Michael Flaherty, Christian Gordon, Christian Lopez Financial Markets & Instruments Professor Goldstein 12/3/2012
Table of Contents Executive Summary 2 Natural Gas Overview 2 Market Structure 4 Factors Affecting Demand 5 Factors Affecting Supply 6 Industry 6 Process 6 Hydraulic Fracturing- Fracking 8 Risks 8 Regulation 9 Statistical Analysis 10 Simple Regression 11 Multiple Regression 12 Conclusion 13 References 14 Regression Analysis 16 Exhibits 23 1
Executive Summary Natural gas is one of the most abundant natural resources found throughout the United States and is a major contributor to the fueling of America s economy. The process of attaining natural gas is very complicated that comes at a high risk, but with an even higher return on investment when operations run smoothly. As technology in the industry has improved and developed over the past 50 years, the operational risk in natural gas is decreasing, however, their exposures are increasing as result of the industry growth. With the earth suffering a great deal from inefficient environmental practices that have been put in place to drive America s economy in the past and present, natural gas shows tremendous upsides when it comes to the carbon footprint that is left behind, while generating a large amount of energy at the same time. In this paper, we analyze the natural gas process and the market that it performs in, along with the key economic factors which affect the price that natural gas is sold and traded at. Additionally, a statistical analysis is performed to show how these economic factors work with one another and what can be expected out of the natural gas industry in the future. Natural Gas Overview Natural gas is a third primary fossil fuel which was originally discovered during scientists search for oil 1. Natural gas, like crude-oil and coal, is an energy commodity which is traded physically but can also be traded as futures. Compared to crude oil, natural gas has a higher octane level and energy content making it a more efficient source of energy as prices for natural gas remain lower 2. At standard temperature and pressure it is in the form of gas and its main chemical component is methane. When it is extracted from the ground it also contains water, butane, propane, ethane, pentanes, sulfur, carbon dioxide and other compounds which must be removed before entering the pipelines 3. When natural gas is burned the only by-products of the combustion are only carbon dioxide and water, therefore, making 1 Boorse, Dorothy, Wright, Richard. Environmental Science, San Francisco, CA, 2011. 2 America s Natural Gas Highway, <http://www.cleanenergyfuels.com/video/anghvideo.html> 3 Background, <http://www.naturalgas.org/overview/background.asp> 2
it the most environmentally desirable energy source in comparison to coal and oil 4. For many years, there was no practical way to transport it from wells to consumers because when oil drillings occurred, the gas was released from the fields and became flared 5. As knowledge and technology improved surrounding natural gas, companies are now able to stop this harm from occurring to the environment, and also make great use out of this valuable fuel 6. A network of pipelines was constructed to connect consumers to wells and as a result, the use of natural gas rapidly increased for heating, cooking, and industrial processes. On top of natural gas being clean and relatively inexpensive, it is also convenient because it does not require storage bins or tanks to be on site, such as at a residence 7. As an energy resource, the United States holds some of the largest natural gas reserves in the world according to a 2007 National Petroleum Council with 1,451 trillion cubic feet of Natural Gas. As a result, the U.S. is capable to be a net exporter of natural gas which will work towards the goal of becoming energy independent. Future projections for natural gas estimate that the worlds demand for natural gas will increase 50 percent from 2005 to 2030 8. The Southeast Asian region is also projected to be the largest consumer of natural gas and it will be imported in the form of Liquefied Natural Gas (LNG) 9. Although natural gas prices in the U.S. are currently low, countries such as Japan are experiencing high LNG prices as the world s top LNG consumer 10. This can be attributed to the low supply but more of the increased demand for electricity as a result of losing the nuclear power plants in 4 Boorse, Dorothy, Wright, Richard. Environmental Science, San Francisco, CA, 2011. 5 Boorse, Dorothy, Wright, Richard. Environmental Science, San Francisco, CA, 2011. 6 Background, <http://www.naturalgas.org/overview/background.asp> 7 Boorse, Dorothy, Wright, Richard. Environmental Science, San Francisco, CA, 2011. 8 Kendall, James. Global Gas Outlook, <http://www.eia.gov/pub/oil_gas/natural_gas/presentations/2008/globalgas/globalgas.html> 9 Kendall, James. Global Gas Outlook, <http://www.eia.gov/pub/oil_gas/natural_gas/presentations/2008/globalgas/globalgas.html> 10 Loh, Bohan. Japan LNG demand has likely peaked, <http://www.reuters.com/article/2012/10/23/us-japan-meti-lng-idusbre89m08720121023> 3
