Capacity
allocation
in
cross-border
intraday
electricity trading: Should future intraday trading
arrangements make cross-border capacity available
for over-the-counter (OTC) transactions or not?
© 2010 E-Bridge Consulting GmbH ii 07 October 2010
Contents
1 Executive Summary ... 1
2 Introduction ... 3
2.1 The role of transmission capacity in electricity trading ... 3
2.2 The advent of liberalisation and cross-border power trading ... 3
2.3 Capacity allocation ... 4
2.4 Capacity at different time horizons ... 4
2.5 The inefficiency of explicit capacity allocation ... 5
2.6 Approaching real time ... 5
2.7 Intraday electricity trading ... 5
2.8 Moving towards more efficient methods for allocating capacity ... 6
2.9 Current developments with respect to intra-day markets... 7
2.10 Outlook on the report ... 8
3 Objective of the present report ... 9
3.1 Scope ... 9
3.2 Terms used... 10
3.3 Stylized trading arrangements - common features ... 10
3.3.1 Capacity management is a natural monopoly ... 11
3.3.2 Design options regarding shared order mechanism ... 11
3.3.3 Operation of within-bidding-area intraday OTC markets unchanged ... 12
3.4 Stylized trading arrangements - with OTC trade ... 12
3.5 Stylized trading arrangements - Excluding OTC trade ... 14
3.6 Digression: implementing OTC trades via an exchange-based trading system such as ELBAS ... 15
3.7 Summary ... 16
4 Earlier work on the question of interest ... 17
4.1 Regulation (EC) No 714/2009 ... 17
4.2 Congestion management guidelines ... 17
4.3 CWE Regulators' Position ... 18
4.4 Project Coordination Group (PCG) Proposal on Target Model ... 19
4.5 Draft Framework Guideline on Capacity Allocation and Congestion Management ... 19
© 2010 E-Bridge Consulting GmbH iii 07 October 2010
4.7 Summary ... 20
5 Criteria for evaluation ... 21
5.1 Approach ... 21
5.2 Maximisation of social welfare ... 22
5.2.1 Efficient utilisation of cross border capacity... 22
5.2.2 Market based allocation of additional intraday capacity ... 25
5.2.3 Implicit maximisation of social welfare by facilitating trades ... 27
5.2.3.1Technical constraints ... 27
5.2.3.2Lack of liquidity on the exchange-based market ... 29
5.2.3.3Increase in liquidity on the exchange-based market ... 30
5.2.4 Effect on intensity of competition in generation ... 32
5.2.4.1OTC trade is more useful to larger market parties ... 32
5.2.4.2Does explicit capacity allocation make it easier to hoard capacity? ... 32
5.2.4.3Transaction costs narrowly defined ... 33
5.2.4.4Search costs ... 33
5.2.4.5Ease of monitoring ... 34
5.2.5 Effect on intensity of competition between trading venues ... 34
5.2.6 Cost of implementation and operation... 35
5.3 Ease of use of the solution ... 36
5.3.1 Accessibility for all market parties ... 37
5.3.2 Visibility / tradability of all offers Europe-wide ... 37
5.3.3 Complexity and number of systems: user-friendly and "single screen" ... 37
5.3.4 Transparent for the market parties ... 38
5.3.5 Quickly and easily implementable consistently across all EU ... 40
5.3.6 Robustness and possible fallback solutions ... 41
5.4 Efficiency in solving the balance of market parties (hedging risk) ... 41
6 Conclusion and recommendation ... 43
7 APPENDIX 1: Project Coordination Group's Target Model for the Intraday Market... 44
8 APPENDIX 2: Gauging the loss due to impaired market functioning ... 45
9 APPENDIX 3: Estimating capacity lost due to unscheduled outages ... 47
1
E
XECUTIVES
UMMARYAn exchange-based trading system that makes cross-border capacity available implicitly will be at the core of European intraday electricity markets in future. The present report addresses the question whether - in addition to implicit allocation - cross-border capacity should also be made available explicitly for over-the-counter (OTC) transactions. The time horizon being considered is the near-term future up to around 2015. There is little point in attempting to make a recommendation for the longer term. Such a recommendation should not be made now, but rather should be made in the light of future experience.
Based on the findings described in this report - which draw on a desk study and in-depth discussions with stakeholders - we come to a clear answer to the question and thus recommendation: an exchange-based trading system offering a block-trading functionality similar to that provided today by the ELBAS trading system would be an adequate solution; direct access to capacity for market parties is not required. This recommendation is qualified somewhat in that the loss in social surplus due to unavailability of OTC trade could not be established with accuracy because our interviewees represented only a sample that cannot be fully representative of the entire market. Based on our research, we have come to the conclusion that this loss is small and we have not seen convincing evidence to the contrary. However, we urge regulators to listen to market parties who may provide such evidence. In order to answer the above question, the report compares two stylized trading arrangements - one featuring only an exchange-based trading system, the other also allowing for OTC trade via direct access to an explicit capacity platform. We evaluate the two trading arrangements against a number of criteria that we discussed in detailed interviews with a number of stakeholders. The criteria are the following
• the efficiency with which cross border capacity is utilised
• the ease with which a market based allocation of additional intraday capacity can be implemented
• the extent to which trades are facilitated and social surplus is thus maximised
• the effect on the intensity of competition in generation as well as on competition between trading venues
• the cost of implementation and operation
• accessibility and visibility of offers; complexity and user-friendliness; transparency • the speed with which the trading arrangement would be consistently implementable
across Europe and • robustness.
