RENEWABLE PORTFOLIO STANDARDS: PENNSYLVANIApassed Pennsylvania’s first Renewable Portfolio...

RENEWABLE PORTFOLIO STANDARDS:PENNSYLVANIAFINAL REPORT | APRIL 2016

TYLER BROUGH PH.DUtah State University

JACOB FISHBECKStrata Policy

RANDY T SIMMONS PH.DUtah State University

RYAN M. YONK PH.DUtah State University

Institute of Political Economy, Utah State University | 1

RENEWABLE PORTFOLIO STANDARDS: PENNSYLVANIA

PRIMARY INVESTIGATORS:

Randy T Simmons, PhD Utah State University

Ryan M. Yonk, PhD Utah State University

Tyler Brough, PhD Utah State University

Jacob Fishbeck Strata Policy

STUDENT RESEARCH ASSOCIATES:

Daniel Mikkelsen Josh Blotter Brian Isom

2 | Renewable Portfolio Standards: Pennsylvania

TABLE OF CONTENTS Executive Summary ............................................................................................................................................................. 3 History and Background ...................................................................................................................................................... 3 Empirical Economic Analysis .............................................................................................................................................. 7 State Coincident Event Study ............................................................................................................................................. 7 The Structural Panel VAR-X Model .................................................................................................................................. 10 Conclusions from the Empirical Analysis ......................................................................................................................... 11 Institutional Analysis ........................................................................................................................................................ 11 Conclusions from the Institutional Analysis ............................................................................................................... 15 Conclusion ......................................................................................................................................................................... 16 Explanation of Empirical Study Methodology .................................................................................................................. 17 0.1. The Panel VAR-X Model ...................................................................................................................................... 17 0.2. Model Estimation ................................................................................................................................................. 18 0.3. Dynamic Multiplier Analysis ................................................................................................................................ 20 Graphical Analysis of the Dynamic Multiplier Analysis ................................................................................................... 22


THE COST OF RENEWABLE PORTFOLIO STANDARDS IN PENNSYLVANIA

EXECUTIVE SUMMARY

The U.S. has no federal mandate for “renewable” power production. Instead, a majority of states, including Pennsylvania, have created their own state laws called Renewable Portfolio Standards (RPS). These laws mandate that electric utilities provide a certain percentage of power from renewable sources. This report analyzes how changes in electricity markets caused by RPS alter the functioning of a state’s economy and institutions, with a specific focus on Pennsylvania. Our report uses an empirical analysis and a survey of legal rules, or institutional analysis. The following are our key findings:

• Our empirical analysis discovered significant harmful effects on the economies of all states with RPS. States that have adopted an RPS have seen a drop in industrial electricity sales by almost 14 percent. Real personal income has fallen by almost four percent, which for Pennsylvania figures to a loss of $19.8 billion in 2009, or $4,002 less per family. Non-farm employment has declined by nearly three percent. Lastly, RPS is correlated with an increase of 10 percent in a state’s unemployment rate, equaling a loss of 49,898 jobs in Pennsylvania.

• Our analysis of the legal rules surrounding RPS in Pennsylvania suggests that the regulatory climate it creates is counterproductive and contradictory, making RPS an even worse venture for taxpayers than the empirical analyses suggests.

HISTORY AND BACKGROUND

Pennsylvania has been known as “The Coal State,” and is home to one of the few major deposits of high-quality anthracite coal in the U.S.1,2 At the turn of this century, 57 percent of electricity in Pennsylvania was generated from coal, while 36 percent came from nuclear, 4 percent from other non-renewable sources, and

1 NETSTATE. (2015, March 29). Pennsylvania. Retrieved from http://www.netstate.com/states/intro/pa_intro.htm

2 ExplorePAhistory.com. (n.d.). Mining anthracite. Retrieved from http://explorepahistory.com/story.php?storyId=1-9-B


2 percent from renewable sources.3 As part of a growing national trend, however, in 2004 legislators passed Pennsylvania’s first Renewable Portfolio Standard (RPS), in an attempt to shift the balance of electricity generation away from coal and toward renewable sources. Prior to this, two other RPS had been proposed in the state legislature. First, Senate Bill 962 of 2003 was supported by several environmental groups who touted it as the most progressive RPS ever created in the United States—yet the bill never left committee.4,5 Weeks later, House Bill 2250 introduced a different approach to RPS: Electric utilities would purchase a certain percentage of their total electricity from two separate lists of alternate sources, designated “Tier 1” and “Tier 2.” Uniquely, however, the bill listed waste coal as a qualifying electricity source, which some consider dirty.6 This bill also did not pass through committee.7 Senate Bill 1030 was the most recent and only successful RPS to be proposed in Pennsylvania. The legislature passed this bill, and in November of 2004 Governor Edward Rendell signed the bill into law as Act 213, establishing the Alternative Energy Portfolio Standard (AEPS). Act 213 borrowed heavily from language in HB 2250, but included even more “dirty” electricity sources as qualifying for meeting the AEPS requirement, causing concern among the groups who had supported RPS from the beginning.8,9 The AEPS requires that 18.5 percent of electric utilities’ sales come from alternative sources by 2021. Utilities have been meeting this requirement gradually over a 15-year period that started in 2007, the first year of compliance. In that first year, utilities needed to provide 5.7 percent of their total electricity from alternative sources (they met this requirement).10 Further, utilities must draw from three categories of sources: Tier 1 and Tier 2, as well as solar photovoltaic (PV). By 2021, eight percent of electricity must come from Tier 1 sources, ten percent from Tier 2 sources, and a half percent from solar PV. Tier 1 sources include solar thermal, wind, low-impact hydroelectric, geothermal, biologically derived methane gas (including landfill gas), fuel cells, biomass, coal mine methane, black liquor (from Pennsylvania

3 Commonwealth Economics. (2013, February). Energy in Pennsylvania: Past, present, and future. Retrieved from

http://www.elibrary.dep.state.pa.us/dsweb/Get/Document-96943/Final%20PA%20Comprehensive%20Energy%20Analysis.pdf, p. 59 4 Pennsylvania General Assembly. (n.d.). Senate Bill 962. Retrieved from

http://www.legis.state.pa.us/cfdocs/billinfo/bill_history.cfm?syear=2003&sind=0&body=S&type=B&bn=962 5 ActionPA. (n.d.). Pennsylvania’s “alternative” energy law. Retrieved from http://www.actionpa.org/cleanenergy/#history

6 Waste coal is the excavated soil from coal mines. It contains some combustible material. 7 Pennsylvania General Assembly. (n.d.). House Bill 2250. Retrieved from

http://www.legis.state.pa.us/cfdocs/billinfo/bill_history.cfm?syear=2003&sind=0&body=H&type=B&bn=2250 8 Ibid.

9 Pennsylvania General Assembly. (n.d.). Senate Bill 1030. Retrieved from

http://www.legis.state.pa.us/cfdocs/billinfo/bill_history.cfm?syear=2003&sind=0&body=S&type=B&bn=1030 10 Pennsylvania Public Utility Commission (PUC). (2008, May). 2007 annual report: Alternative Energy Portfolio Standards Act of 2004. Retrieved from http://www.puc.state.pa.us/electric/pdf/AEPS/AEPS_Ann_Rpt_2007.pdf, p. 6


only), and under certain restrictions, large-scale hydroelectric. Tier 2 sources include waste coal, distributed generation systems, demand-side management, large-scale hydroelectric, municipal solid waste, generation of electricity utilizing by-products of the pulping process and wood, and integrated combined coal gasification technology (our institutional analysis will show these classifications to be counterproductive and environmentally harmful.)11 Six years into the life of the AEPS, the makeup of electricity generation in Pennsylvania had transformed: Now 48 percent of electricity came from coal, 34 percent came from nuclear, 15 percent came from natural gas, and 2 percent came from renewable sources, as shown in Figure 1.12 Figure 1: Shares of Electricity Generation by Fuel Type in Pennsylvania, 2000 and 2010

Source: Commonwealth Economics13

Between 2000 and 2010, the most significant change was the increase in electricity generation from natural gas. Commonwealth Economics has estimated that natural gas will continue to grow its share of generation, reaching 23 percent by 2017, and renewable sources will constitute five percent of generation at that time.14 Pennsylvania’s AEPS set several national precedents: It was the first RPS to include fossil fuels as a qualifying electricity source, and the first to require a higher percentage of electricity from a category of sources that includes fossil fuels and fossil fuel wastes than from a category that includes only renewables.

11 United States Department of Energy (DOE). (2015, March 30). Database of state incentives for renewables and efficiency.

Retrieved from http://programs.dsireusa.org/system/program/detail/262 12

Commonwealth Economics (2013), op. cit., p. 59 13 Ibid. 14

Ibid.


As a result, ActionPA, an environmental advocacy organization, expressed displeasure with Act 213, stating that the law “could do serious environmental damage, while posing as an innovative clean energy bill.”15 The AEPS remains mostly unchanged since its enactment in 2004. It was amended slightly in 2007 with the passage HB 1203, which fixed several semantic loopholes and more rigorously defined “customer generation.”16 Some of the substantive new legislation pertaining to Act 213 has been designed to increase AEPS requirements.17,18 On the other hand, multiple legislators have tried to expand the list of Tier 1 sources, which would make the law easier to comply with and therefore less meaningful.19,20,21,22,23,24 Three similar bills would have lessened the AEPS requirements overall, while strengthening the requirement for solar PV in particular.25,26,27 And finally, one attempt to repeal Act 213 took place in 2013, but failed.28 Of all the proposed substantive legislation, only SB 1015 of 2013 passed, which moved municipal waste from Tier 2 to Tier 1, allowing utilities to comply with the law using dirtier electricity and making the law less meaningful for pursuing environmental goals. Several studies have attempted to quantify the costs of Pennsylvania’s AEPS. The Pennsylvania Public Utilities Commission reports their own cost projections annually, and most recently says it will cost $165

15 ActionPA (n.d.), op. cit.

16 Pennsylvania General Assembly. (n.d.). House Bill 1203. Retrieved from

http://www.legis.state.pa.us/cfdocs/billinfo/bill_history.cfm?syear=2007&sind=0&body=H&type=B&bn=1203 17

Pennsylvania General Assembly. (n.d.). House Bill 100. Retrieved from http://www.legis.state.pa.us/cfdocs/billinfo/bill_history.cfm?syear=2013&sind=0&body=H&type=B&bn=100 18


Pennsylvania General Assembly. (n.d.). Senate Bill 1015. Retrieved from http://www.legis.state.pa.us/cfdocs/billinfo/bill_history.cfm?syear=2013&sind=0&body=S&type=B&bn=1015 20



Pennsylvania General Assembly. (n.d.). House Bill 1151. Retrieved from http://www.legis.state.pa.us/cfdocs/billInfo/bill_history.cfm?syear=2013&sind=0&body=H&type=B&bn=1151 23



Pennsylvania General Assembly. (n.d.). Senate Bill 1171. Retrieved from http://www.legis.state.pa.us/cfdocs/billInfo/bill_history.cfm?syear=2013&sind=0&body=S&type=B&bn=1171 26

Pennsylvania General Assembly. (n.d.). House Bill 100. Retrieved from http://www.legis.state.pa.us/cfdocs/billInfo/bill_history.cfm?syear=2015&sind=0&body=H&type=B&bn=100 27


Pennsylvania General Assembly. (n.d.). House Bill 1912. Retrieved from http://www.legis.state.pa.us/cfdocs/billInfo/bill_history.cfm?syear=2013&sind=0&body=H&type=B&bn=1912


million for utilities to comply with the AEPS in 2021. This is the low end of all available cost estimates.29 An economic study by the Beacon Hill Institute , however, argued that the AEPS would cost $2.55 billion in 2021, which would be born by utilities and ratepayers.30 In the middle of those estimates, a study from Pennsylvania State University puts the AEPS at a cost of $434 million in 2021 in its worst-case scenario. Lastly, regional electricity transmitter PJM analyzed several scenarios of renewable penetration into the electricity market and predicted that generation costs will generally decrease on net.31,32 Out of all these studies, only the Beacon Hill Institute looks at the economy as a whole, as opposed to looking only at the effects on electricity prices or generation costs. This is useful when trying to measure the effects of renewable electricity mandates, which often contain clauses that prevent electricity prices from rising. Beacon Hill’s theoretical analysis only predicts the future, though, and does not report on what has actually happened in the state of Pennsylvania. Our economic analysis uses empirical data to show the comprehensive effects of the AEPS since its enactment.

EMPIRICAL ECONOMIC ANALYSIS

STATE COINCIDENT EVENT STUDY

In this section, we present the results of an event study for state coincident indices—a methodology first fashioned by the Federal Reserve Bank of Philadelphia.33 The event study indexes the economic conditions of all states across multiple points in time, and assigns as “point zero” each state’s economic conditions on the dates of their respective RPS implementations. The study compares said economic conditions over a span from 48 months before to 48 months after that enactment date. The indices of each state RPS policy, therefore, while enacted in different calendar months and years, can thus be lined up in this so-called “event time,” and the economic conditions in each state can be averaged. Given that RPS have been implemented in many states over a long period, this will minimize the effects of anomalies such as recessions and the enactment of other energy-related laws. For these reasons, the event study has become a time-honored empirical methodology in finance and economics and a standard course of analysis for the

29 Pennsylvania Public Utility Commission (PUC). (2014). 2014 annual report: Alternative Energy Portfolio Standards Act of 2004.

Retrieved from http://www.puc.state.pa.us/electric/pdf/AEPS/AEPS_Ann_Rpt_2014.pdf, p. 14 30

Beacon Hill Institute. (2012, December). The economic impact of Pennsylvania’s Alternative Energy Porfolio Standard. Retrieved from http://www.beaconhill.org/BHIStudies/PA-AEPS2012/PA-AEPS-study-BHI-Dec-2012.pdf, p. 8 31

Blumsack, S. A., Kleit, A. N., & Idrisu, B. (2011, October). The cost of alternative energy portfolio standards in Pennsylvania. Retrieved from http://www.ems.psu.edu/sites/default/files/u5/Cost_of_PA_AEPS_Study.pdf, p. 41 32

GE General Consulting. (2014, March 3). PJM renewable integration study (PRIS): Final project review. Retrieved from http://www.pjm.com/~/media/committees-groups/committees/mic/20140303/20140303-pjm-pris-final-project-review.ashx 33

Federal Reserve Bank of Philadelphia. (2015, January 29). State coincident indexes. Retrieved from http://www.philadelphiafed.org/research-and-data/regional-economy/indexes/coincident/


Philadelphia Fed. It is a simple but powerful method for measuring the effect of an exogenous shock to an economic variable of interest. MacKinlay gives an in-depth discussion of the event study methodology.34 Table 1: The Dates (Month and Year) of the 31 States That Have Enacted an RPS Policy to Date State RPS Enactment Date Arizona July, 2007 California January, 2003 Colorado December, 2004 Connecticut July, 1998 Delaware July, 2005 Hawaii December, 2003 Iowa January, 1983 Illinois August, 2007 Kansas July, 2009 Massachusetts April, 2002 Maryland January, 2004 Maine March, 2000 Michigan October, 2008 Minnesota February, 2007 Missouri November, 2008 Montana April, 2005 North Carolina January, 2008 New Hampshire July, 2007 New Jersey September, 2001 New Mexico September, 2007 Nevada January, 1997 New York September, 2004 Ohio May, 2008 Oregon January, 2007 Pennsylvania February, 2005 Rhode Island June, 2004 South Carolina June, 2014 Texas September, 1999 Washington November, 2006 Wisconsin December, 2001 West Virginia July, 2009

34 MacKinlay, A.C. (1997). Event studies in economics and finance. Journal of Economic Literature, 35(1), 13-39.


The results of the event study are presented in Figure 1, wherein we see the response of the state coincident index to the enactment of RPS policies. The coincident index is a summary measure of the strength of a st ate economy, and is comprised of four economic indicators: nonfarm payroll employment, average hours worked in manufacturing, the unemployment rate, and wage and salary disbursements deflated by the Consumer Price Index (CPI).35 Figure 2: The Response of The State Coincident Index to the Enactment of RPS Policies.

The horizontal axis shows months before and after point zero (RPS enactment). The vertical axis shows an indexed scale measuring the average reaction of states in terms of several economic indicators.

As can be seen in Figure 2, the average effect on the state coincident index is a precipitous drop surrounding the enactment of an RPS policy. This evidence is suggestive of a negative effect of an RPS

35 Federal Reserve Bank of Philadelphia. (2015, January 29). State Coincident Indexes. Retrieved from

http://www.philadelphiafed.org/research-and-data/regional-economy/indexes/coincident/


policy on a state economy. While suggestive, the evidence from the event study warrants further exploration into the effects, since state economies also appear to decline several months prior to the enactment of an RPS. The next section presents the structural panel VAR-X model, which provides further evidence of the negative economic effects of an RPS.

THE STRUCTURAL PANEL VAR-X MODEL

The VAR model takes into account the nature of the state macroeconomic variables that could provide unwanted feedback into the model, and considers their dynamic interactions. By including a panel dimension to the model we can include the data for multiple states in a single model. We include fixed effects to control for state-level heterogeneity. We impose a recursive causal ordering on the VAR-X model to allow for structural interpretation of dynamic multiplier analysis of the RPS policy variable. Table 2 presents the cumulative effects of an RPS on the state economy via structural policy simulations. Table 2: The Long-Run Effects on State Macroeconomic Variables

State Economic Variable Long-Run Effect

Electricity Sales -13.7075% Real Personal Income -3.6369% Non-farm Employment -2.8416% Manufacturing Employment 3.7454% Unemployment Rate 9.6841% The cumulative effect of the enactment of an RPS policy on state electricity sales is a staggering 13.7-percent decline. This is, perhaps, not surprising as the RPS increases the cost of electricity generation. Real personal income declines in the long run by 3.6369 percent, which figures to a loss of $19.8 billion in 2009, or $4,002 less per family.36,37 Non-farm employment declines in the long run by 2.8 percent. Only one analyzed component of non-farm employment, manufacturing employment, does not experience a long-term suppression in response to an RPS policy, although as we see in the later graphical analysis, it does still experience a sharp decline in the short term. Most significantly, the state unemployment rate increases by 9.6 percent. This means that, in February of 2009, Pennsylvania had 49,898 fewer jobs than it would have had without the RPS.38,39 There can be no doubt that the combined economic effect on an RPS enactment, as measured

36 Bureau of Economic Analysis. (n.d.). Regional Data, Annual State Personal Income and Employment. Retrieved from

http://www.bea.gov/iTable/iTable.cfm?reqid=70&step=1&isuri=1&acrdn=1#reqid=70&step=1&isuri=1 37

United States Census Bureau. (2015, December 1). State and county quick facts. Retrieved from http://quickfacts.census.gov/qfd/states/42000.html 38

Bureau of Labor Statistics (n.d.). Pennsylvania. Retrieved from http://www.bls.gov/eag/eag.pa.htm


by the structural panel VAR-X model, is a severe decline in the Pennsylvania economy. A graphical representation of the analysis, showing the changes over time that lead to these results, can be found in Appendix B.

CONCLUSIONS FROM THE EMPIRICAL ANALYSIS

We demonstrate strong empirical evidence that a Renewable Portfolio Standard has a lasting negative effect on a state economy. We present this evidence from both an event study of the state coincident index as measured by the Federal Reserve Bank of Philadelphia, as well as from structural policy simulations from a panel VAR-X model. The long-run effect of an RPS on state industrial production, as measured through industrial electricity sales, is almost a 14-percent decline. Real personal income declines in the long run after an RPS by almost four percent. The cumulative effect of an RPS on non-farm employment is nearly three percent. While the effect of an RPS on manufacturing employment is not as severe in the long run, it too demonstrates an initial sharp decline lasting for several years. Finally, the state unemployment rate increases in the long run in response to an RPS by nearly 10 percent. These are strong and lasting effects in four of the five variables measuring the state economy. The combined econometric evidence makes clear that an RPS policy has a severely negative economic effect on a state that enacts such.

INSTITUTIONAL ANALYSIS

In addition to the AEPS’ negative economic effects, judging the law in the context of its own legislative goals reveals inherent and serious structural problems. Like all RPS, Pennsylvania’s AEPS is purportedly meant to help the environment by increasing the proportion of electricity generated from renewable sources. The AEPS creates a regulatory structure, however, that makes compliance exceedingly difficult. Further, even if utilities achieve compliance, the AEPS does not help the environment in any meaningful way. One of the obvious inherent problems with the AEPS is that it sanctions electricity generation from sources that are more environmentally detrimental than conventional fossil fuels. 40 While eight percent of Pennsylvania’s electricity by 2021 must come from Tier 1 sources—which are themselves not the

39 We arrived at this number by applying the results of the empirical model to Pennsylvania’s employment statistics. Using arithmetic, we calculated what the unemployment rate would have been absent the AEPS. We subtracted the number of employed under the AEPS from the number of employed under a no-AEPS scenario (what would have been required to produce the new unemployment rate). 40

ActionPA (n.d.), op. cit.


cleanest—ten percent of electricity must come from the even dirtier Tier 2 sources, such as waste coal.41 For those who wish to combat global warming or air pollution in general, the AEPS may be unsatisfactory. One example of a dirty Tier 1 source is animal waste. While burning animal waste for electricity may reduce carbon dioxide (CO2) emissions, it may increase emissions of other pollutants. As Baranyai and Bradley note, “due to the high concentration of nitrogen in the poultry manure, there may be significant fuel NOx emissions. Nitrogen-oxides dissolved in the moisture of the atmosphere cause acid-rain and immediately end up in surface waters.”42 “Biologically” derived methane (including landfill gases) is another Tier 1 source that poses environmental issues. First, only about 19 percent of all landfill gases are captured over a landfill’s lifetime.43 This means that about four-fifths of the gas, which consists of roughly half CO2 and half methane, is released into the atmosphere, almost completely offsetting the emissions reductions from burning the gas in place of fossil fuels.44 Instead of trying to capture the gas, therefore, Peter Anderson of the Center for a Competitive Waste Industry advises that governments simply phase out the landfilling of organic waste.45 Even if we could capture more of the gas, burning it may create dioxins, furans, and other compounds, which are hazardous to human health and potentially environmentally damaging.46,47 Another environmentally questionable Tier 1 source is coal mine methane (CMM). CMM remains trapped in coal mines until it is “liberated,” at which point it must be intercepted, or “captured,” and prevented from escaping into the atmosphere. Once it is captured, CMM can be transported and burned to generate electricity. Although capturing the gas has become easier over recent decades, the vast majority of CMM escapes into the atmosphere once released.48 This once again offsets the desired reductions in greenhouse gases. Lastly, fuel cells are included in Tier 1, but they do nothing to help us avoid using fossil fuels. Most hydrogen in the United States (greater than 90 percent) is created by “steam reforming” natural gas, and

41 DOE (2015), op. cit.

42 Baranyai, V., & Bradley, S. (2008). Turning Chesapeake Bay Watershed poultry manure and litter into energy. Chesapeake Bay

Program. Retrieved from http://www.chesapeakebay.net/documents/cbp_17018.pdf 43

Anderson, P. (2006). The landfill gas energy recovery hoax. Retrieved from http://www3.epa.gov/lmop/documents/pdfs/conf/9th/anderson_abs.pdf 44

Ibid. 45

Ibid. 46

Ewall, M. (2013). Primer on landfill gas as green energy. Retrieved from http://www.energyjustice.net/lfg 47

United States Environmental Protection Agency (EPA). (1994, June). Estimating exposure to dioxin-like compounds – volume II: Properties, sources, occurrence, and background exposure. Retrieved from http://nepis.epa.gov/Exe/ZyPDF.cgi/200099OM.PDF?Dockey=200099OM.PDF 48

The National Coal Council. (2003, May). Coal-related greenhouse gas management issues. Retrieved from http://www.nationalcoalcouncil.org/Documents/fpb.pdf, p. 7, 40


the energy output of hydrogen is actually less than the energy required in the reforming process (85 percent at best).49,50 It is therefore disingenuous to claim that electricity from fuel cells is clean or renewable. Despite being counted towards the purportedly environmentally friendly AEPS, Tier 2 sources emit even more than Tier 1 sources or conventional fossil fuels. For example, burning waste coal in fluidized bed combusters (FBCs), a type of low-temperature coal power plant, emits more carbon monoxide (CO) and nitrous oxide (N2O) than traditional coal plants.51,52 According to the EPA's Toxic Release Information (TRI) database, burning waste coal emits more mercury and other harmful pollutants than regular coal.53 The Pennsylvania Department of Environmental Protection, however, claims that waste coal mercury emissions are a fraction those from regular coal.54 Similar disputes arise for sulfur dioxide (SO2) and hazardous organic compounds.55 Regardless of the environmental consequences, though, the AEPS has certainly spurred the use of waste coal facilities: Thirteen of the sixteen waste coal-burning facilities that have qualified for the AEPS reside in Pennsylvania.56 If waste coal is dirtier than regular coal, then the AEPS will propel Pennsylvania to the top of all pollution emitting states.