the tsunami disaster 11. However, if the U.S. decides to export LNG in the future, an increase in the price of natural gas within the U.S. must be avoided to maintain efficiency. Market Structure In the United States there are over 210 natural gas pipeline systems, 305,000 miles of interstate and intrastate transmission pipelines, and 49 locations where natural gas can be imported or exported via pipelines 12. In 2010, natural gas consumption in the United States totaled 24.71 quadrillion BTUs, which makes up 25% of the total fuel energy consumption in the country 13. Natural gas annual consumption is forecasted by the year 2035 to reach 27.24 quadrillion BTUs with only renewable energy predicted to compare to this same level of increase in additional energy consumed over the same period 14. Natural gas is consumed and demanded by different market segments throughout the United States. Electric generation is responsible for 30% of natural gas consumption, 27% is from Industrial, 21% is from Residential, 13% is from Commercial, and 9% is classified as other 15. The Electric Generation sector is predicted to continue as the highest consumer of natural gas with 60% of new electric generation capacity built by 2035 predicted to be natural gas combined-cycle or combustion turbine generation 16. The retirement of old nuclear, petroleum and coal powered generation plants will be replaced by natural gas-fired combined cycle generation plants that require relatively low capital 11 Loh, Bohan. Japan LNG demand has likely peaked, <http://www.reuters.com/article/2012/10/23/us-japan-meti-lng-idusbre89m08720121023> 12 About Natural Gas Pipelines <http://www.eia.gov/pub/oil_gas/natural_gas/analysis_publications/ngpipeline/index.html> 13 Beauman, Jeffrey. Natural Gas Storage Transmission and Distribution. June 5, 2012. FM Global. 14 Beauman, Jeffrey. Natural Gas Storage Transmission and Distribution. June 5, 2012. FM Global. 15 Beauman, Jeffrey. Natural Gas Storage Transmission and Distribution. June 5, 2012. FM Global. 16 Beauman, Jeffrey. Natural Gas Storage Transmission and Distribution. June 5, 2012. FM Global. 4
requirements 17. The consumption of natural gas in the Industrial Sector is also predicted to maintain growth at an average rate of 0.9% per year through 2035 18. Factors Affecting Demand Similar to any other commodity traded, natural gas prices are determined based upon the supply and demand. In the short run, demand for natural gas is cyclical and therefore so are prices. In its history, natural gas prices were always affected by the seasonal weather patterns 19. This is due to the fact that natural gas is used for heating homes during the winter which means that the prices will increase with demand in the months of January and February. Although this still plays a large role in the determinants of demand today, the uses for natural gas have been diversified which has slightly interrupted that cycle. Advances in technology has made it effective to use natural gas to power trucks and buses, to produce electricity, and run manufacturing plants. As a result, the demand with those new uses is more consistent throughout, and as a result it will bring more consistent demand. One reason for more constant demand is due to increased electricity in the slower months (summer) where natural gas is now the source of producing the electricity to cool commercial buildings and homes. Also with new technologies, the ability to switch fuels is much easier and is not as expensive, so this gives the electricity producers the ability to switch from natural gas to the alternative coal in times of high natural gas prices. As a result, this ends up helping the price lower back down to the equilibrium, ideally. These switches made by some electricity producers will have a significant effect on the short term demand because in 2011, power generation accounted for 34.1 percent of US natural gas consumption 20. Short term demand can be affected by the state of the economy. In a recession, output decreases which will lower the demand for energy in the industrial sector, and increase the demand for energy when output is up. 17 Beauman, Jeffrey. Natural Gas Storage Transmission and Distribution. June 5, 2012. FM Global. 18 Beauman, Jeffrey. Natural Gas Storage Transmission and Distribution. June 5, 2012. FM Global. 19 Natural Gas Demand, <http://www.naturalgas.org/business/demand.asp> 20 Capital IQ; NetAdvantage, Industry Surveys: Natural Gas Distribution, July 12, 2012. 5
Long Term Demand is determined by the long term adoption of natural gas uses throughout the economy. Investment in Natural gas public transportation fleets, new technology expanding to new industries, and government/environmental regulations in favor of clean energy will all increase the long term demand of natural gas. Factors Affecting Supply In comparison to the changes in short term demand, short term supply on average experiences less fluctuation. This is due to the nature of the process and the longer lead times. Most often, companies aim to produce the optimal amount of natural gas to meet demand. However, it is not economically efficient to shut down active wells because of the investment and costs of operation. As a result, the supply can t easily be reduced and consequently if there was a need to increase supply, the lead times are no shorter than 6 months in many cases 21. To offset this, natural gas is stored in locations near the end users to meet sudden/unforeseen increases in demand. Factors that affect long term supply are more obvious being the actual natural gas reserves. Other determinants are the total cost of extracting to delivering the natural gas, investments in pipeline infrastructure, and capital resources for new drilling sites 22. Industry Process The process of delivering natural gas to the end user consists of several steps beginning with physically removing the gas from beneath the Earth s surface. Enabled by modern technology, scientists are able to survey the rock formations to know how much natural gas can be extracted from the particular location before investment 23. Once a location s potential is realized, water and chemicals are pumped at high pressure deep into the rock formations which will enable the natural gas to be 21 Factors Affecting Gas Prices, <http://www.eia.gov/energyexplained/index.cfm?page=natural_gas_factors_affecting_prices> 22 Natural Gas Supply, <http://www.naturalgas.org/business/analysis.asp> 23 Natural Gas, <http://naturalgas.org/naturalgas/naturalgas.asp> 6
recovered 24. The natural gas is extracted in its raw form which contains water and impurities that must be removed before it can be used for energy. Once the natural gas reaches the wellhead, it is then piped into a processing plant which separates the methane from water, excess hydrocarbons, carbon, sulfur, and crude oil (if present) to make dry and usable natural gas 25. Once it reaches the form of dry natural gas, also known as pipeline quality, it is then transported through interstate transmission pipelines to local distribution companies (LDCs) 26. These transmission pipelines are made of carbon steel and range in diameter from two to 60 inches 27. The newest pipes in the distribution channel are externally coated with cathodic protection to prevent rusting 28. In order to keep the natural gas flowing, compressor stations are located along the pipelines where pressure of the gas is boosted in order to compensate for pressure losses along the pipeline 29. LDC s transport the natural gas to city gates, where local utility companies take ownership of the natural gas and then it is either stored or delivered to the end user 30. The entire process can also be categorized by four basic processes, which are accomplished through a company s gathering systems, transmission systems, gas distribution, and power generation 31. Gathering systems includes raw natural gas from production wells on land, offshore drilling, liquefied natural gas from overseas tankers and raw gas storage 32. Transmissions systems involve gas processing, pipeline systems transporting natural gas thousands of miles across country 33. The equipment is typically tanks, reservoirs, storage facilities, pipes, pumps, valves, racks, compressors and loading 24 Natural Gas, <http://naturalgas.org/naturalgas/naturalgas.asp> 25 Natural Gas, <http://naturalgas.org/naturalgas/naturalgas.asp> 26 Natural Gas, <http://naturalgas.org/naturalgas/naturalgas.asp> 27 Beauman, Jeffrey. Natural Gas Storage Transmission and Distribution. June 5, 2012. FM Global. 28 Beauman, Jeffrey. Natural Gas Storage Transmission and Distribution. June 5, 2012. FM Global. 29 Beauman, Jeffrey. Natural Gas Storage Transmission and Distribution. June 5, 2012. FM Global. 30 Natural Gas, <http://naturalgas.org/naturalgas/distribution.asp> 31 Beauman, Jeffrey. Natural Gas Storage Transmission and Distribution. June 5, 2012. FM Global. 32 Beauman, Jeffrey. Natural Gas Storage Transmission and Distribution. June 5, 2012. FM Global. 33 Beauman, Jeffrey. Natural Gas Storage Transmission and Distribution. June 5, 2012. FM Global. 7
facilities 34. Gas distribution includes branches and extensions that distribute to homes and businesses. Lastly, the power generation process occurs when the gas combusts to generate energy. Hydraulic Fracturing- Fracking Hydraulic fracturing, also known as fracking, is a well stimulation process used to maximize the extraction of underground resources including oil, natural gas, geothermal energy, and water. The oil and gas industry uses hydraulic fracturing to enhance subsurface fracture systems to allow oil or natural gas to move more freely from the rock pores to production wells that bring the oil or gas to the surface 35. The fracking process takes pressurized liquid, known as fracturing fluid, and introduces it to a source rock through a borehole to induce a fracture 36. Petroleum or natural gas is released through the fractures and then captured. Although the burning of natural gas is considered environmentally friendly, the process of fracking is not because of the amount of water which gets wasted. An individual well can require up to seven million gallons water during the fracking process, 30% of which is unrecoverable 37. Operations make every effort to reuse water throughout the process, however, geological conditions, such as the Marcellus, make water recovery impossible 38. Risks The extraction of energy from the ground combined with the transfer of gases is a complicated process where malfunctions have serious repercussions to a company s assets, the environment, and people s health. Distribution and power generation companies pay over $10 million in property 34 Beauman, Jeffrey. Natural Gas Storage Transmission and Distribution. June 5, 2012. FM Global. 35 Hydraulic Fracturing Background Information, <http://water.epa.gov/type/groundwater/uic/class2/hydraulicfracturing/wells_hydrowhat.cfm> 36 Hydraulic Fracturing, The Gale Encyclopedia of Science. 37 Hydraulic Fracturing, The Gale Encyclopedia of Science. 38 Hydraulic Fracturing, The Gale Encyclopedia of Science. 8
insurance premiums annually to transfer the risk of their company s assets to the insurance underwriter, however, many exposures still remain 39. Natural gas equipment is susceptible to natural hazards such as earthquakes, floods, and windstorm buildings. Emphasis on proper operational integrity is vital to the successful functioning of a natural gas plant, however, operator error and improper training of employees have been cited in natural gas explosions 40. General exposures include control equipment failure, nearby drilling, fracking, and natural gas processing. Risks involved in the transmission and distribution processes include pipe material, weld defect, pipeline impact, pipeline corrosion, black powder accumulation in pipelines, poor workmanship, fires or explosions at compressor stations, compressed liquid damage to equipment, sulfide stress cracking in equipment and pipelines, and secondary dehydration using ignitable liquids 41. Regulation The natural gas industry is constantly developing and when problems have occurred in the past, it has provokes stricter controls to be implemented across the country. Regulation policies vary from country to country, but the federal regulators in the United States include the Federal Energy Regulatory Commission (FERC), Department of Transportation (DOT), DOT Pipeline and Hazardous Materials Safety Administration (PHMSA), DOT Office of Pipeline Safety (OPS), Occupational Safety and Health Administration (OSHA), Transportation Safety Administration (TSA), and Environmental Protection Agency (EPA) 42. Energy generation has taken a toll on the environment, and with the depletion natural resources, regulations are becoming stricter, ultimately improving the efficiency of the natural gas process. Intrastate natural gas transmission pipeline systems and local distribution pipelines are regulated by state public service and public utility commissions along with many local authorities being 39 Beauman, Jeffrey. Natural Gas Storage Transmission and Distribution. June 5, 2012. FM Global. 40 Beauman, Jeffrey. Natural Gas Storage Transmission and Distribution. June 5, 2012. FM Global. 41 Beauman, Jeffrey. Natural Gas Storage Transmission and Distribution. June 5, 2012. FM Global. 42 Beauman, Jeffrey. Natural Gas Storage Transmission and Distribution. June 5, 2012. FM Global. 9