One finding is that a purely exchange-based trading system makes it more likely that intertemporal inefficiencies are reduced via arbitrage. In other words, if capacity allocated to one transaction later turns out to have higher value if allocated to another transaction, we believe that this correction would be more likely to occur if all trade were concentrated on an exchange-based trading system.
would be left with the inefficiencies arising from the separate purchasing of capacity and energy (the primary reason why day-ahead capacity will in the future be allocated implicitly). The block-bid functionality of the exchange-based trading system ought to be sufficient to cover unexpected outages until the exchange-based market has become sufficiently mature and liquid. A general statement on whether trade should be centralised is beyond the scope of the present study. However, concentration of trading in an exchange-based trading system likely leads to increased liquidity and lower spreads.
An exchange-based trading system is more conducive to competition in generation. It is inherently less prone to hoarding of capacity; search costs are lower; and it should also be easier to monitor. Direct (explicit) transaction costs are higher trading via an exchange-based system and we would ask regulators to ensure that these do not inappropriately burden smaller market parties. In a similar vein, should all cross-border intraday trading take place on an exchange-based system, appropriate safeguards need to be in place that compensate for the lack of competition from OTC trading.
Setting up and / or maintaining a parallel explicit capacity allocation channel will lead to additional costs of implementation and operation. We would not expect these costs to be prohibitive, but they ought to be taken into account.
An exchange-based trading system ought to be more easily accessible to all market parties and it ensures maximum visibility and tradability of offers regardless of the bidding area in which load and generation are physically connected to the grid (subject to availability of capacity, of course). It is less complex and more user-friendly and would conveniently centralise information in a single venue ("single screen"). Conversely, a system offering continued access to explicit capacity via a capacity platform would not be as quickly and easily implementable consistently across all of Europe because of the additional complexity such a setup entails. As an upside to this complexity, such a setup offers a ready-made fallback solution in case the exchange-based platform ever fails.
The most important argument for keeping OTC trade is that some transactions have characteristics that are so non-standard that it would not be possible to implement them via exchange-based trading, which has to rely on products that are always to some extent standardised. Based on our research and interviews we believe that this argument has substance in theory, but that the volume of trading that would be lost in practice would be tolerable, provided that a block-bid functionality along the lines of the ELBAS system were available. However, as noted above, regulators should be open to the possibility that market parties present convincing evidence to the contrary.
2
I
NTRODUCTIONThe present chapter is meant to provide an introduction to the subject. It also puts the question the report aims to answer into a broader context. That broader context is the process of increasing integration of Europe's markets for electricity. Even ten years ago a largely monopolistic, vertically integrated industry, the European electricity sector has seen considerable change in the recent past towards more integration and thus more competition. Much of this was driven by European legislation.
2.1 THE ROLE OF TRANSMISSION CAPACITY IN ELECTRICITY TRADING
In order to trade electricity one has to transport it via the transmission network. The transmission grid, however, has only a limited capacity. Historically, European power systems developed as national islands, although from the very beginning of the then-UCTPE1 interconnections were used to optimise dispatch across national borders. Apart from - initially quite limited - commercial exchanges, the interconnections also served the purpose of mutual support by transmission system operators (TSOs) and stabilisation of the interconnected power system.
2.2 THE ADVENT OF LIBERALISATION AND CROSS-BORDER POWER TRADING
As liberalisation opened up the electricity sector in the European Union to competition, electricity increasingly turned into a regular commodity with power plant operators trying to sell the power at the highest possible price - which may well be paid abroad.
If the price of electricity is low in area A and high in area B, the power plant owners in area A would want to sell power at the more advantageous price of area B. If such exports were not limited, the resulting load flows might exceed the available capacity and the overloaded grid elements would eventually trip (i.e., be taken out of operation in order to prevent lasting damage to the equipment).
Therefore in order to ensure the safe operation of the power system, load flows are carefully monitored by the system operators and the capacity on the borders2 is managed.
1
"Union pour la coordination de la production et du transport de l’électricité", the interconnected synchronous electricity transmission system in Western Continental Europe; in 1999 renamed "Union for the Co-ordination of Transmission of Electricity" (UCTE); the UCTE was wound up and became part of the "European Network of Transmission System Operators for Electricity" (ENTSO-E) with effect from 01 July 2009
2
2.3 CAPACITY ALLOCATION
The capacity on the interconnections is usually scarce in the sense that if it were given away for free, demand would exceed the available capacity. Simplifying slightly, the reason for this is that the capacity holder is able to take advantage of the price difference between the two markets; i.e., buy power in the low price area, sell it in the high price area, and lock in the difference as a profit3. The public policy objective in making this capacity available to market parties must be to maximize the social welfare that the gains from trade create for society. In concrete terms, such gains from trade are obtained, for example, when a power plant with low marginal cost substitutes a more expensive one.
In practical terms, the typical4 approach to maximizing social welfare has been to sell the capacity at auction so that those market parties who value it the most can obtain it. The sellers of the capacity are typically the transmission system operators (TSOs) who own / operate the grid, but they have to follow fairly restrictive European and national rules on how they can sell interconnection capacity and what they are allowed to do with the proceeds (see documents cited in footnote 14).
2.4 CAPACITY AT DIFFERENT TIME HORIZONS
Like energy itself capacity is traded at different time horizons at present. For the purposes of the present report it is acceptable to think of the capacity available on an interconnection as a fixed number in terms of megawatts [MW], the available transfer capacity (ATC). In reality the way power flows in a power system is quite complicated and the actual availability will often differ significantly from the stated value. The available capacity is usually divided into three maturities and sold at auction. At the long-term auctions, capacity is typically auctioned for one year resp. one month at a time. A successful bid for one MW of yearly capacity entails the right to transport one MW of power in one direction for a whole year. The monthly auction functions in the same manner, but the capacity is only given out for one month at a time. Finally, some capacity may also be made available to the market via an auction on D-1, the day prior to delivery5.
3
Note that this reasoning implies that there are zonal prices - i.e., different price zones for each of which is established a separate price.
4
In geographical terms, Western Continental Europe should be thought of as the primary reference area for this report. Practices are not uniform across this area but they are often sufficiently similar to allow a degree of generalisation.
5
2.5 THE INEFFICIENCY OF EXPLICIT CAPACITY ALLOCATION
The greatest drawback of this type of arrangement, known as an explicit auction because market parties bid for capacity and energy separately, is that empirical evidence shows that quite frequently this leads to adverse flows - power flowing from the high-price area to the low-price area. For example, a generator located in area A may expect the energy price in A to be EUR 50 / MWh and the energy price in area B to be EUR 70 / MWh. If the market party has obtained capacity for transporting power from A to B, then given the expectations it looks like a good deal to sell power from A to B. However, assume that the prices ultimately realised are EUR 50 / MWh in area A and EUR 40 / MWh in area B, respectively. Clearly, it would have been both more profitable and more efficient to sell in area A rather than B. There is a social loss because a more expensive power plant was running even though it could have been replaced by a cheaper one in the other area. In section 2.8 below we briefly discuss how this problem of adverse flows can be avoided.
2.6 APPROACHING REAL TIME
"D-1" is a very important time in electricity trading. The reason for this is that in any electric power system, at any moment in time, power consumed has to be approximately equal to power generated and conversely. Since a large deviation could ultimately trigger a black-out, maintaining that balance is very important and is one of the key tasks of the system operator. However, the first line of defence against aggregate imbalances is the obligation or incentive for each market party to be balanced at the planning stage. In principle every market participant (or balance-responsible party) is responsible for generating (or purchasing) the same amount of power that it consumes (or sells). This is not easily done close to real time; it requires planning (forecasting, finding suitable counterparties etc.) and so most electricity is traded on the so-called day-ahead market or via long-term contracts.
Generally speaking, electricity is traded either on an exchange (where trading is typically anonymous) or over-the-counter (we will refer to this as "OTC" throughout the report; the parties will often know each other and agree on trades bilaterally via telephone although any particular transaction might be arranged by a broker).
A fairly typical arrangement is for a power exchange to have an auction at 11am or noon for a number of products for delivery on the following day (individual hours, blocks, base, peak...). In other words interested buyers and sellers submit offers to buy and sell and after the gate closure (in this case, the power exchange's matching gate closure) these bids are matched and the resulting transactions (successful matches) are scheduled for the next day.
2.7 INTRADAY ELECTRICITY TRADING
day-ahead gate closures of relevance: the first being the time at which the power exchange6 stops accepting bids for the day-ahead auction; the second is the time at which the local transmission system operator ceases to accept nominations (schedules) for within-area OTC transactions for the next day (normally somewhat later in the afternoon). Shortly after the second gate closure, the market reopens and it becomes possible, once again, to buy and sell power for delivery until the end of the following day7.
While most energy is traded on the day-ahead market, it is clearly useful to be able to adjust one's position as more information about likely consumption, meteorological conditions etc becomes available. In fact, trading volumes on the intraday markets have generally increased substantially in recent years and with an ever-greater share of generation being provided by intermittent, hard-to-predict energy sources such as wind, intraday trading is bound to become more and more important.
Cross-border trade exists at the intraday timescale even today (though not yet on all interconnections). The standard arrangement in Continental Europe is explicit allocation of intra-day capacity for free, albeit with mandatory use. The details of the allocation rules vary a lot by border (e.g., on some interconnections matching gates are used; on others trading is continuous). In the Nordpool area the intraday market ELBAS is part of Nordpool Spot and bids submitted in one price area will be automatically considered in other price areas if sufficient capacity is available. Three additional current initiatives will be discussed in section 2.9 below.
2.8 MOVING TOWARDS MORE EFFICIENT METHODS FOR ALLOCATING CAPACITY
In the above it was noted that auctioning capacity explicitly (separately) is inefficient because this would sometimes cause adverse flows (energy sent from the high-price area to the low-price area). One way to avoid this problem is to auction capacity and energy together; a so-called implicit auction because market parties only need to submit energy bids to the power exchange in their local price area. The power exchanges (or an auction office) see to it that the capacity is used in an efficient manner8 via a suitable matching algorithm.
As far as the day-ahead markets are concerned, the implicit auction model sketched out above is likely to be the future target design for, eventually, the entire European Union. One group that has been contributing to this work is the so-called Project Coordination Group
6
In most bidding areas there is only one power exchange. Great Britain (APX Power UK; N2EX) is one example of an exception; Germany is another (Epex Spot; NPS).
7
one way to think of the interruption is that the TSO needs some time to process the new information, run updated contingency analyses etc.
8
(PCG) which has developed a set of proposals for electricity trading arrangements9 with European scope under the auspices of the Florence Forum. The Project Coordination Group (PCG) has also drafted a proposal for cross-border intraday trading arrangements which is summarised in an appendix to this report (see Appendix 1; chapter 7) and discussed in chapter 4.
2.9 CURRENT DEVELOPMENTS WITH RESPECT TO INTRA-DAY MARKETS
In order to complete the discussion of the background for this study, it should be noted that the regulators of the Central West Europe (CWE) region - the Central West region being one of seven regional initiatives for furthering electricity market integration in the European Union - have included the intraday market in their plans for regional market development early on. The Central-West Europe (CWE) Electricity Regional Initiative (ERI) consists of Belgium, France, Germany, Luxembourg, and The Netherlands and the regulators from these countries have produced, as early as February 2007, an Action Plan explicitly aiming at, among other objectives, an integrated cross-border intraday market. In November 2009, an extensive consultation exercise (including a joint workshop with stakeholders; two discussion papers by the regulators; a number of written statements submitted by stakeholders in response) resulted in a position paper titled "Regional intraday trade model for the Central West Region" (see section 4.3).
In parallel with these conceptual developments, concrete pilot projects for improving integration of intraday markets are underway. Three of these should be mentioned in particular:
• The intraday markets of Belgium and the Netherlands are to be integrated based on the joint use of Nordpool's intraday ELBAS system; the initial integration effort will encompass these two countries and there are plans for further expansion. The market is due to become operational in the autumn of 2010. This initiative will, like ELBAS, be based on an allocation of cross-border capacity exclusively via the power exchanges and is characterized by continuous implicit allocation of capacity.
• On the interconnections linking France and Germany, available intraday capacity will be made available to the market via both explicit first-come-first-served (FCFS) allocation (the status quo) as well as implicitly via EPEX Spot's ComXerv system (projected start date 2010 Q4).
• In addition, a project initiated in June 2010 is linking the CWE region to Nordpool (Scandinavia) and the British (GB) market. This last initiative is currently being designed and discussed between relevant stakeholders as a practical way to integrate markets even beyond the currently stipulated regions.
9
2.10 OUTLOOK ON THE REPORT
The above brief descriptions of three market integration initiatives in the area of intraday trading of electricity complete the introduction to the present report. We have described the major institutional features of electricity trading, especially cross-border trade in electricity, and sketched the process of European market integration.
3
O
BJECTIVE OF THE PRESENT REPORTThis report ultimately seeks to answer a very specific question: Should future intraday trading arrangements make cross-border capacity available for OTC transactions or not?
The time horizon being considered is the near-term future up to around 2015. There is little point in attempting to make a recommendation for the longer term ("enduring solution"). Such a recommendation should not be made now, but rather should be made in the light of future experience.
One way to read this report is with the proposed intraday trading arrangements on the Belgian-Dutch border in mind. The question is relevant beyond this particular interconnection, but it is particularly urgent with respect to the proposed integration of these two countries' intraday markets.
In the present chapter the above question guiding our research is explained in more detail. First, a few words on the specificity of the question may be in order. Next, some of the terms used are explained with a view to avoiding confusion. We then present two stylized representative trading arrangements. These are the two models that we will have in mind in chapter 5 of the report when we apply a number of criteria to evaluate the respective advantages (benefits) and disadvantages (costs).
3.1 SCOPE
The question set out above makes it clear that the aim of the present report is not to provide a comprehensive assessment of all aspects of intraday trading arrangements. Nor will we comprehensively investigate the subset of cross-border intraday trading arrangements. Instead we shall analyse these arrangements with respect to one issue: availability of cross-border capacity for OTC transactions or lack thereof.
The specificity of the question is easily understood when one recalls that the regulators in the CWE region and beyond have been working on how intraday trading arrangements or, indeed, trading arrangements in general ought to be designed, for some time now and have found much common ground (as documented, for example, in the position paper of November 2009; cf. footnote 15).
Exchange-based Over-the-counter Exchange-based Over-the-counter Longer-term Day-ahead Intraday Balancing T im e s c a le Within-area Cross-border Classification by geographical scope and type of counterparty
3.2 TERMS USED
An over-the-counter (OTC) trade involves two market parties (who might have been matched by a broker or by telephone) agreeing on the terms of a sale of power directly; i.e., the two parties are typically each other's counterparties.
"OTC" is normally contrasted with "exchange-based"; this is the term that will be used in the present report. In practical terms we will be thinking of the power exchanges existing today as the institutions providing a matching service for buyers and sellers of power. The power exchange is also the counterparty to both sides of the transaction. The key differences between OTC and exchange-based trading are summarised below.