Overall, the lists of qualifying resources for the AEPS may be troubling to those who generally view RPS as a way to promote renewable energy technologies like wind and solar. But aside from a 0.5-percent “carve-out” for electricity generated from solar photovoltaic (PV), the AEPS has no requirements for electricity to come from any particular source in either of its two tiers. This means that utilities could hypothetically meet all AEPS requirements without building a single wind turbine. Far from moving the law toward environmental friendliness, however, the aggressive solar PV carve-out has actually created problems in Pennsylvania’s solar PV market. The only way for utilities to comply with the AEPS is to retire a certain number of renewable energy credits (REC), each of which represents one megawatt-hour (MWh) of electricity. These credits come into existence when a power plant (e.g., a solar farm) generates one MWh of electricity, and are bought and sold by utilities on an exchange. The prices of

49 Ogden, J. M. (1999, November). Prospects for building a hydrogen energy infrastructure. Annual Review of Energy and the

Environment, 24, 227-279. DOI: 10.1146/annurev.energy.24.1.227. Retrieved from EBSCO Academic Search Premier., p. 234 50

Ibid. 51

Mastral, A. M., Callén, M. S., & Garcia, T. (2000, March). Toxic organic emissions from coal combustion. Fuel Processing Technology, 67(2000), 1-10. Retrieved from http://www.energyjustice.net/files/coal/wastecoal/pah/2000FPT.pdf, p. 2 52

The National Coal Council (2003), op. cit., p. 7 53

Pennsylvania Department of Environmental Protection. (n.d.). Analysis of waste coal-fired combustion units in Pennsylvania. Retrieved from http://www.portal.state.pa.us/portal/server.pt/document/1421185/emissions_analysis_pdf&usg=AFQjCNGYDlYEGiYr-Cr5BCJJgXFDnKIpag&sig2=6p0qAociBy7zJ5xInUxm4w&bvm=bv.113370389,d.cGc, p. 1 54

Ibid. 55

Ibid. 56

PUC. (n.d.). Reports. Retrieved from http://pennaeps.com/app/publiccontroller/download_QFs/


RECs fluctuate, though, and solar PV manufacturers will only continue to build new panels if they expect the price of solar RECs to sustain their industry. Because the solar carve-out was so aggressive, however, this caused a “flood of new [solar] installations” early on that led to an oversupply and a decrease in price of the photovoltaic energy credits.57 Because those credits serve as a critical funding source for solar projects, this actually endangered future solar energy projects in Pennsylvania, and puts thousands of solar-related jobs at risk.58 The crisis in solar REC prices also reaffirmed that the solar PV industry is dependent on financial incentives. Solar PV is a relatively expensive form of electricity generation, and would probably not be widely adopted without support from the state of Pennsylvania.59,60 If the political climate changes, therefore, and that support were removed, the technology would likely be abandoned. If a goal of the AEPS is to create a self-sustaining solar industry, the law cannot currently fulfill its objective.61 Act 213 (which created the AEPS) also includes a counterintuitive force majeure clause. The PUC can modify or temporarily waive the obligations of an individual utility under the AEPS, provided that the utility can show a good-faith attempt to comply with the AEPS.62 Ohio’s RPS includes a similar clause that illustrates the contradictory effects of force majeure. In that state, during the force majeure determination process, “utilities will be uncertain as to their liability—not only for the preceding year, but for the following year,” and the Public Utilities Commission of Ohio can increase a utility’s future obligations by the amount of its missed past obligations.63

57 HB 1580. Hearing before the Consumer Affairs Committee, Pennsylvania House of Representatives, Session of 2011. (2012,

January 11). (Multiple Testimonies). Retrieved from http://www.legis.state.pa.us/cfdocs/legis/tr/transcripts/2012_0002T.pdf, p. 8 58

Ibid., p. 9 59

U.S. Energy Information Administration (EIA). Levelized cost and levelized avoided cost of new generation resources in the Annual Energy Outlook 2015. Retrieved from https://www.eia.gov/forecasts/aeo/electricity_generation.cfm 60

HB 1580 Hearing (2012), op. cit., p. 8 61

Pennsylvania Public Utilities Commission (2013) 2013 Annual Report: Alternative Energy Portfolio Standards Act of 2004. pgs 11-12. Retrieved from: http://paaeps.com/credit/getFilenouser.do?file=AEPSReport_13.pdf&docdir=true 62

General Assembly of Pennsylvania. (2007) House Bill 1203. Amendment to Section 1. Retrieved from: http://www.legis.state.pa.us/CFDOCS/Legis/PN/Public/btCheck.cfm?txtType=HTM&sessYr=2007&sessInd=0&billBody=H&billTyp=B&billNbr=1203&pn=1995 63

Simmons, R. T., Yonk, R. M., Brough, T., Sim, K., Fishbeck, J. (2015, April). Renewable Portfolio Standards: Ohio. Institute of Political Economy, Utah State University. Retrieved from http://www.strata.org/wp-content/uploads/2015/06/RPS-Ohio-Report.pdf, p. 18


The language in Pennsylvania’s AEPS emulates this feature. Act 213 reads,

“[…] the Commission may require the electric distribution company or electric generation supplier to acquire additional alternative energy credits in subsequent years equivalent to the obligation reduced due to a force majeure declaration […]”64

This makes utilities’ investment decisions difficult and risky. A utility could hypothetically over-invest in renewable electricity in order to ensure compliance in years after a force majeure determination, which could cause financial hardship down the road. A utility that overinvests in renewable electricity could be at a competitive disadvantage when negotiating for wholesale electricity contracts in the future.

CONCLUSIONS FROM THE INSTITUTIONAL ANALYSIS

This institutional analysis shows that the structure of Pennsylvania’s AEPS is counterproductive and contradictory. In many cases it either fails to promote environmental benefits, or works directly against the clean energy ideals it was designed to attain. Environmental protection, therefore, cannot be used as a justification for the harsh economic effects that the AEPS causes.

64 Pennsylvania General Assembly. (2004, November). Act 213. Retrieved from

http://www.legis.state.pa.us/cfdocs/legis/li/uconsCheck.cfm?yr=2004&sessInd=0&act=213


CONCLUSION

The results from these analyses paint a clear picture about the effects of RPS. Our empirical analysis finds a reduction in industrial electricity sales by almost 14 percent, real personal income by almost four percent, non-farm employment by nearly three percent and an increase of 10 percent in the unemployment rate, which for Pennsylvania equals a loss of 49,898 jobs. Our analysis of the legal rules surrounding Pennsylvania’s AEPS describes the barriers that make it difficult for utilities to comply and for bureaucracies to enforce it, finding that Pennsylvania’s tiered structure of qualifying electricity sources does little to achieve its environmental goals. Any state currently deliberating on implementing a new RPS, or strengthening an existing one, should heed these results. Finally, states should seriously weigh the costs of these regulations against the gain of following a politically popular trend.


Appendix A

EXPLANATION OF EMPIRICAL STUDY METHODOLOGY

Methodology Constructed by Tyler Brough, Ph.D. at Utah State University In this technical appendix, we outline the details of the structural panel VAR-X model, its estimation, and its use for policy simulation.

0.1. THE PANEL VAR-X MODEL

The vector autoregressive (VAR) model is the standard work horse model in empirical macroeconomics. The basic p-lag VAR model can be written as:

𝑦" = 𝑎% + 𝐴(

)

(*+

𝑦",( + 𝜀"

where 𝑦" for t = 1, . . . , T is an M vector of observations on M time series variables, 𝜀" is an M × 1 vector of errors, 𝑎% is an M × 1 vector of intercepts and the 𝐴( are M × M matrices containing model

coefficients.65 This is the reduced-form VAR model. For the present study 𝑦" = (𝑦+",...,𝑦.")ʹ , M = 5, and the 𝑦(" are the five state macroeconomic variables presented in the main body of the paper, namely electricity

sales, real personal income, non-farm employment, manufacturing employment, and the unemployment rate. Thus, the VAR model is a system of M equations, with one equation for each variable in the system. Each of the M = 5 variables is treated as endogenously determined. The present model also includes an exogenous policy variable that represents the enactment of an RPS by a given state. Thus, we can now write the VAR-X (a VAR model with the exogenously determined variable) as follows:

𝑦" = 𝑎% + 𝐴(

)

(*+

𝑦",( + 𝐵",0

1

0*+

𝑋",0 + 𝜀"

where the 𝑋",0 vectors contain the exogenous variables and their lags, and the B matrices contain the coefficients respectively. The variables in the X vector affect the state of the other variables, but are not themselves determined by the system of equations, and thus are considered exogenous. In addition, the present model adds a cross-sectional dimension to the basic VAR(p) model by essentially stacking the VAR models for the different states on top of each other. In other words, the panel VAR-X model is:

65 Lutkepohl, H. (2005, Spring). New introduction to multiple time series analysis, Chapter 2.


𝑦+," = 𝑎+,% + 𝐴(

)

(*+

𝑦+,",( + 𝐵",0

1

0*+

𝑋",0 + 𝜀+,"

⋮

𝑦.," = 𝑎.,% + 𝐴(𝑦.,",(

)

(*+

+ 𝐵",0

1

0*+

𝑋",0 + 𝜀𝑀, 𝑡

See Canova and Ciccarelli for an excellent survey of panel VAR methods.66 One thing of note is that in the present study we do not focus on estimating dynamic heterogeneities between the different state economies as is done in many panel VAR models for large macroeconomic studies. Instead we focus on the average effect of an RPS enactment on a state economy. To that end, we estimate the model with fixed effects to control for possible heterogeneities across states.67 It is possible to recover the state fixed effects, though we make no effort to do so here as the focus of the study is on the average effect of an RPS and not on individual state effects.

0.2. MODEL ESTIMATION

We use Bayesian techniques, namely the Gibbs sampler, to estimate the structural panel VAR-X model. See Ciccarelli and Rebucci for a review of Bayesian methods for VAR models.68 See also Ocampo and Rodríguez for a very practical tutorial.69 We follow their Algorithm 3 for Bayesian estimation, which for simplicity we reproduce below.

66 Canova, F., & Ciccarelli, M. (2013). Panel vector autoregressive models: A survey.

67 Greene, William. (2012). Econometric Analysis. Upper Saddle River, NJ: Prentice Hall.

68 Ciccarelli, M., & Rebucci, A. (2003). Bayesian VARs: A survey of recent literature with an application to the European monetary

system. 69

Ocampo, S., & Rodríguez, N. (2012). An introductory review of a structural VAR-X estimation and applications. Revista Colombiana de Estadística, 3, 479-508.


