The Performance of U.S. Wind and Solar Generators.

AuthorSchmalensee, Richard
PositionReport - Statistical data
  1. INTRODUCTION

    The rapid growth of wind and solar electricity generation globally in recent decades has been driven in large part by a diverse set of subsidies and regulations. In the U.S. these have involved all levels of government, and they have primarily rewarded investment in generation facilities and megawatt-hours (MWh) generated. (1) The main federal subsidy for solar generation has been a 30% investment tax credit, and both solar and wind facilities have been eligible for accelerated depreciation. These and similar investment-based state and local tax credits clearly provide weaker incentives to minimize production costs than generation-based subsidies.

    The main federal subsidy for wind generation has been a $23/MWh production tax credit (PTC). (2) At the state level, 29 states and the District of Columbia have Renewable Portfolio Standard (RPS) programs that generally require affected utilities to purchase specified quantities of renewable energy credits (RECs), which are produced in proportion to their outputs by certified generating facilities powered by renewable energy. To be certified in half of these programs, a generator must deliver power in the state in question, and the other 15 programs either limit certification to nearby generators or provide incentives for in-state generation. (3) Thus the PTC and each state's RPS program treat all certified wind-generated MWh as equally valuable regardless of when and where within the state or region they are produced; all certified solar-generated MWh are treated similarly; (4) and RPS programs constrain the siting of wind and solar generators.

    Ignoring the facts that the marginal value of electricity varies substantially over time and space and that sites for wind and solar generators vary on several dimensions of quality simplifies these policies but also reduces their efficiency. It is not clear a priori that this is a bad tradeoff, of course. If the PTC and state RPS program also covered nuclear plants, for instance, it would seem unlikely that site choice would affect performance much or that these baseload plants would be able to be very responsive to changes in the value of electricity over time except, perhaps, via scheduling maintenance outages.

    Wind and solar plants are different from nuclear plants, of course. Their performance on several dimensions is affected by the weather and thus by location, as well as by operator decisions. Some places are obviously sunnier than others and some are windier, but other aspects of weather also matter. Since the spot market price of electricity varies over time, the average social value of wind and solar generators are importantly affected by when the wind blows and when the sun shines at their sites, as well as by when facility maintenance is scheduled. Moreover, the outputs of wind and solar facilities are intermittent--variable over time and imperfectly predictable--so the observed output of a solar or wind facility over time is a realization of a stochastic process. Both the variability of that process and its correlations with electricity demand and the output of other intermittent facilities will affect the facility's contribution to system capacity and the costs of integrating it and facilities like it.

    The importance of the variation over space and time along multiple dimensions of wind and solar generators' performance has implications for academic research, market design, and public policy. Can wind generators' performance in a single region be treated as representative of the nation as a whole? Should all solar plants receive identical treatment in a regional capacity market? Are the costs of constraining the siting of renewable generators acceptably small? Developing the factual basis necessary to address these and other related questions requires quantitative analysis of the performance of wind and solar generators within and between multiple U.S. regions and over time. This paper presents a first such analysis.

    When essentially all US generating plants were owned by regulated public utilities, data on their performance were publicly available and were employed in numerous empirical studies. Many wind and solar facilities are now unregulated and operate in competitive markets, however, and data on the operation of those facilities are commonly treated as proprietary and confidential. This study employs confidential data supplied by the seven U.S. Independent System Operators (ISOs) on hourly output and nodal spot prices for a sample of 25 wind and 9 solar (photovoltaic) generating plants across the U.S. for 2011 and up to 12 adjacent months. In order to maintain confidentiality most characteristics of those facilities were not provided. This essay presents an analysis of the performance of those 34 plants. As far as I know, this study is the first to employ data of this sort from all seven U.S. ISOs and thus to examine performance differences both within and between regions as well as, to a limited extent, over time.

