy-intensive industry is also at its highest may seem paradoxical to the results of this paper in which the manufacturing variable just missed the cutoff of marginal significance, however that result may be due to the fact that this paper is a cross-sectional of the year 2006. If the same paper were to be researched in 1920, the results may have been vastly different. So what should be taken away from this paper? Well,for one,thatclimate, fuelprice, and standard of living are by far the most impactful and significant factors to look at when dissecting the energy efficiency of a country. Unfortunately this brings me to another conclusion: That the future prospectsare bleak for an underdeveloped, extreme weather nation to improve on its energy efficiency due to the intractability of the standard of living and climate variables. These countries were likely put in low standard of living situations in the first place because they lacked the required resources to develop into global forces. Sometimes, itcomes down to the luck of the draw. The United States has been blessed with tremendous amounts of resourcessince its nascent stages, which has ultimately led to ahigh standard oflivingand our assertion as a dominant power. We also have relatively tepid temperatures overall. Hopefully,we can continually improve in other areas so that we may one 12 http://www.economist.com/blogs/dailychart/2011/01/energy_use

PAGE 18

18 day be a global leader in energy efficiency, a hot-button issuebound to be inextricably linkedto allnationsas we head into the unforeseeable future.

PAGE 19

19

PAGE 20

20