1 The Economic Impacts of the Proposed Natural Gas Severance Tax in Pennsylvania by Timothy J. Considine, Ph.D. Natural Resource Economics, Inc. April 22, 2015 This study was funded under a consulting agreement between Natural Resource Economics, Inc. and the American Petroleum Institute. The author is also a Distinguished Professor of Energy Economics at the University of Wyoming. The analysis, findings, and conclusions in this paper are solely those of the author and do not necessarily reflect those of the American Petroleum Institute or the University of Wyoming.
2 Key Findings In March 2015, Pennsylvania Governor Tom Wolf formally proposed a natural gas severance tax of 5% on the gross market value of production plus a fixed fee of 4.7 cents per thousand cubic feet produced. This tax would repeal the existing natural gas impact fee. This study conducts an economic impact analysis of this proposed tax. The estimated impacts are as follows: 1) The proposed severance tax is estimated to reduce the number of wells drilled by 1,364 from 2016 to 2025 leading to a cumulative drilling investment loss of $11.5 billion. With fewer wells drilled cumulative losses in natural gas and liquids production amount to $11.2 billion over the same period. 2) These investment and production losses lead to cumulative losses in value added or gross state product to the Pennsylvania economy of over $20 billion from 2016 to ) The proposed severance tax is estimated to reduce supported employment in the state by nearly 18,000 by 2025 relative to projected levels without the tax. The relatively high paying construction and oil and gas sectors would be hardest hit. 4) The severance tax revenues will be offset by the elimination of the impact fee and lower sales and income tax receipts from reduced economic activity resulting from the severance tax. Net revenue to the state of Pennsylvania in 2016 is estimated to be $515 million, 1/3 less than the $765 million projected by the Governor s office. This study projects that net revenue to the state is estimated at $492 million in 2020 and $565 million in 2025, significantly less than the $1 billion per year estimated by the tax supporters. 5) The proposed severance tax is estimated to reduce Pennsylvania s natural gas output by over 900 million cubic feet per day by as early The cumulative loss in natural gas production is 2.86 trillion cubic feet from 2016 to ) The proposed severance tax will transfer hundreds of millions per year from one of Pennsylvania s most productive sectors to the government. 7) The proposed price floor for calculating the tax will increase the burden of the severance tax when natural gas prices are low, which are times when the industry is least capable of absorbing a cost increase. These estimates are conservative, based upon projections of declining well productivity. If the productivity if Marcellus and Utica wells holds steady or increases, the economic losses from the severance could be much larger. ii
3 Table of Contents Key Findings... ii List of Figures... iv List of Tables... iv 1. Introduction Proposed Severance Tax Pennsylvania Oil & Gas Production Analysis of Drilling and Production Baseline Forecast Impacts on Oil & Gas Industry Economic Impacts Sensitivity Analysis Summary and Conclusions References Appendix A: Analysis of Price Impact of Severance Tax iii
4 List of Figures Figure 1: Production profiles of horizontal wells... 7 Figure 2: Real price projections for natural gas and propane Figure 3: Number of well spuds and producing wells Figure 4: Projected production decline curves Figure 5: Baseline drilling and production forecast Figure 6: Impacts on drilling and production Figure 7: Impacts on gross output of drilling and extraction Figure 8: Summary of Tax Impacts Figure 9: Losses in drilling and production due to the severance tax Figure 10: Losses in employment and value added due to the severance tax List of Tables Table 1: Oil and Gas Production, Table 2: Well spuds... 5 Table 3: Drilling model parameter estimates... 9 Table 4: Summary of economic impacts Table 5: Direct, indirect, and induced impacts on real gross output by sector Table 6: Total impacts on employment and value added by sector Table 7: Impacts on state and local tax revenues iv
5 1. Introduction Energy reserves offer an appealing target for governments to collect tax revenues, presumably without inducing companies to shift activities to other jurisdictions. Once an oil or gas well is constructed it cannot be moved across state or local boundaries. Economic reality, however, is never that simple. This is particularly true for natural gas and oil production in the 21 st century. In conventional fields, oil and gas is pushed to the surface by natural pressures, either from gas or water. Many of these fields are being depleted around the world and now require enhanced production techniques to produce additional oil and gas. For nearly a century, substantial reserves of oil and gas have been known to exist outside these conventional fields in so-called unconventional or tight formations. Over the past twenty years, several technological innovations, such as hydraulic fracturing, directional drilling, and advanced seismic technologies have enabled cost effective production from these unconventional reservoirs. As a result, natural gas production has increased sharply from many new regions within the United States, such as the Marcellus and Utica in Pennsylvania, the Barnett in Texas, the Fayetteville in Arkansas, and the Haynesville in Louisiana, among many others. These technologies also increase the flexibility of oil and gas production, allowing producers to quickly produce more oil and gas as prices rise. If costs rise in one area, however, producers shift drilling equipment to other regions to pursue relatively more profitable opportunities. Unlike conventional fields in which production rises for several years and then gradually declines, unconventional wells have high initial production rates but steep declines in production thereafter. Hence, to maintain production from an
6 Pennsylvania page 2 unconventional field, producers must improve their recovery techniques or drill more wells. Given these features of unconventional natural gas production, fewer wells drilled due to greater taxation reduces oil and gas production over time and, given the steep production decline curves for unconventional production, these production losses mount over time. Lower investment in drilling reduces economic activity associated with the construction and completion of wells. Fewer jobs are supported throughout the supply chain for building natural gas industry infrastructure, such as steel for casing pipe, aggregate materials for access roads, legal and engineering services, and many other supplies. In addition, fewer wells reduce natural gas production, which has similar ripple effects through the local economy. While state and local governments may receive higher tax revenues, some of these gains are offset by lost income and sales taxes from reduced economic activity caused by the tax. Hence, the severance tax on natural gas and oil recently proposed by Pennsylvania Governor Tom Wolf must be evaluated with these economic realities in mind. Accordingly, the objective of this study is to estimate how this proposed severance tax would affect drilling investment, oil and gas production, and economic and fiscal conditions in the Commonwealth of Pennsylvania. The next section briefly summarizes the key provisions of the proposed legislation. Section three provides an overview of recent trends in oil and gas production in Pennsylvania. The fourth part examines the economic forces driving oil and gas investment and production. With this analytical framework, a baseline forecast for drilling and production over the next decade is then discussed. The following two sections provide an analysis of the impacts of the proposed severance tax on the oil and
7 Severance Tax page 3 gas industry, the economy, and tax revenues for the Commonwealth of Pennsylvania. The study concludes with a summary of the key findings and a discussion of the economic trade-offs that should be considered in the severance tax debate. 2. Proposed Severance Tax In his Executive Budget, Governor Wolf proposed legislation enacting the Pennsylvania Education Reinvestment Act (PERA), which includes a severance tax on natural gas and related liquids produced within the Commonwealth. The severance tax is 5% on the gross value of production plus 4.7 cents per thousand cubic feet of volume extracted, which adds roughly another percentage point of tax, making the government take on gross revenues of approximately 6%. The legislation also stipulates a minimum gross value of $20 per barrel for natural gas liquids and $2.97 per thousand cubic feet for natural gas. This provision has the effect of establishing a minimum price for taxation purposes. If market prices for natural gas fall below $2.97 per thousand cubic feet, then producers must pay the severance tax on this price not the market price. Hence, if market prices fall below $2.97 the tax rate is above the combined rate of 5% plus the 4.7 cents per thousand cubic feet. For example, if natural gas prices are $2.26 per thousand cubic feet, the government take on gross revenues increases to 8.5%. This study assumes that operators of gas wells pay this tax after royalties are paid to holders of mineral property rights. In addition, if an effective tax rate is calculated on net revenues after expenses, the calculated rate will be higher Pennsylvania currently imposes an impact fee on every unconventional gas well drilled within the state. This fee has generated more than $630 million between 2011 and Under Act 13, which implemented the impact fee, if a severance tax is enacted, the
8 Pennsylvania page 4 impact fee would be automatically repealed. Hence, this study assumes the impact fee would cease upon imposition of the severance tax. In lieu of the impact fee, PERA has provisions to allocate a portion of the severance tax revenues to local communities. Since the amount and especially the timing of disbursements are not clear at this time, this study does not consider how severance tax revenues are re-spent. 3. Pennsylvania Oil & Gas Production The oil industry started in Pennsylvania in 1859 and oil and natural gas production has continued ever since. Only recently, however, has Pennsylvania reemerged as a major producer. In 2011 Pennsylvania produced 3.66 billion cubic feet (BCF) of natural gas per day and 9.49 thousand barrels of crude oil and lease condensate per day (see Table 1). In 2014 these production levels were up considerably with natural gas production more than tripling to BCF per day and liquids production rising to over 17 thousand barrels per day (see Table 1). All of this increase is attributed to unconventional production from the Marcellus, Upper Devonian, and Utica shale gas plays. Conventional oil and natural gas production declined over the same period. Significant advances in technology enhancing well productivity and a sharp increase in the number of unconventional wells drilled have contributed to this production success. Between 2006 and 2014, 7,704 unconventional wells were drilled in Pennsylvania with 7,074 horizontal wells and 630 vertical wells. Horizontal wells cost on average roughly $8 million to drill and complete, which is considerably more than the cost of drilling vertical wells, but of course, horizontal wells have much higher output than vertical wells. Over the same period, 23,028 conventional wells were drilled but the pace of conventional drilling slowed dramatically from 4,683 wells drilled during 2006 to
9 Severance Tax page wells in These wells on average cost approximately $230,000 to drill and complete. So far during the first 10 weeks of 2015, unconventional drilling is down 35% from the same period in 2014 while conventional drilling is down 53% most likely the result of much lower prices for oil and natural gas (see Table 2). Table 1: Oil and Gas Production, Conventional Unconventional Total Natural Gas in BCF / day Oil & Condensate in MBBL / day BCF = Billion cubic feet, MBBL = thousand barrels Source: Pennsylvania Department of Environmental Protection (PADEP) Table 2: Well spuds Unconventional Period Conventional Horizontal Vertical Total , , , , ,054 1, , ,571 1, , ,174 1, , ,016 1, , , ,309 Cumulative 23,028 7, ,704 Jan-March 15, Jan-March 15, Percentage Change -53% -33% -100% -35%
10 Pennsylvania page 6 The productivity of unconventional horizontal wells has been increasing sharply in recent years. One way to measure productivity is to compare average production decline curves for wells that go into production over time. For instance, in 2009 there were 137 unconventional horizontal wells that went into production with average annual initial production of 242 million cubic feet. During 2014 there were 1,121 horizontal wells that went into production with initial production that averaged 902 million cubic feet. So the class or vintage of wells from 2014 was more than 3.7 times more productive during the first year than wells drilled during Initial production and output in subsequent years are plotted for vintages of wells from 2008 through 2014 in Figure 1. Notice that output during the second year is considerably higher than initial production. On average second year production is 40% percent higher than initial production for vintages of wells between 2009 and After the second year, production declines on average 34% during the third year, 31% during the fourth, 22% during the fifth, 10% during the sixth year, and 9% during the seventh year after initial production. As the diagram illustrates, the production decline curve for horizontal Marcellus wells have been shifting up over time. Higher productivity and a larger number of wells drilled translate to the considerably higher natural gas production reported in Table 1. Hence, projecting future natural gas production involves predicting the future number of wells drilled, future decline rates for wells already in production, and how average production decline curves for future wells will adjust over time. The models used to develop these forecasts are discussed in the following section.
11 Severance Tax page 7 Figure 1: Production profiles of horizontal wells 4. Analysis of Drilling and Production As the previous discussion demonstrates, the number of producing wells and their productivity determine oil and gas production. By reducing the financial returns from drilling, a severance tax reduces drilling activity and the number of producing wells, which diminishes the capacity of the industry to produce oil and gas both now and in the future. To estimate the magnitude of these adjustments, the response of drilling activity to financial returns must be estimated and the link between drilling, producing wells, and actual production must be forged. To quantify how drilling is affected by financial returns, this study estimates an economic model of drilling activity similar to previous studies by Epple (1985) and Deaton et al. (1990). Like this previous literature, the model developed here hypothesizes
12 Pennsylvania page 8 that the number of wells drilled is a function of prices for natural gas and petroleum liquids. Wells drilled should increase (decrease) with higher (lower) prices. Following Deaton et al. (1990) drilling costs are excluded from the model because drilling costs can rise and fall when the demand for drilling equipment shifts. In effect, drilling costs are determined by crude oil prices so that by including costs in the model the response of drilling to price would be over-stated, according to Deaton et al. (1990). By excluding drilling costs, the price terms are allowed to capture any impact that crude oil prices exert the costs of drilling in the short run. In applying this model to the Pennsylvania Marcellus, an additional variable should be included. Mineral leases typically include a clause that drilling should commence within a few years of signing the agreement and if drilling does not take place, the lease is freed for other bidders. As a result, many companies drill wells to hold leases. To capture this effect, this study uses acreage under lease as an explanatory variable. As the number of acres under lease increases, drilling should increase. This model is estimated using survey data collected by Considine (2009, 2010, and 2011). The surveys provide 84 quarterly observations that capture the variation in costs and drilling activity for the top seven firms operating in the Pennsylvania Marcellus. The model is estimated in log linear form so that the estimated coefficients are elasticities. Four variations of the model are estimated, fixed and random effects, pooled, and means. The model estimated pooling all the observations using ordinary least squares is the preferred model based upon the Schwarz Bayesian Information Criteria. The econometric results for this model appear in Table 3 below. All of the estimated elasticities have signs consistent with economic theory and intuition. Moreover,
13 Severance Tax page 9 their magnitudes seem reasonable. A 10 percent increase in prices for natural gas liquids (ethane, propane, butane, iso-butane, and natural gasoline) leads to an 8.6 percent increase in drilling. The elasticity for natural gas is 0.68, which is similar in magnitude, but the probability value indicates 31 percent chance this coefficient is insignificantly different from zero. The estimated coefficient on undrilled acreage is highly significant, indicating that the number of wells drilled increase 7.8 percent for every 10 percent increase in acreage under lease. The R-squared coefficient for this regression is 0.72 indicating that 72% of the variation in drilling activity observed for the seven firms over this three year period is associated with variation in the explanatory variables. Based upon the R-squared coefficient and the elasticity estimates, the model appears to provide a reasonable basis for estimating how a severance tax would affect drilling activity in Pennsylvania. Table 3: Drilling model parameter estimates Explanatory Factor Estimated Coefficient Standard Error t-statistic Probability Value Intercept Natural Gas Price Prices for NGLs Acres under lease The production decline curves discussed above allow a basis for predicting future production from wells drilled prior to The average decline rates cited above are used to extrapolate decline curves from 2008 through 2014 into the future. For example, the second year production rate for wells that went into production during 2014, which is 2015 production from these 2014 wells, is based upon the average second year rate of change of 40%. Decline rates for subsequent years are also based upon the averages
14 Pennsylvania page 10 observed over the sample period Decline rates beyond 7 years are assumed to be 5%. For vintages of wells drilled from 2015 to 2025, the initial and second year production levels for the average decline curves are adjusted to reflect diminishing marginal productivity of wells because the number of high output areas or sweet spots diminishes as development of the field matures. The details of these productivity changes are discussed below in the following section. 5. Baseline Forecast To provide a basis of comparison for analysis of the impacts of the severance tax, a baseline forecast is required. The drilling and production models developed here provide a means to generate such a forecast given projections for prices of natural gas and related liquids and future acres under lease. Forecasts of oil and gas prices from the U.S. Energy Information Administration are used to develop the baseline forecast. The price projections used in this study are based upon the March 10, 2015 Short Term Energy Outlook for 2015 and 2016 and the Annual Energy Outlook (EIA, 2014) for years beyond These price projections are plotted in Figure 2. The natural gas price represents the prevailing market price in the Northeast United States, which shows a substantial discount relative to spot prices for natural gas at the Henry Hub trading location on the U.S. Gulf Coast. Prices for propane are a proxy used to represent prices for all natural gas liquids because propane prices were used in the econometric estimation of the drilling model. Prices for other natural gas liquids, such as ethane, butane, and natural gasoline, were unavailable for this analysis.
15 Severance Tax page 11 Figure 2: Real price projections for natural gas and propane The sharp drop in oil prices during late 2014 and early 2015 are reflected in Figure 2. Prices for propane after 2016 are based upon the low oil price scenario from the EIA Annual Energy Outlook for This approach provides a rather conservative basis for projecting future drilling activity in the Marcellus. The last assumption required for the drilling forecast involves future acres under lease. Since most of the major companies operating in the Marcellus have already purchased leases in their core production areas, this study assumes acreage under lease will remain constant over the forecast horizon. Like the price projections, this assumption is conservative and acts to limit the projection of future drilling activity. How many wells drilled determines the number of producing wells. During the early years of Marcellus development, the number of wells drilled exceeded the number of wells going into production due to leasing incentives to drill wells to hold production and from delays in constructing pipelines and other related gas infrastructure (see Figure 3). Recently, however, the number of new producing wells seems to be tracking drilling
16 Pennsylvania page 12 activity (see Figure 3). Consequently, this study assumes the number of producing wells during the baseline forecast from 2015 to 2025 moves with predicted drilling activity. Figure 3: Number of well spuds and producing wells Together with wells that went into production in previous years and the associated production decline curves, future production can be predicted based upon projected production decline curves. So as to not over-state the economic impacts of the severance tax, this study assumes that Marcellus well productivity peaks in 2015 and steadily declines thereafter to the end of the forecast horizon Initial production is assumed to decline 8.8 percent per year and the growth in second year output from initial production is assumed to gradually decline from 30% to 3%. These assumptions result in production forecasts that are roughly in line with projections made by the US Energy Information Administration for Marcellus production. An illustration of the projected decline curves appears in Figure 4.
17 Severance Tax page 13 Figure 4: Projected production decline curves The baseline forecast of drilling and production from horizontal wells appears in Figure 5. The number of wells drilled declines sharply during 2015 but then recovers after 2016 as real prices for natural gas increase. Production growth slows during 2015 and 2016 and peaking just after 2020, reaching 14.6 million cubic feet of gas equivalents. Production of gas and liquids then slightly declines after 2021, essentially hitting a plateau during the early 2020s. Overall, total gas and liquids production is slightly more than 20 percent higher in 2025 than in Figure 5: Baseline drilling and production forecast
18 Pennsylvania page Impacts on Oil & Gas Industry As mentioned above, the proposed severance tax has three main components, a fixed 5 percent rate, a fixed charge of 4.7 cents per thousand cubic feet, and a minimum price. These provisions interact to reduce the net value of natural gas and propane production that in turn affects drilling decisions. In 2016 natural gas prices are projected to fall below the minimum gross value of $2.97 per thousand cubic feet stipulated in the proposed legislation. Hence, the burden of the severance tax increases when the natural gas price falls below this minimum. The fixed charge of 4.7 cents per thousand cubic feet is from 1.5% to 0.9% of market prices. This component raises a similar amount of revenue as the current impact fee. If royalties to the mineral owner, gathering, transportation, and other production expenses are taken into account, the combined tax will be higher. These severance taxes reduce investments in drilling and, as a consequence, lead to fewer wells in production and escalating losses in production over time. For example, if the severance tax is enacted in 2016, there are 113 fewer wells drilled than under the baseline forecast. Since the baseline forecast trajectory of drilling is up after 2016, the lost number of wells drilled increases to 158 in 2025 compared to the baseline projection (see Figure 6). Under a severance tax, the number of producing wells drops each year as the number of wells drilled are reduced relative to the baseline forecast. As a result, production losses mount over time. For instance, production is 0.24 billion cubic feet per day lower than in the baseline forecast in 2016 and 0.93 billion cubic feet per day lower in Note that these losses are not cumulative but are annual production losses. By discouraging investment in new wells, the severance tax reduces production of natural
19 Severance Tax page 15 gas and associated liquids. The cumulative loss in natural gas production is 2.86 trillion cubic feet from 2016 to Figure 6: Impacts on drilling and production These impacts in terms of fewer wells drilled and lower physical output can be translated to monetary terms. The lost number of wells drilled are multiplied by the average per well cost of drilling and completion to estimate the reduction in the gross value of output for the oil and natural gas drilling sector. The proposed severance tax is estimated to reduce the amount spent on drilling in the state by $900 million in 2016 rising to $1.4 billion by 2025.
20 Pennsylvania page 16 Likewise, the value of resulting losses in natural gas and liquids production is estimated by multiplying the losses in physical volumes by market prices. The results of these calculations are summarized in Figure 8. Notice that most of the $1.1 billion loss in total real gross output in 2016 arises from reduced drilling expenditures of $0.9 billion while only $0.2 billion is from lower gross output of gas and liquids extraction. As Figure 7 illustrates, by 2025 more than half the total loss in real gross output is from lower volumes of natural gas and liquids. Figure 7: Impacts on gross output of drilling and extraction
21 Severance Tax page Economic Impacts These reductions in real gross output from the severance tax will reduce production in other sectors of the economy. While the drilling rig may be the most widely associated symbol of natural gas development, there are many economic activities before and after drilling that generate significant economic impacts. Many people are required to identify lease properties, write leases, and conduct related legal and regulatory work. Seismic surveys also require manpower, local business services, and other provisions. Once a prospective site is identified, site preparation and drilling begins and a range of services, labor, and other locally supplied inputs are required. Once a natural gas well is completed, infrastructure, such as well production equipment and pipelines are installed, which again stimulates local business activity. Finally, as production flows from the well, royalties are paid to landowners. These expenditures stimulate the local economy and provide resources for community services, such as health care, education, and charities. The initial purchases of goods and services made by natural gas companies during well construction and production are known as direct economic impacts. The firms supplying these goods and services in turn purchase supplies from their vendors. These impacts are known as indirect economic impacts. For example, in developing mineral leases natural gas drilling companies employ the services of land management companies that in turn purchase goods and services from other businesses. The wages earned by these employees increase household incomes, which then stimulates spending on local goods and services. These impacts associated with household spending are called induced impacts. The total economic impacts are the sum of the direct, indirect, and induced spending. Comparing value added, tax revenues, and employment in the local economy with and without the severance tax provides an estimate of these economic impacts.
22 Pennsylvania page 18 To estimate these impacts, this study employs the input-output model of the Pennsylvania economy developed by the Minnesota IMPLAN Group, Inc. developed from data collected by the Bureau of Economic Analysis in the US Department of Commerce. 1 Based upon detailed accounting data collected by Considine et al. ( ) this study assumes that 82 percent of spending for vendors supplying goods and services to drilling operations is paid to Pennsylvania companies. Hence, the net changes in real gross output for the drilling and extraction sectors due to the severance tax are equal to the change in real gross output in Figure 8 multiplied by Changes in royalty income, which is a component of other property income within the IMPLAN model, are captured by these changes in gross output. The changes in Pennsylvania real gross output after deducting spending on goods and services provided by businesses outside of Pennsylvania appear Table 4. The direct change in real gross output is -$939 million in 2016, which is 82% of the -$1,145.4 million presented in Figure 8. These losses occur in two sectors: drilling and extraction of oil and gas and reduce business activity throughout the supply chain, resulting in indirect losses of $263 million in Induced changes in real output, arising from a reduction in labor and property income, add another $355 million to losses. As a result, the total reduction in real gross output is $1.588 billion in Gross output includes expenditures on labor and capital services and purchases from other industries. Value added, which is equivalent to gross domestic or, in this case, gross state product, is an more accurate measure of the change in net economic activity because it does not include inter-industry purchases and instead just includes wages and salaries, proprietor income, other property income, and taxes on production and imports. 1
23 Severance Tax page 19 The impacts on real value added appear in the second panel of Table 4. During the first year under the proposed severance tax, real value added in the Pennsylvania economy declines over $1 billion. Gross state product declines almost $2 billion in 2020, and $2.6 billion in By reducing drilling activity and associated construction activities, reductions in oil and gas production accumulate and correspondingly the losses in real economic output also build over time. Labor income declines $544 million in 2016 and more than $1.6 billion in 2025 from the baseline forecast. Table 4: Summary of economic impacts Millions of 2015 Dollars Gross Output Direct Indirect Induced Total , , , , ,056-4,228 Value Added , , , , ,617 Labor Income , ,611 Employment Number of Supported Jobs ,548-1,185-2,601-6, ,240-2,365-5,789-13, ,940-3,120-7,734-17,795 The severance tax also reduces employment. During 2016, more than 6 thousand fewer jobs are supported by oil and gas in Pennsylvania. These job losses increase to 13,394 in 2020 and to 17,795 in As Table 4 indicates, more than half of these
24 Pennsylvania page 20 losses occur from the direct and indirect impacts, illustrating the adverse impacts on the natural gas industry supply chain. A disaggregation of the impacts of the severance tax on real gross output is displayed in Table 5 for The direct and indirect effects for mining and construction are the largest among the specific impacts by sector, accounting for 68% of the total impacts. Induced impacts account for about 25% of the reduction in gross output, affecting a broad swath of the economy. The remaining 7% include indirect effects across other industries affected by lower output in gas extraction and drilling. Table 5: Direct, indirect, and induced impacts on real gross output by sector Sector Real Gross Output in 2020 Direct & Indirect Induced Total Ag, Forestry & Fishing Mining Construction -1, ,224 Manufacturing Trade, Finance, Inform Prof. & Tech Management Educational Health & social Entertainment Hotel & Food Other services Government & other Total -2, ,194 The impacts of the severance tax on employment and value added by sector are presented in Table 6. Like the change in gross output, impacts are concentrated in the mining and construction sectors. For example, of the total 6,334 fewer jobs supported
25 Severance Tax page 21 under the severance tax in 2016, 3,202 or 50 percent of this loss occurs in mining and construction. Unlike value added and gross output, the employment losses are more distributed across many sectors, with the largest losses occurring across a range of service industries (see Table 6). This is illustrative of the broad impacts that the oil and gas industry has on local economies, generating employment for many people who work outside the mining and construction sectors. Table 6: Total impacts on employment and value added by sector Number of Jobs Supported Real Value Added in Millions Sector Ag, Forestry & Fishing Mining ,886-4, Construction -2,664-3,680-4, ,072 Manufacturing Trade, Finance, Inform Prof. & Tech Management Educational Health & social ,484-1, Entertainment Hotel & Food Other services -1,200-2,618-3, Government & other Total -6,334-13,394-17,795-1,012-1,987-2,617 The impacts of the severance tax on fiscal conditions in Pennsylvania are summarized in Figure 8. During 2016, the severance tax is estimated to collect more than $840 million. As mentioned above, however, the imposition of the severance tax would trigger repeal of the impact fee so the state would lose almost $227 million. In addition, lower economic activity due to the effects of the severance tax reduces state and local income taxes by almost $99 million in The net change in state and local taxes
26 Pennsylvania page 22 during 2016 is $515.6 million (see Figure 8). Even though severance taxes rise to $1.1 billion in 2020 and $1.3 billion by 2025, the offsets in terms of foregone impact fees and lower other state and local taxes results in net tax revenue gains of only $491.7 and $564.3 million in 2020 and 2025 respectively (see Figure 8). Figure 8: Summary of Tax Impacts The sources of changes in state and local taxes are detailed in Table 7. More than 70 percent of the reductions in other state and local taxes are arising from lower sales tax revenues. For example, sales and property taxes decline an estimated $38.2 and $32.8 million respectively in These losses increase to $110.1 and $94.6 million in 2025.
27 Severance Tax page 23 Table 7: Impacts on state and local tax revenues Millions of 2015 Dollars State & Local Tax Revenues Dividends Social Insurance Taxes Employee Contribution Employer Contribution Tax on Production and Imports Sales Taxes Property Tax Motor Vehicle Licenses Other Taxes Fees Corporate Profits Tax Personal Taxes Income Tax Fines- Fees Motor Vehicle License Property Taxes Other Tax (Fish/Hunt) Total State and Local Tax Sensitivity Analysis In any study, there are a myriad of assumptions and judgments that affect the results. This study has deliberately made a number of conservative decisions to avoid an upward bias in the estimated impacts of the severance tax. Conducting the analysis under a range of different assumptions for key parameters would unnecessarily complicate the presentation of the results. Two assumptions, however, are important and deserve discussion. The first involves the shape of the production decline curves from 2015 to Sensitivity analysis of the model indicates that the results are quite sensitive to assumptions for initial production and the second year production levels. If diminishing
28 Pennsylvania page 24 marginal productivity does not set in as assumed above and, instead, output per well even stays at current levels, output of natural gas and associated liquids could be much higher. As a result, the economic losses from not drilling wells due to the severance tax could be much larger than the estimates presented above. The second assumption working in the opposite direction involves the effect of the severance tax on natural gas prices. The above analysis assumes that natural gas prices are unaffected by the tax. In theory, lower gas production in Pennsylvania could put upward pressure on natural gas prices. The size of the potential price increase depends upon the price elasticities of supply and demand for natural gas. The price elasticity of supply developed in this study is 0.67 in the short-run and 1.13 after 10 years. Using estimates developed by Considine et al. (2011), the price elasticity of demand for natural gas is in the short run and after 10 years. Using the methods described in Appendix A., this study finds that natural gas prices would increase between 1.18 and 0.7 percent over the baseline price forecasts. So a small offset due to higher natural gas prices is possible. This offset would reduce the estimated economic impacts of the severance tax. Since these two key assumptions work in opposite directions, this study chooses not to complicate the presentation with a detailed discussion of the results under these possible scenarios. The implication is that the results presented in this study are likely to fall in between these two extremes. Given that the natural gas price impact is likely to be rather small but the possible upside surprise on the production side could be quite large, especially in light of the unexpected size of Marcellus wells, the economic impacts of the severance tax presented above should be considered rather conservative.
29 Severance Tax page Summary and Conclusions Governments find energy reserves an appealing target for raising tax revenue because presumably once an oil or gas well is constructed it cannot be moved. Economic reality, however, is never that simple. This is particularly the case with unconventional oil and gas production. Oil and gas producers now enjoy a much larger playing field and the flexibility to move drilling rigs from state to state to maximize financial returns. While production can be increased quickly over the near term, continuous drilling is required to maintain or expand production over the long run. These technological features and their associated economic realities are important to consider when evaluating the economic impacts of the proposed severance tax on oil and gas in Pennsylvania by Governor Tom Wolf. The proposed tax involves a 5% levy on the gross value of wellhead production plus a fixed fee of 4.7 cents per thousand cubic feet produced. The proposed severance tax also has a minimum per-unit gross value provision or price floor. If prices fall below this minimum threshold, taxes are assessed at the minimum gross value, in effect increasing the burden of the severance tax. Finally, under Pennsylvania law, if a severance tax is enacted, the existing impact fee, which is a fixed fee assessed on each well drilled, would be repealed. This study conducts an economic impact analysis of this proposed tax. The model captures how the severance tax affects drilling and production activity. Production decline curves are generated from detailed production data from thousands of individual oil and gas wells. Projections of future drilling with and without the tax, therefore, are grounded in empirically based drilling and production relationships.
30 Pennsylvania page 26 The proposed severance tax reduces investment in drilling. During the first year of the tax, this study estimates that 113 fewer horizontal wells are drilled. More than 150 fewer wells are drilled in 2025 under the tax compared to the baseline without the severance tax. Fewer wells drilled mean a lower number of producing wells and, therefore, lower production of natural gas and petroleum liquids. These production losses mount over time. In 2016, the production losses are 0.24 billion cubic feet per day. By 2025, the annual volume of production lost due to the tax increases to 0.93 billion cubic feet per day (see Figure 9). Figure 9: Losses in drilling and production due to the severance tax These reductions in oil and gas industry investment and production have economic consequences. During 2016, over 6 thousand fewer jobs are supported by the oil and gas industry in Pennsylvania. Job losses occur not just in the oil and gas sector but also across a range of industries that are part of the gas industry supply chain and from service industries dependent upon spending by workers employed in these industries.
31 Severance Tax page 27 Like the physical production losses above, the economic losses grow larger over time. More than 17.8 thousand fewer jobs are supported by the industry in 2025 under the severance tax. Value added, which is the broadest measure of economic output for the state, is reduced by $1.0 billion in 2016 and by $2.6 in 2025 (see Figure 10). Cumulative losses in value added are over $20 billion. Figure 10: Losses in employment and value added due to the severance tax While the severance tax is viewed as a way to raise revenue, these broader economic losses overwhelm any revenue the Commonwealth gains from a severance tax. Furthermore, with lost impact fee revenue and lower state and local taxes due to the severance tax, the net change in tax revenue from the severance tax is less than 50% of the revenue expected by its supporters. After conducting this analysis and discovering how assumptions on production decline curves and economic relationships influence the results, the above estimates should be considered conservative. If Marcellus production continues to exceed
32 Pennsylvania page 28 expectations, the economic impacts of the severance tax could be much larger than those estimated above. Overall, this study confirms economic intuition. Taxing a highly productive industry that supports thousands of jobs and generates billions in economic output and millions in tax revenues has negative economic consequences for the Commonwealth of Pennsylvania. These results indicate there are economic trade-offs associated with the proposed severance tax. Even though state and some local governments in Pennsylvania may have more tax revenues if the proposed legislation is passed, the net tax revenue gain is considerably less than the $1 billion expected by its supporters. This study finds that the reason for these unmet expectations is that the severance tax has significant negative impacts on the oil and gas industry and the state s economy. The size of these economic losses also suggest that expectations for offsetting the lost jobs and economic output by spending the severance tax revenues are also unlikely to be realized.
33 Severance Tax page 29 References Considine, T.J., R. Watson, R. Entler, J. Sparks (2009) An Emerging Giant: Prospects and Economic Impacts of Developing the Marcellus Shale Natural Gas Play, The Pennsylvania State University, Dept. of Energy and Mineral Engineering, August 5, 39 pages. Considine, T.J., R. Watson and S. Blumsack, (2010) The Economic Impacts of the Pennsylvania Marcellus Shale Natural Gas Play: An Update, The Pennsylvania State University, Energy and Mineral Engineering, May 2010, 21 pages. Considine, T.J., R. Watson and S. Blumsack (2011) The Pennsylvania Marcellus Shale Natural Gas Industry: Status, Economic Impacts, and Future Potential, The Pennsylvania State University, July, 59 pages. Deacon, R., S. DeCanio, H.E. Frech, M.B. Johnson (1990) Taxing Energy: Oil Severance Taxation and the Economy, The Independent Institute, 161 pages. Epple, D. (1985) The Econometrics of Exhaustible Resources: A Theory and an Application, in Sargent, T, editor. Energy Foresight and Strategy, Resources for the Future, Washington, DC. Miller, R.E. and P.D. Blair (1985) Input-output Analysis Foundations and Extensions, Cambridge University Press, 464 pages.
34 Pennsylvania page 30 Appendix A: Analysis of Price Impact of Severance Tax Consider the equilibrium condition for the natural gas market: Q d = Q o + Q r, (1) where Q d is the total demand for natural gas, Q r is production of natural gas from the region under study, andq o is natural gas supply from other regions. Recognizing that each quantity in (1) is a function of price, taking the total differential of (1) and re-arranging terms yields: dq d dp dp = dq o dq r dp dp dq r +1 (2) Factoring equation (2) and transforming to express in terms of elasticities provides: dp é æ P dq Q d ö d PdQ r è ç Q d dp ø - Q æ P dq o ö ù ê o è ç Q o dp ø ú = 1 ë û d ln P é æ d lnq Q d ö æ d lnq d dq r è ç d ln P ø - Q o ö ù ê o è ç d ln P ø ú ë û = 1 dq d ln P = r Q d e d - Q o e o [ ] where e d is the elasticity of total market demand and e o is the elasticity of supply from other regions.