    Section 2 provides a brief description of the data used in this study. More detail is provided in Sections A.1 and A.2 of the Appendix. Taking advantage of our uniquely rich price data, Section A3 of the Appendix provides new information that may be of independent interest on the distributions of spot prices over time at individual network nodes and across space within ISOs. Nodal price distributions are generally right-skewed and have considerably fatter tails than normal distributions, and prices often differ substantially within some ISOs.

    Section 3 considers the distributions of capacity factors among the generators in our sample and of their value factors: the ratios of the average spot-price value of these generators' outputs to the unweighted average spot prices they faced. Variation in cross-section and over time in capacity factors and in wind generators' value factors are substantial. Some correlates of the cross-section differences in value factors are discussed, and evidence consistent with their expected decline over time is presented. Finally, this section explores generators' reactions to the negative spot prices that all units outside ISONE faced in 2011 and shows that negative prices generally did not induce generators to reduce output. (5)

    Section 4 presents evidence on three quantities related to the intermittency of wind and solar generation: capacity factors during hours of high spot prices, hour-to-hour and day-to-day variability in output at the facility and ISO levels, and the incidence of low or zero generation. Capacity factors during high-price periods are found to vary substantially over space and time. Measures of variability are developed to deal with the predictable diurnal changes in the output of solar facilities, and a new measure of the gains from geographic averaging across facilities is presented and employed. Variability differs substantially in cross-section, as do gains from geographic averaging. Spells of zero ISO-level wind output appear much more common in the Northeast than elsewhere. Both inter- and intra-regional differences are important, and some quantities vary substantially over time.

    This analysis reveals that there is a great deal of variation around overall and regional means of many performance measures, and some vary substantially over time. Some of that variability reflects the poor design of current U.S. subsidy policies, which also leads to socially inefficient operating decisions. Section 5 provides a brief discussion of some implications of our substantive findings for research, market design, and public policy.

  2. DATA EMPLOYED

    As Joskow (2011), Borenstein (2012), and others have stressed, in the absence of identifiable externalities, the best measure of the marginal social value of the output of any particular generator is given by the location-specific spot prices that generator faces. Unfortunately, in the U.S. location-specific spot wholesale prices exist only in the regions served by the seven Independent System Operators (ISOs), which manage organized wholesale electricity markets and regional transmission systems. (6) These systems meet around 2/3 of U.S. electricity demand and serve around 2/3 of U.S. electricity customers.

    The sample of wind and solar generation plants analyzed here is accordingly drawn from those systems, and data on their hourly outputs and the corresponding spot prices were kindly provided by all of the seven U.S. ISOs: (7)

    The Electric Reliability Council of Texas (ERCOT), which serves most of Texas. ISO-New England (ISONE), which serves the six New England States.

    The Midcontinent ISO (MISO), which serves North Dakota, Minnesota, and Iowa, as well as most of South Dakota, Illinois, and Indiana, and small parts of several adjacent states.

    The New York ISO (NYISO), which serves New York State

    The PJM Interconnection (PJM), which serves Pennsylvania, New Jersey, Maryland, Delaware, Virginia, West Virginia, and the District of Columbia, as well as most of Ohio and parts of Illinois, Indiana, and other adjacent states.

    The Southwest Power Pool (SPP), which serves Nebraska, Kansas, and Oklahoma, as well as parts of Texas, New Mexico, and other adjacent states.

    The California ISO (CAISO), which serves most of California.

    For all but SPP, the spot price data are Locational Marginal Prices (LMPs) or nodal prices for the network nodes at which each generator in the sample is located. These LMPs are defined as the short-run marginal cost of meeting an additional MWh of demand at the node in the transmission system at which the generator is located, taking into account transmission losses, transmission line capacity constraints, and the (as-bid) costs of incremental generation. (8) The SPP prices are not LMPs since they do not take into account transmission losses, but they are the spot prices each generator in fact faced. (As this was written SPP was in the process of moving to a...

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT