Great Britain’s carbon emissions from electricity generation fell by two-thirds between 2012 and 2019, providing an important example for other nations. This rapid transition was driven by a complex interplay of policies and events: investment in renewable generation, closure of coal power stations, raising carbon prices and energy efficiency measures. Previous studies of the impact of these simultaneous individual measures miss their interactions with each other and with exogenous changes in fuel prices and the weather. Here we use Shapley values, a concept from cooperative game theory, to disentangle these and precisely attribute outcomes (CO2 saved, changes to electricity prices and fossil fuel consumption) to individual drivers. We find the effectiveness of each driver remained stable despite the transformation seen over the 7 years we study. The four main drivers each saved 19–29 MtCO2 per year in 2019, reinforcing the view that there is no ‘silver bullet’, and a multi-faceted approach to deep decarbonisation is essential.
Electricity market manipulation enforcement actions have moved from conventional analysis of generator market power in real-time physical markets to material allegations of sustained crossproduct price manipulation in forward financial markets. A major challenge is to develop and apply forward market analytical frameworks and models. This task is more difficult than for the real-time market. An adaptation of cross-product manipulation models from cash-settled financial markets provides an existence demonstration under uncertainty and asymmetric information. The implications of this analysis include strong empirical predictions about necessary randomized strategies that are not likely to be observed or sustainable in electricity markets. Absent these randomized strategies and other market imperfections, the means for achieving sustained forward market price manipulation remains unexplained.
Keywords: market manipulation; electricity markets; limits to arbitrage; asymmetric information
This paper examines carbon tax design options in the United States using an intertemporal computable general equilibrium model of the world economy called G- Cubed. Four policy scenarios explore two overarching issues: (1) the effects of a carbon tax under alternative assumptions about the use of the resulting revenue, and (2) the effects of a system of import charges on carbon-intensive goods (“border carbon adjustments”).
Summary Efficient electricity day-ahead market designs include virtual transactions. These are financial contracts awarded at day-ahead prices and settled at real-time prices. In PJM these virtual transactions include incremental offers (INCs) that are like generation offers, decremental bids (DECs) that are like demand bids, and up-to-congestion bids (UTCs) that are like transmission price spread bids. Virtual transactions offer potential benefits to improve the efficiency of electricity markets, mitigate market power, enhance price formation, hedge real-time market risks, and price those risk hedging benefits.
The role and performance of virtual transactions has been a subject of controversy. A report by PJM addresses some of these controversies, identifies possible problems in the present implementation of virtual transactions with the associated settlement rules, and makes recommendations for changes in the treatment of virtual transactions. The PJM report is generally supportive of the contribution of virtual transactions as improving overall market performance. Illustrative examples in the report highlight these contributions and add to the general understanding of the benefits and some of the problems with its current rules for treating virtual transaction.
Although these examples help in explaining the mechanics of virtual transactions, and the interactions with the underlying physical market, the examples do not provide a framework for evaluating the overall cost and benefits of virtual transactions. The PJM analysis is not alone in this regard, because the evaluation task is not easy. There is no readily available template waiting to be applied to the PJM case. The limited available analyses from other regions indicate that the benefits are material and outweigh the costs, but no available studies cover all the relevant issues. However, going beyond examples of particular outcomes to consider, the broader context is important. Looking to the broader framework can change both the diagnosis of the symptoms and the prescriptions for the cures.
Under the current PJM market rules, there is an asymmetry in the settlement treatment of different types of virtual transactions, applying residual uplift charges to INCs and DECs but not to UTCs. One of the PJM recommendations is to eliminate this asymmetry by extending the same uplift treatment to UTCs. The argument is based on allocation of uplift costs according to the deviations between real-time quantity and day-ahead schedules. This approach is particularly problematic for virtual transactions, which by design involve a 100% deviation.
There is no simple connection between deviations, uplift costs and market efficiency. Under a broader equilibrium analysis there can be conditions where there is no relationship between any of these components. Furthermore, the allocation of properly defined residual costs according to a cost causation argument can in itself be a contradiction. More importantly, the focus on uplift cost causation is misplaced. The important question is the aggregate net benefit of virtual transactions, not the residual cost. If virtual transactions increase the net benefits in the market, then there is no incentive-based reason to assign additional costs to virtual transactions. The iii criterion for assigning residual costs would then turn to doing the least damage to the performance of the market.
A better symmetric solution is to avoid any uplift allocation to virtual transactions. The residual cost allocation would then apply to real load; liquidity and entry in financial day-ahead virtual transactions would be enhanced; market power would be reduced; accurate price formation would be supported; and the efficiency of the overall PJM electricity market system should be improved. This reversal of the conventional wisdom follows from a broader framework than that applied by PJM for consideration of the costs and benefits of virtual transactions.
This broader framework builds directly on the basic principles of efficient electricity market design. Stepping back to consider first principles makes it easier to see the connections among the components of market design, in order to consider the function and benefits of virtual transactions from the perspective of aggregate market performance. PJM’s own analysis provides many examples of the contributions and effects of virtual bidding, but does not connect the examples to the broader framework of electricity market design principles. Furthermore, going beyond the uplift allocations, the PJM recommendations restricting the use of virtual transactions do not follow necessarily even from a narrower evaluation perspective. The principal problem PJM identifies with virtual transactions is a computational burden that would be only indirectly affected by uplift allocations, and could be addressed through other means with fewer negative consequences for the broader market design, such as by continuation of bidding budgets that allowed flexibility in the choice of virtual transactions.
Restricting explicit virtual bidding, as PJM proposes, creates market power for those who can make implicit virtual bids. Explicit virtual bidding mitigates or eliminates this market power, provides liquidity, improves price formation, allows hedging, connects naturally with longer term financial transmission rights, helps reveal defects in market design, and on average should improve system operations.
The PJM report appears in a context where virtual bidding is under attack. While a complete cost benefit analysis is not available, the PJM analysis can be expanded to enhance both the understanding of the role of virtual bidding and the policies that support overall electricity market efficiency.
In one of the first laws establishing regulation of the electric utility industry, the Federal Power Act of 1920 (FPA) there was a recognition that there were two types of transactions commonly entered into in the industry that would be subject to regulation - with a different regulatory regime for each. Retail sales, or sales directly to customers who consumed the power themselves were deemed to be intra-state sales to be regulated by the states. But any sale for resale, i.e. a sale from any generating entity to a second entity that resold the power, was deemed to be an inter-state sale subject to regulation by the Federal Energy Regulatory Commission (FERC). This paper deals with the latter type of electric sale – wholesale sales regulated by the federal government. Up until the mid-1990’s, most wholesale sales were between vertically-integrated and state-franchised utilities, either short-term to take advantage of one utility having cheaper power at a moment in time than another utility, or longer term to provide needed capacity to the purchasing utility. Both of these types of transactions were mostly conducted under bilateral contracts between the buyer and the seller – the contracts being submitted to FERC for approval according to the statutory framework of the FPA.1 Until the mid-1990’s, short-term transactions were typically conducted on a split-savings basis, meaning the savings resulting from the transaction were evenly split between the buyer and seller. Longer-term transactions were typically cost-based, with the seller allowed to earn a regulated return on the sale.
For a variety of reasons beginning in the mid-1990’s there was a development of a new type of market, made possible by the deregulation or restructuring process which for the first time allowed retail customers in some states to choose their electric supplier. It was thought at the time that effective retail competition required utilities to divest all or some of their power plants to third parties. At the same time, changes in Federal law and regulation were making it considerably easier for third parties to enter generation markets and have guaranteed access to utility transmission systems. Thus new wholesale markets began to be developed in many regions of the country
Because of concerns about fairness, these new markets formed around independent system operators or regional transmission organizations independent of the transmission and generation owners in their regions. These regional operators also adopted a new form of wholesale market for their regions, a centralized market based on bids submitted to the market operator from individual generators. These bid-based centralized markets utilized locational marginal pricing (“LMP”), whereby generators bid at their location into a centralized market and bids are accepted or rejected based on projected electricity needs for the relevant period. While
generators are dispatched from lowest-cost bid to highest-cost bid up until the point that expected demand plus a reasonable margin is satisfied (and reliability constraints are recognized,) all successful bidders receive the highest priced successful bid at their location.
Another feature of these new LMP markets is that rather than charging for transmission service based on a contract path, users of the transmission system were to pay congestion charges based on the difference in locational prices between the point of injection and the point of receipt (i.e., the location of the seller and the location of the buyer). Market participants were either allocated, or had to buy through auctions, so called financial transmission rights (“FTRs”) which give them rights to use the transmission system without paying congestion charges. In this way, market participants could hedge their transactions by owning FTRs.
The theoretical basis of LMP markets is that individual generators bid into the market at their marginal cost (the cost of producing their next kilowatt-hour) because to bid less would result in their losing money if they were to win the bid and have to generate and to bid more might mean that they don’t get dispatched even though the transaction would be profitable to them. The market operator (RTO or ISO) chooses winning bidders based on the lowest cost combination of bids that can be dispatched in real time within reliability constraints. Thus, in theory, generators presumably will have incentives to operate as efficiently as possible, because only the lowest cost generators get paid, and their profitability depends on getting dispatched and having costs below the LMP. Profits are simply the difference between the LMP paid to all generators at a given location and the generators actual cost for the period for which its bid was submitted. These bid-based LMP markets are most often referred to as “organized” wholesale markets or “centralized” wholesale markets. This paper refers to centralized markets, as the term “organized’ gives a false impression that other markets types are not organized.