Algorithm 1 Bayesian Estimation

1. Select the specification for the reduced form VAR-X, that is to choose values of p (endogenous variables lags) and q (exogenous variables lags) such that the residuals of the VAR-X (𝜀) have white noise properties. With this the following variables are obtained: T, p, q, k, where:

𝑘 = 1 + 𝑛𝑝 +𝑚 𝑞 + 1

2. Calculate the values of Γ, S with the data (Y , Z) as:

Γ = (𝑍@𝑍),+𝑍′𝑌 𝑆 = (𝑌 − 𝑍Γ)′(𝑌 − 𝑍Γ)

3. Generate a draw for matrix Σ from an inverse Wishart distribution with parameter S and T − k degrees of freedom.

Σ~𝑖𝑊)JK 𝑆, 𝑇 − 𝑘

4. Generate a draw for matrix Γ from a multivariate normal distribution with mean Γ and covariance matrix Σ 𝑍@𝑍 ,+

Γ|Σ~𝑀𝑁)JK(Γ, Σ⨂ 𝑍@𝑍 ,+)

5. Repeat steps 2-3 as many times as desired, save the values of each draw.

The draws generated can be used to compute moments of the parameters. For every draw the corresponding structural parameters, impulse response functions, etc. can be computed, then their moments and statistics can also be computed. The algorithm for generating draws for the inverse Wishart and multivariate normal distributions are presented in Bauwens et al., Appendix B.70 Observe that in this notation:

70 Bauwens, L., Lubrano, M., & Jean-Francois, R. (2000). Bayesian inference in dynamic econometric models. Oxford, UK: Oxford

University Press.


,

,

and finally,

Then the VAR-X model can be written simple as

𝑌 = 𝑍Γ + 𝐸 We set 𝑝 = 3 and 𝑞 = 1 for simplicity.

0.3. DYNAMIC MULTIPLIER ANALYSIS

During the Gibbs sampling simulation, which we run for 5, 000 replications with 500 burn-in steps, we also conduct dynamic multiplier analysis for the exogenous RPS policy variable. We follow Algorithm 2 in Ocampo and Rodrıguez to conduct this analysis.71 This algorithm is as follows:

71 Ocampo, S., & Rodríguez, N., op. cit.

Y =

y '1!y 't!y 'T

!

"

#######

$

%

&&&&&&&

Z =

1 y '0 ! y '1−p x '1"1 y 't−1 ! y 't−p x 't"1 y 'T−1 ! y 'T−p x 'T

"

#

$$$$$$$

%

&

'''''''

E =

ε '1!ε 't!ε 'T

!

"

#######

$

%

&&&&&&&

Γ = v A1 … Ap B0 ! Bq"#$

%&'


Algorithm 2 Identification by Long-Run Restrictions

1. Estimate the reduced form of the VAR-X model. 2. Calculate the VMA-X representation of the model (matrices 𝛙T ) and the convariance matrix of the

reduced form disturbance 𝜀 (matrix Ʃ). 3. From the Cholesky decomposition of 𝛙(1) and Ʃ𝛙(1) calculate matrix C(1)

𝐶 1 = chol 𝛙 1 Ʃ𝛙@ 1

4. With the matrices of long run effects of the reduced form, 𝝍(1), and structural shocks, C(1), calculate the matrix of contemporaneous effects of the structural shocks, 𝐶%.

𝐶% = 𝛙 1 ,+𝐶 1

5. For i = 1, . . . , R with R sufficiently large, calculate the matrices 𝐶T as:

𝐶T = 𝛙T𝐶%

Identification is completed since all matrices of the structural VMA-X are known. We set R = 120 months after an RPS to estimate the cumulative, or long-run effects of an RPS enactment for dynamic multiplier analysis.


Appendix B

GRAPHICAL ANALYSIS OF THE DYNAMIC MULTIPLIER ANALYSIS

Below we present in graphical form the dynamic multiplier analysis for each of the five state macroeconomic variables. This analysis strengthens the evidence of a severely deleterious effect of an RPS policy. For electricity sales, real personal income, and non-farm employment the response to an RPS is an initial sharp decline lasting for several years and the long-run effect is a large and lasting decline. Manufacturing demonstrates the same initial sharp decline in response to an enacted RPS, but does show some recovery, after several years, though still never returns to levels prior to the RPS. However, the unemployment rate demonstrates a steadily increasing rate that cumulates into a large increase in state unemployment.


Dynamic Multiplier Analysis for Electricity Sales.


Dynamic Multiplier Analysis for Real Personal Income.


Dynamic Multiplier Analysis for Non-farm Employment.


Dynamic Multiplier Analysis for Manufacturing Employment.


Dynamic Multiplier Analysis for Unemployment Rate.

RANDY T SIMMONS PH.DUTAH STATE UNIVERSITY

TYLER BROUGH PH.DUTAH STATE UNIVERSITY

PRINCIPAL INVESTIGATORS

JACOB FISHBECKSTRATA POLICY

RENEWABLE PORTFOLIO STANDARDS: PENNSYLVANIApassed Pennsylvania’s first Renewable Portfolio...

Documents

Transcript of RENEWABLE PORTFOLIO STANDARDS: PENNSYLVANIApassed Pennsylvania’s first Renewable Portfolio...