involved with land use regulation 43. With a major source of harm stemming from pipe leaks, it is vital to the sustainability of the natural gas and energy industry to successfully transfer the liquefied gas. In order to assure pipes were responsibly being monitored, the U.S. Congress passed the Pipeline Safety Improvement Act of 2002 44. These regulations are commonly known as the natural gas Transmission Integrity Management Program (TIMP), which require operators to use a risk-based approach to manage pipelines and specify how pipeline operators must identify, prioritize, assess, evaluate, repair and validate a pipeline s integrity. Statistical Analysis The statistical analysis on natural gas focuses on how demand of natural gas affects the Henry Hub spot price. The data is annual data from 1997-2011. The variables we chose to use are residential consumption, industrial consumption, commercial consumption, electricity power consumption (EPC), crude oil price (per barrel), and housing starts. Electricity power consumption represents the total amount of electricity consumed throughout the United States. The data is measured in kilowatt hours and is sourced from the US Energy Information Administration 45. The housing starts measure the number of new private owned housing units started in a given year. This includes both single unit and multi-unit developments. The data is sourced from the US Census Bureau and is the sum of unadjusted monthly data 46. The residential, industrial and commercial consumptions are all measured millions of cubic feet. The crude oil prices come from Cushing, OK WTI Spot Price FOB (Dollars per Barrel) 47. The reason why demand variables were used alone instead of using them with supply is because increase in demand drives the price up and increase in supply drives the price of the product 43 Beauman, Jeffrey. Natural Gas Storage Transmission and Distribution. June 5, 2012. FM Global. 44 Beauman, Jeffrey. Natural Gas Storage Transmission and Distribution. June 5, 2012. FM Global. 45 Natural Gas Distribution in the US, IBISWorld Industry Report 22121. <http://clients1.ibisworld.com.ezproxy.babson.edu/reports/us/bed/default.aspx?bedid=88193> 46 Housing Starts, IBISWorld Business Environment Profiles. <http://clients1.ibisworld.com.ezproxy.babson.edu/reports/us/bed/default.aspx?bedid=2598> 47 Petroleum & Other Liquids, Cushing, OK WTI Spot Price FOB. <http://www.eia.gov/dnav/pet/hist/leafhandler.ashx?n=pet&s=rwtc&f=a> 10
down. Since demand and supply are inversely correlated you cannot use them in a regression equation and still get a high R-Sq. Simple Regression 1997-2011 Variables R-Sq Residential Consumption 0.1% Housing Starts 2.1% Commercial Consumption 5.9% GDP 36.0% Crude Oil 38.5% Industrial Consumption 42.0% EPC 43.7% p-value>0.1 p-value<0.1 Table 1.1 Table 1.1 shows that individually, residential consumption, housing starts, and commercial consumption do now have a p-value lower than 0.1, which means that they are not significant and we reject the null hypothesis. They also have very low coefficient of determination (R-Sq) which means they are not good indicators of what future spot prices will be. The simple regressions for GDP, crude oil, industrial consumption and EPC all have p-values less than 0.1. Therefore, we reject the null hypothesis. There is sufficient evidence of a linear relationship between these variables and the spot price of natural gas. Electrical power consumption has the highest coefficient of determination, which means it is the best indicator of what future natural gas spot prices will do. This is understandable because as the economy grows more people more will be likely to spend money on utilities and industrial companies will be more be making more products and will require more natural gas. If crude oil goes up in price, more people will start to buy natural gas because it can be used as a substitute. This will drive the price up for natural gas as the demand increases. Increase in the electric power consumption means that power plants need 11
more energy in the form of natural gas to power their plants to keep up with their demand. This will also increase the demand, driving up the spot price for natural gas. Multiple Regression 1997-2011 R-Sq (adju Industrial Consumption Crude Oil Pirce 35.50% Residential Consumption Industrial Consumption Commercial Con 51.60% Industrial Consumption Commercial Consumption Residential Consumption EPC 61.30% GDP Housing Starts Residential Consumption EPC 63.50% GDP Residential Consumption Housing Starts 66.10% Commercial Consumption Industrial Consumption EPC 68.50% GDP Residential Consumption Housing Starts 68.50% Residential Consumption Industrial Consumption Commercial Consumption EPC GDP Crude Oil Price Housing Starts 90.90% Table 1.2 p-value>0.1 p-value<0.1 Average 63.24% The regression data from Table 1.2 shows us that as the number of variables in the regression equation the coefficient of determination (adjusted) also increases. The adjusted version of the coefficient of determination is similar to the normal coefficient of determination except it has been adjusted for the number of predictors in the model 48. As the number of predictor variables increase, so does the coefficient of determination. The regression equation with all three types of end user consumptions is the only equation where are the p-values are less than 0.1, but it only has and R-Squared of 51.60%. The full model has a much higher R-Squared than all the other models at 90.90%. This goes to show that there are a lot of different determinates of demand that affect the spot price of natural gas. All of the regression models has adjusted coefficient of determinations above 60% except for industrial consumption and crude oil price. As well as the consumption equation that was mentioned earlier. A reason why these functions could have low coefficient of determinations is due to multicollinearity. If 48 Interpretation: Using Minitab, <http://mtweb.mtsu.edu/stats/regression/level3/indicator/useminitabinterp.htm> 12
you look at the time series graph (Exhibit 1) of residential, industrial consumption, and commercial consumption you can see that they all follow a similar trend. That negative trend is not a good determent of the high variance in the graph of (Exhibit 2). If consumers from a specific sector increase their consumption of natural gas it causes the consumers from other sectors to increase their consumption as well. The most reasonable explanation for the high variance in Exhibit 2 is due unexpected weather, which increased the demand in natural gas or slowed down production. Conclusion The recent discovery of the natural gas in the Appalachian Basin and new technologies, such as fracking, should allow for low natural gas prices in the future with such large supply. The development of the infrastructure to move natural gas will also help lower the cost to transport natural gas. Barring any natural disasters, economic crisis, or abnormal temperatures natural gas prices should be fairly stable in the long run due to regulation and technological developments. Our regressions showed us that there are many demand factors that affect the Henry Hub spot price of natural gas. We have learned that although the largest consumers for natural gas are the industrial consumers, they do not have a big impact on the market because the price they pay is so low. Residential customers have the highest impact on the market because they pay the most for natural gas. Due to the residential customers, weather has a significant impact on the short term demand because natural gas is an efficient energy source for heating many commercial buildings and homes in the United States in the winter months. 13
References Beauman, Jeffrey. Natural Gas Storage Transmission and Distribution. June 5, 2012. FM Global. Boorse, Dorothy, Wright, Richard. Environmental Science, San Francisco, CA, 2011. Kendall, James. Global Gas Outlook, <http://www.eia.gov/pub/oil_gas/natural_gas/presentations/2008/globalgas/globalgas.html> Loh, Bohan. Japan LNG demand has likely peaked, <http://www.reuters.com/article/2012/10/23/us-japan-meti-lng-idusbre89m08720121023> About U.S. Natural Gas Pipelines, <http://www.eia.gov/pub/oil_gas/natural_gas/analysis_publications/ngpipeline/index.html> America s Natural Gas Highway, <http://www.cleanenergyfuels.com/video/anghvideo.html> Background, <http://www.naturalgas.org/overview/background.asp> Factors Affecting Gas Prices, <http://www.eia.gov/energyexplained/index.cfm?page=natural_gas_factors_affecting_prices> Housing Starts, IBISWorld Business Environment Profiles. <http://clients1.ibisworld.com.ezproxy.babson.edu/reports/us/bed/default.aspx?bedid=2598> Hydraulic Fracturing, The Gale Encyclopedia of Science. <http://ic.galegroup.com.ezproxy.babson.edu/ic/scic/referencedetailspage/referencedetailswi ndow?failovertype=&query=&windowstate=normal&contentmodules=&mode=view&displaygr oupname=reference&limiter=&currpage=&disablehighlighting=true&source=&sortby=&display Groups=&search_within_results=&action=e&catId=GALE OWJADB227635689&activityType=&s canid=&documentid=gale CV2644042543&> Hydraulic Fracturing Background Information, <http://water.epa.gov/type/groundwater/uic/class2/hydraulicfracturing/wells_hydrowhat.cfm> Interpretation: Using Minitab, <http://mtweb.mtsu.edu/stats/regression/level3/indicator/useminitabinterp.htm> National Data, Bureau of Economic Analysis. <http://www.bea.gov/itable/itable.cfm?reqid=9&step=1> Natural Gas, <http://naturalgas.org/naturalgas/naturalgas.asp> Natural Gas, Henry Hub Gulf Coast Natural Gas Spot Price. <http://www.eia.gov/dnav/ng/hist/rngwhhda.htm> Natural Gas, Natural Gas Consumption by End Use. <http://www.eia.gov/dnav/ng/ng_cons_sum_dcu_nus_a.htm> Natural Gas Demand, <http://www.naturalgas.org/business/demand.asp> 14
Natural Gas Supply, <http://www.naturalgas.org/business/analysis.asp> Natural Gas Distribution, <http://naturalgas.org/naturalgas/distribution.asp> Natural Gas Distribution in the US, IBISWorld Industry Report 22121. <http://clients1.ibisworld.com.ezproxy.babson.edu/reports/us/bed/default.aspx?bedid=88193> Petroleum & Other Liquids, Cushing, OK WTI Spot Price FOB. <http://www.eia.gov/dnav/pet/hist/leafhandler.ashx?n=pet&s=rwtc&f=a> Capital IQ; NetAdvantage, Industry Surveys: Natural Gas Distribution, July 12, 2012. The authors of this paper hereby give permission to Professor Michael Goldstein to distribute this paper by hard copy, to put it on reserve at Horn Library at Babson College, or to post a PDF version of this paper on the internet. I pledge my honor that I have neither received nor provided any unauthorized assistance during the completion of this work. 15
Regression Analysis: Spot Price versus Residential Consumption Spot Price = 3.3 + 0.000000 Residential Consumption Constant 3.32 15.93 0.21 0.838 Residential Consumption 0.00000033 0.00000332 0.10 0.923 S = 2.23797 R-Sq = 0.1% R-Sq(adj) = 0.0% Regression 1 0.049 0.049 0.01 0.923 Residual Error 13 65.111 5.009 Total 14 65.160 Unusual Observations Residential Spot Obs Consumption Price Fit SE Fit Residual St Resid 10 4368466 6.730 4.754 1.532 1.976 1.21 X X denotes an observation whose X value gives it large leverage. Regression Analysis: Spot Price versus Industrial Consumption Spot Price = 18.3-0.000002 Industrial Consumption Constant 18.304 4.392 4.17 0.001 Industrial Consumption -0.00000186 0.00000061-3.07 0.009 S = 1.70488 R-Sq = 42.0% R-Sq(adj) = 37.5% Regression 1 27.374 27.374 9.42 0.009 Residual Error 13 37.786 2.907 Total 14 65.160 Regression Analysis: Spot Price versus Commercial Consumption Spot Price = 21.0-0.000005 Commercial Consumption Constant 20.97 17.77 1.18 0.259 Commercial Consumption -0.00000521 0.00000575-0.91 0.382 16
S = 2.17146 R-Sq = 5.9% R-Sq(adj) = 0.0% Regression 1 3.862 3.862 0.82 0.382 Residual Error 13 61.298 4.715 Total 14 65.160 Unusual Observations Commercial Spot Obs Consumption Price Fit SE Fit Residual St Resid 10 2832030 6.730 6.220 1.568 0.510 0.34 X 12 3152529 8.860 4.551 0.677 4.309 2.09R R denotes an observation with a large standardized residual. X denotes an observation whose X value gives it large leverage. Regression Analysis: Spot Price versus EPC Spot Price = - 22.2 + 0.00733 EPC Constant -22.158 8.521-2.60 0.022 EPC 0.007330 0.002306 3.18 0.007 S = 1.67934 R-Sq = 43.7% R-Sq(adj) = 39.4% Regression 1 28.497 28.497 10.10 0.007 Residual Error 13 36.662 2.820 Total 14 65.160 Unusual Observations Spot Obs EPC Price Fit SE Fit Residual St Resid 15 3998 4.000 7.148 0.831-3.148-2.16R R denotes an observation with a large standardized residual. Regression Analysis: Spot Price versus GDP Spot Price = - 1.82 + 0.000566 GDP 17
Constant -1.818 2.524-0.72 0.484 GDP 0.0005662 0.0002093 2.71 0.018 S = 1.79077 R-Sq = 36.0% R-Sq(adj) = 31.1% Regression 1 23.471 23.471 7.32 0.018 Residual Error 13 41.689 3.207 Total 14 65.160 Regression Analysis: Spot Price versus Crude Oil Price Spot Price = 2.57 + 0.0471 Crude Oil Price Constant 2.5716 0.9327 2.76 0.016 Crude Oil Price 0.04706 0.01651 2.85 0.014 S = 1.75629 R-Sq = 38.5% R-Sq(adj) = 33.7% Regression 1 25.061 25.061 8.12 0.014 Residual Error 13 40.099 3.085 Total 14 65.160 Unusual Observations Crude Oil Spot Obs Price Price Fit SE Fit Residual St Resid 9 56.6 8.690 5.237 0.469 3.453 2.04R R denotes an observation with a large standardized residual. Regression Analysis: Spot Price versus Housing Starts Spot Price = 4.04 + 0.00059 Housing Starts Constant 4.037 1.727 2.34 0.036 Housing Starts 0.000591 0.001123 0.53 0.607 S = 2.21532 R-Sq = 2.1% R-Sq(adj) = 0.0% 18
Regression 1 1.360 1.360 0.28 0.607 Residual Error 13 63.799 4.908 Total 14 65.160 Unusual Observations Housing Spot Obs Starts Price Fit SE Fit Residual St Resid 12 905 8.860 4.572 0.838 4.288 2.09R R denotes an observation with a large standardized residual. Regression Analysis: Spot Price versus Residential, Industrial C,... Spot Price = 22.4 + 0.000010 Residential Consumption - 0.000002 Industrial Consumption - 0.000018 Commercial Consumption Constant 22.38 12.30 1.82 0.096 Residential Consumption 0.00001032 0.00000435 2.37 0.037 Industrial Consumption -0.00000175 0.00000055-3.17 0.009 Commercial Consumption -0.00001761 0.00000789-2.23 0.047 S = 1.50023 R-Sq = 62.0% R-Sq(adj) = 51.6% Regression 3 40.402 13.467 5.98 0.011 Residual Error 11 24.758 2.251 Total 14 65.160 Source DF Seq SS Residential Consumption 1 0.049 Industrial Consumption 1 29.126 Commercial Consumption 1 11.227 Unusual Observations Residential Spot Obs Consumption Price Fit SE Fit Residual St Resid 12 4892277 8.860 5.677 0.598 3.183 2.31R R denotes an observation with a large standardized residual. Regression Analysis: Spot Price versus Crude Oil Price, Industrial Consu Spot Price = 12.6 + 0.0211 Crude Oil Price - 0.000001 Industrial Consumption 19
Constant 12.617 8.686 1.45 0.172 Crude Oil Price 0.02109 0.02764 0.76 0.460 Industrial Consumption -0.00000122 0.00000104-1.16 0.267 S = 1.73293 R-Sq = 44.7% R-Sq(adj) = 35.5% Regression 2 29.123 14.562 4.85 0.029 Residual Error 12 36.037 3.003 Total 14 65.160 Source DF Seq SS Crude Oil Price 1 25.061 Industrial Consumption 1 4.062 Regression Analysis: Spot Price versus GDP, Housing Star, Residential Spot Price = - 23.7 + 0.000982 GDP + 0.00292 Housing Starts + 0.000003 Residential Consumption Constant -23.71 10.20-2.32 0.040 GDP 0.0009820 0.0001810 5.43 0.000 Housing Starts 0.0029179 0.0007683 3.80 0.003 Residential Consumption 0.00000265 0.00000191 1.39 0.193 S = 1.25575 R-Sq = 73.4% R-Sq(adj) = 66.1% Regression 3 47.814 15.938 10.11 0.002 Residual Error 11 17.346 1.577 Total 14 65.160 Source DF Seq SS GDP 1 23.471 Housing Starts 1 21.315 Residential Consumption 1 3.028 Unusual Observations Spot Obs GDP Price Fit SE Fit Residual St Resid 12 14292 8.860 5.952 0.570 2.908 2.60R R denotes an observation with a large standardized residual. 20
Regression Analysis: Spot Price versus GDP, Housing Starts,... Spot Price = - 16.2 + 0.00129 GDP + 0.00317 Housing Starts + 0.000003 Residential Consumption - 0.00333 EPC Constant -16.23 19.58-0.83 0.427 GDP 0.0012875 0.0006990 1.84 0.095 Housing Starts 0.0031749 0.0009783 3.25 0.009 Residential Consumption 0.00000282 0.00000202 1.40 0.193 EPC -0.003326 0.007329-0.45 0.660 S = 1.30368 R-Sq = 73.9% R-Sq(adj) = 63.5% Regression 4 48.164 12.041 7.08 0.006 Residual Error 10 16.996 1.700 Total 14 65.160 Source DF Seq SS GDP 1 23.471 Housing Starts 1 21.315 Residential Consumption 1 3.028 EPC 1 0.350 Unusual Observations Spot Obs GDP Price Fit SE Fit Residual St Resid 12 14292 8.860 5.992 0.598 2.868 2.48R R denotes an observation with a large standardized residual. Regression Analysis: Spot Price versus Residential, Industrial C,... Spot Price = 22.2 + 0.000006 Residential Consumption - 0.000002 Industrial Consumption - 0.000008 Commercial Consumption - 0.00299 EPC - 0.00079 GDP + 0.129 Crude Oil Price + 0.00289 Housing Starts Constant 22.19 13.97 1.59 0.156 Residential Consumption 0.00000624 0.00000395 1.58 0.158 Industrial Consumption -0.00000163 0.00000166-0.98 0.358 Commercial Consumption -0.00000831 0.00000789-1.05 0.327 EPC -0.002986 0.004877-0.61 0.560 GDP -0.000790 0.001100-0.72 0.496 Crude Oil Price 0.12906 0.03028 4.26 0.004 Housing Starts 0.0028924 0.0007544 3.83 0.006 21
S = 0.651351 R-Sq = 95.4% R-Sq(adj) = 90.9% Regression 7 62.1900 8.8843 20.94 0.000 Residual Error 7 2.9698 0.4243 Total 14 65.1598 Source DF Seq SS Residential Consumption 1 0.0493 Industrial Consumption 1 29.1262 Commercial Consumption 1 11.2266 EPC 1 6.7633 GDP 1 3.1234 Crude Oil Price 1 5.6648 Housing Starts 1 6.2363 Regression Analysis: Spot Price versus Industrial C, Commercial C,... Spot Price = - 14.9 + 0.000000 Industrial Consumption - 0.000022 Commercial Consumption + 0.000012 Residential Consumption + 0.00764 EPC Constant -14.92 22.16-0.67 0.516 Industrial Consumption 0.00000003 0.00000104 0.03 0.980 Commercial Consumption -0.00002163 0.00000735-2.94 0.015 Residential Consumption 0.00001213 0.00000400 3.03 0.013 EPC 0.007641 0.003941 1.94 0.081 S = 1.34143 R-Sq = 72.4% R-Sq(adj) = 61.3% Regression 4 47.165 11.791 6.55 0.007 Residual Error 10 17.994 1.799 Total 14 65.160 Source DF Seq SS Industrial Consumption 1 27.374 Commercial Consumption 1 0.368 Residential Consumption 1 12.660 EPC 1 6.763 Unusual Observations Industrial Spot Obs Consumption Price Fit SE Fit Residual St Resid 12 6670182 8.860 5.954 0.554 2.906 2.38R R denotes an observation with a large standardized residual. 22
Data Natural Gas Exhibit 1 Time Series Plot of Residential, Industrial C, Commercial C 9000000 8000000 Variable Residential Consumption Industrial Consumption Commercial Consumption 7000000 6000000 5000000 4000000 3000000 1 2 3 4 5 6 7 8 9 Index 10 11 12 13 14 15 23
Spot Price Natural Gas Exhibit 2 Time Series Plot of Spot Price 9 8 7 6 5 4 3 2 1 2 3 4 5 6 7 8 Index 9 10 11 12 13 14 15 24