3.3 STYLIZED TRADING ARRANGEMENTS - COMMON FEATURES
In the present section we describe the common features of the two stylized trading arrangements to be contrasted. Summarizing the common features upfront allows us to be clearer about the differences below. What is described as "common features" here is by no means confirmed - it represents our understanding of the intraday trading arrangements
Counterparty
Transparency
Competition
Trading cost
Nomination
In principle, market parties are each other's counterparties (use of PX for clearing optional) and
know each other's identity In principle, transaction data no public information; can be kept
confidential
Competition is less automatic; counterparties can be approached in a targeted
manner
Manual nomination for every interconnection that needs to
be used
Differences between transaction characteristics Exchange-based Over-the-counter
Implicit fixed cost of trading team and possibly broker
screen; implicit cost of counterparty risk (unless PX used for clearing); additional transaction fee is incurred if a
broker is used Power exchange; the actual
market parties typically remain anonymous
Anonymous summary transaction data published by
power exchange Orders compete against all other orders in the market; matching is typically automatic
Fixed cost of exchange membership; variable cost per
trade [per MWh]
being envisaged for the future as sketched out in, for example, the sources cited in chapter 4. Clearly, a number of questions remain to be resolved10.
However, it can be stated that it is planned to have a Capacity Management Module (CMM) which would administer the cross-border capacities between all the bidding areas that are part of a given trading system. It would be the TSOs' responsibility to design and operate the CMM.
This CMM is linked to a Shared Order Mechanism (SOM). This refers to an arrangement whereby bids and offers are matched (see below). The aim of a shared order mechanism is to pool liquidity from several sources, so the design of a shared order mechanism is typically in the competence of power exchanges. One design parameter is the degree of centralisation.
Note that it would be theoretically possible to allow several order books resp. SOMs to access capacity via the CMM independently of each other. As this last option, which would clearly disperse liquidity, does not seem to be pursued for the time being, it will not be discussed further in the following.
3.3.1 CAPACITY MANAGEMENT IS A NATURAL MONOPOLY
Transmission system operators (TSOs) make capacity left over at the day-ahead stage (and any increases or decreases in available transmission capacity between day-ahead gate closure and the last gate closure before real-time) available to the market via a capacity management module (CMM).
The capacity management module administers the capacity between all the participating bidding areas. All capacity (including that made available D-1 via the implicit auction) would be managed via this single system; this is the logical arrangement ensuring integration with TSOs' grid management and the tools required to accomplish the latter task (continuously updated load flow and contingency analysis etc.).
TSOs would have to be able to adjust the capacity that the CMM posts as "available" in real-time in order to be able to take into account operational requirements (outages; wind forecast errors etc.).
3.3.2 DESIGN OPTIONS REGARDING SHARED ORDER MECHANISM
Two different models are worth mentioning here. The first is known as "One-to-One" and involves a setup in which there is only a single order book (but this is accessible from all the different participating bidding areas). The ELBAS system is an example. The second is known as "One-to-Many" and refers to several distinct order books which, however, are linked via an IT solution that allows all orders to be pooled.
10
It seems clear that the intraday market is to operate as a continuous market; that use of capacity, once allocated, is mandatory and that capacity would preferably be allocated implicitly. The information about available capacity would have to be instantly available to the shared order mechanism (SOM) which pools and matches buy and sell orders from participating bidding areas subject to availability of capacity. Matching could be automatic or manual.
It seems at any rate agreed that the shared order mechanism should be able to accomodate "sophisticated products" (see below); however, definitive specifications with respect to what constitutes "sophisticated products" are not available yet. We give an illustration of the type of technical constraint that might require "sophisticated products" in chapter 5.
If a match via the shared order mechanism is successful, the back-office systems need to automatically take care of settlement across all participating bidding areas (including for the purpose of later imbalance settlement with the respective TSO or clearing institution).
If there are capacity restrictions with respect to, for example, export from area A to area B, the shared order mechanism needs to know that an offer to supply power in bidding area A is not or not fully available to a counterparty in bidding area B and show only that part (if any) that is available. We would expect market participants to access this information via a screen-based trading system similar to the (intraday) trading systems operated by the power exchanges today.
3.3.3 OPERATION OF WITHIN-BIDDING-AREA INTRADAY OTC MARKETS UNCHANGED
In our understanding, the local (within-bidding area) intraday OTC markets would continue to exist in parallel (and would be accessible via broker screen or telephone as today), but these markets would not be linked with other markets outside the bidding area11. Thus the shared order mechanism (which would also include within-bidding area exchange-based bids) would be competing for trades with the within-bidding area intraday OTC markets. Liquidity in the two markets12 would not be pooled and market participants would need to manage their positions in both markets separately.
3.4 STYLIZED TRADING ARRANGEMENTS - WITH OTC TRADE
In what follows we attempt to provide a more detailed description of what a regime where OTC cross-border trades are possible would look like. The description below strikes us as broadly consistent with the arrangements being envisaged for the French-German interconnection.
11
"not linked" in the sense that direct OTC trades would be possible only with a counterparty physically located in the same bidding area
12
The first important observation is that if OTC trades were possible, this would be in addition to some version of the shared order mechanism described above. The exact way in which capacity is made available for OTC trades would need to be determined in detail. To fix ideas we are thinking of a web-based capacity platform similar to today's. On this platform market parties can reserve explicit capacity, the use of which is mandatory. Note, however, that asking for capacity in the opposite direction leads to the effective cancellation of the first capacity reservation (via netting). This capacity platform would constitute the front end used by market parties.
As for the back end (i.e., the actual IT implementation), given that under the "One-to-Many" model there could be several trading systems (power exchanges) all having access to the capacity management module (CMM) via the shared order mechanism (SOM), a capacity channel for OTC trades via an explicit capacity platform could be thought of as merely one additional system accessing the capacity management module via the shared order mechanism as illustrated below.
Figure 1: Possible back-end solution for link to CMM
A process for allocating capacity and nominating the corresponding schedules, for taking the allocation into account in TSOs' contingency analyses etc. would need to be established resp. maintained, but this may not need to be very different from similar processes to be set up for any other party accessing the capacity management module (via the shared order mechanism). A mechanism would need to be devised that ensures that capacity made available via this channel is immediately reflected in a reduced availability of capacity (and conversely). However, these requirements are the same for all trading systems linking to the CMM via the shared order mechanism and should not be difficult to meet.
An alternative setup for the back-end would be for the capacity platform to access the capacity management module directly; i.e., in parallel to the capacity management module. This is illustrated below:
A BCDCA EF C
CF FC C A D
FC EE FC EE A FC
Figure 2: Alternative back-end solution for link to CMM
We assume that the explicit capacity allocation is also a continuous process; i.e., capacity would be made available on a first-come-first-served (FCFS) basis (in other words, in a situation of congestion that is suddenly relieved because of a recalculation of available capacity, the timestamp of the request would decide on the allocation). It should be noted that allowing for the allocation of capacity to OTC transactions nevertheless entails a mandatory use of the capacity. The "obligation" character of the capacity would most likely be enforced not by direct checks but by giving market parties a financial incentive to use allocated capacity (see 5.2.4.2). A scenario in which capacity is allocated explicitly as a true option seems out of the question; the risk of hoarding under such an arrangement is obvious. Note also that while the implicit assumption in the above is that intraday capacity will continue to be given out for free, we think that it would be possible to devise a pricing mechanism for explicitly allocated capacity. The social opportunity cost of giving the capacity to a market party for use in an OTC transaction is the surplus lost because the capacity is not available to the exchange-based market. In principle it would be possible to determine this value based on the order book(s) linked to the shared order mechanism and impose this cost on the user of the explicitly allocated capacity.
In summary, under the system sketched out here, a market party could either submit an energy bid (for sale / purchase) in the exchange-based trading mechanism or it could try to obtain capacity and separately find a counterparty for an OTC transaction.
Note that because of the way we have framed the issue in the present section (i.e., equating OTC trade with access to explicit capacity via a capacity platform) the question that this report asks is largely equivalent to the question of whether capacity should be allocated exclusively via the exchange-based trading system or not.
3.5 STYLIZED TRADING ARRANGEMENTS -EXCLUDING OTC TRADE
Even under the alternative assumption that OTC cross-border trades are in principle not possible, note that the shared order mechanism may well be able to accomodate these in practice. Before we discuss this in more detail, we would like to note that exchanges are working on developing "sophisticated products" which we will discuss in more detail below. Therefore it should be noted that even if OTC trade were not possible, these "sophisticated products" are expected to be available.
A BCDCA EF C
CF FC C A D
3.6 DIGRESSION: IMPLEMENTING OTC TRADES VIA AN EXCHANGE-BASED TRADING SYSTEM SUCH AS ELBAS
In the course of the research for this project we were alerted to the possibility of putting bilaterally agreed bids into a shared order mechanism such as ELBAS. If the nature of these bids is such that these would be matched with each other (assuming, of course, that the capacity is available) then even under a system of automatic matching (with no option of manually choosing a bid from the order book posted on the exchange's website) this may be an effective workaround.
To illustrate, assume that a market party immediately needs 500 MW for several hours because a generator has tripped. If this is more than is available in the exchange's order book, the market party could via telephone arrange with a generator in another bidding area to start up a plant and deliver the energy. Both buyer and seller would put a corresponding block bid into the system - since there are no other matching offers, the trade would go through albeit via an exchange-based trading system (always assuming that the capacity is not a constraint).
Since the possibility of organising OTC trades on ELBAS via the block trade functionality has been a topic of much discussion in the interviews and would presumably be available to market parties in Belgium and the Netherlands after the "go live" of the intraday market integration project, we describe this in more detail.
First, note that for the time being the block orders on ELBAS are not matched against hourly bids at all. The NPS programmers are currently working on implementing the cross-matching between the two order books. Even without cross-matching, the markets are linked to some extent because of the possibility of arbitrage between them.
Second, using this "trick" to schedule a prearranged trade is acceptable under NPS's market rules only if other market parties have a fair chance ("sufficient time") to also accept the bid. The minimum requirement seems to be to leave the bid in the system for five minutes. However, NPS has not implemented an automatic (e.g., five-minute) delay before an order can be accepted. One counter-measure traders use is to enter the "buy" side of the transaction, as it is less likely that a generator can respond quickly than that a consumer can respond quickly. Further, if the order were to be accepted by another market party, the price would actually be more favorable than the price that the intended counterparty would have been able to offer and so nothing would be lost.
Some interviewees criticised the use of the block-bid functionality for implementing pre-arranged trades (including trades with one-self). It was noted that this undermines anonymity and impairs the functioning of the market mechanism and thus the quality of the price as a benchmark or reference for other market parties. It was argued that any such trades should be kept outside an exchange's order book.
of dealing with the problem might be to define separate indices or reference prices for different types of transactions (just as there are different grades of crude oil).
We see the value of a solid, reliable price benchmark. However, we are not aware of empirical evidence showing that the kind of block trade being envisaged for implementation on an exchange-based intra-day market as discussed here would impair the functioning of the market.
3.7 SUMMARY
4
E
ARLIER WORK ON THE QUESTION OF INTERESTThe subject of the present study is addressed, directly and indirectly, in a number of sources. Anyone addressing the subject of intraday cross-border OTC trade needs to take this earlier work into account and so does the present study.
While the sources discussed in the present chapter vary in the extent to which they are legally binding, it seems reasonable to argue that if either (i) the question of interest is unambiguously answered in European legislation or (ii) there is a robust consensus among all relevant stakeholders on what this answer is, then additional analysis could be kept to a minimum. Despite assertions to the contrary from interviewees, neither seems to be the case.
In the following the main sources of interest will be briefly discussed and relevant passages highlighted.
4.1 REGULATION (EC)NO 714/2009
Article 16 of the Regulation13 sets "general principles for congestion management". The relevant passage is the following No. 1:
"Network congestion problems shall be addressed with non-discriminatory market-based solutions which give efficient economic signals to the market participants and transmission system operators involved. Network congestion problems shall preferentially be solved with non-transaction based methods, i.e. methods that do not involve a selection between the contracts of individual market participants."
This statement is so general as to be of little help in addressing the question of interest.
4.2 CONGESTION MANAGEMENT GUIDELINES
The so-called Congestion Management Guidelines14 are the most important piece of legislation with respect to the allocation of scarce interconnector capacity in the European Union. If the Guidelines gave an unambiguous answer to the question that this report is
13
REGULATION (EC) No 714/2009 OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL of 13 July 2009 on conditions for access to the network for cross-border exchanges in electricity and repealing Regulation (EC) No 1228/2003. Technically, this Regulation enters into force on 03 March 2011; however, since the formulation is unchanged from Article 6 No. 1 in the preceding Regulation (EC) 1228/2003 it seems reasonable to refer to the former.
14
looking into, the question would be settled. For example, it was suggested in the interviews that Article 2.7 of the Guidelines establishes more or less a right to direct access to capacity:
"Capacity allocation shall not discriminate between market participants that wish to use their rights to make use of bilateral supply contracts or to bid into power exchanges. The highest value bids, whether implicit or explicit in a given timeframe, shall be successful."
With respect to intraday capacity allocation we do not share the view that this constitutes an obligation for TSOs to make both routes - OTC and some kind of exchange-based arrangement - possible. More likely, Article 2.7 says that if both channels are available, then market parties should compete for capacity via both channels on equal terms.
Conversely, it was claimed that first-come-first-served allocation of explicit capacity is not consistent with the Guidelines because it is not market-based. Actually, Article 2.1 reads as follows
"Congestion-management methods shall be market-based in order to facilitate efficient cross-border trade. For that purpose, capacity shall be allocated only by means of explicit (capacity) or implicit (capacity and energy) auctions. Both methods may coexist on the same interconnection. For intra-day trade continuous trading may be used."
It is not clear that first-come-first-served allocation is inconsistent with this under all circumstances. For example, when there is no congestion, the (market) price of capacity ought to be zero and giving capacity to the first market party who asks for it may be as good as any other allocation.
Our point here is simply to illustrate that, in our view, the Guidelines do not unambiguously answer the question of interest.
4.3 CWEREGULATORS'POSITION
The CWE Regulators' Position15 of November 2009 suggests that allowing explicit allocation of capacity to OTC trades was clearly considered an option at least at the time. Section 4 on "Type of intraday products offered and OTC-access" reads as follows:
"Regulators consider that all cross-border intraday trades should be channelled through the intraday mechanism. Regulators regard explicit allocation of capacity to OTC-trades as a pragmatic first step towards a quick implementation of a regional solution. OTC trades should be made transparent regarding prices and volumes. Regulators will monitor the developments in OTC-trade closely. When assessing the implementation of SOM, TSOs are asked to study the introduction of sophisticated products enabling to start new units (fix costs, minimal duration etc.) in order to
15
replace explicit allocation to OTC-trades while providing for the special needs of market participants."
4.4 PROJECT COORDINATION GROUP (PCG)PROPOSAL ON TARGET MODEL
The Project Coordination Group (composed of representatives of all relevant stakeholders) developed, among other proposals, a target model (and roadmap for implementation) for the Intraday Market. This was presented to the December 2009 Florence Forum16. As this was frequently referred to in the interviews with stakeholders that were conducted for the present study, we provide a summary description of the target model in Chapter 7 (Appendix 1). The PCG target model actually is, in our view, clear in its rejection of cross-border OTC trades as they are understood in the context of the present study. However, note that a target model is by definition about the longer-term and not the near-term future17 and that the work of the PCG is non-binding and participants would, in principle, be free to change their mind (presumably based on new evidence18).
4.5 DRAFT FRAMEWORK GUIDELINE ON CAPACITY ALLOCATION AND CONGESTION MANAGEMENT
A revised draft of the Framework Guidelines on Capacity Allocation and Congestion Management was published by ERGEG in early September 2010; this also contains a section titled "Objective #4: To Design Efficient Intraday Market Capacity Allocation". This is still somewhat preliminary and, at the time of writing, ERGEG was accepting comments on the document. The draft version as of 10 September 2010 would not, in our reading, explicitly rule out continued cross-border OTC trade for the near-term future. However, it shows a clear preference for an implicit allocation of available intraday capacity.
4.6 AD HOC ADVISORY GROUP (AHAG) OF STAKEHOLDERS:INTRADAY PROJECT
While the PCG has formally ended its work, in practical terms the cooperation of the European Regulators’ Group for Electricity and Gas (ERGEG) with an ad hoc advisory group of stakeholders (AHAG) on the drafting of a Framework Guideline on Capacity Allocation and Congestion Management serves a similar function. Work on the intraday market continues as
16
PCG Proposal for Target Model and Roadmap for Intraday Market; A PCG Report to the XVIIth Florence Forum 10&11 December 2009, Rome; PCG Proposal for Target Model and Roadmap for Capacity Allocation and Congestion Management; 2nd of December 2009 (presentation; slides 19 - 23)
17
The PCG's own roadmap envisages implementation by 2015. 18
one of three implementation projects under the AHAG umbrella19. The Intraday Group's status report to the June 2010 Florence Forum flags a number of still-open issues yet does not rule out continued cross-border OTC trade for the near-term future.
4.7 SUMMARY
In addition to the sources cited above, we were alerted to additional minutes of meetings also addressing the issue of cross-border OTC trade. Also, some of the interviewees had first-hand knowledge of the discussions in the various groups. However, in summary this material is not clear enough to come to a definitive conclusion on cross-border intraday OTC trade. Neither do we see a clear prohibition against it in European legislation; nor do we see a consensus for phasing out OTC trade in the near term.
For this reason, the next section will evaluate the option of trading OTC against a number of criteria.
19
5
C
RITERIA FOR EVALUATION5.1 APPROACH
The question of what the appropriate view on OTC transactions is does not, in our opinion, have a straightforward answer. In practical terms, the approach chosen to tackle this question was the following: possible evaluation criteria developed as part of a desk study were collected in a first version of the present report. This was then distributed to a number of stakeholders suggested by NMa. The stakeholders are from the NWE region and jointly cover all market roles (market facilitators / exchanges; market parties narrowly defined / trading; regulators; transmission system operators). The individuals approached are known to NMa experts and have shown themselves to be knowledgeable about the issues of interest in the past. All-in, about twenty organisations contributed and many individuals gave very generously of their time to discuss the preliminary version of the report by telephone interview, sometimes also making notes and comments available to us in writing. The telephone interviews typically lasted at least one hour and contributed greatly to the revision of the report including the evaluation criteria described in the present chapter. All participants were promised anonymity.
Having described the two types of cross-border intraday trading arrangements - one also allocating cross-border capacity to OTC trades, but not the other - in the present section we provide an overview of the criteria that will be applied in order to evaluate the alternatives. In the following sections the criteria are discussed one-by-one.
In developing these criteria, one could draw on relevant statements in the Congestion Management Guidelines20 or in the draft Framework Guideline on Capital Allocation and Congestion Management21. We were also given many suggestions as to what the "correct" criteria are by the interviewees. In the list below we organise the discussion around the set of "main criteria for target model choice" as defined by the Project Coordination Group (see footnote 18; slide 2). Our emphasis is on the application of the criteria being concrete and operational. We have added some criteria of our own and we trust that the list covers the relevant issues; some criteria that have been deliberately given less attention are stated below.
20
e.g., "methods adopted for congestion management shall give efficient economic signals to market participants and TSOs, promote competition and be suitable for regional and Community-wide application"; Article 1.5
21
5.2 MAXIMISATION OF SOCIAL WELFARE
We do not define an explicit social welfare function. Instead the criterion is broken down into six sub-criteria in order to make it operational. We first list these criteria and explain them in detail in the remainder of the section.
• Efficient utilisation of cross border capacity
• Market based allocation of additional intraday capacity • Implicit maximisation of social welfare by facilitating trades • Effect on intensity of competition in generation
• Effect on intensity of competition between trading venues • Cost of implementation and operation
5.2.1 EFFICIENT UTILISATION OF CROSS BORDER CAPACITY
Subject to the assumption that explicit allocation for OTC trades and implicit allocation for exchange-based trades both occur on a first-come-first-served basis22 and subject to the further assumption that capacity continues to be allocated for free, is the capacity that is allocated at any moment in time put to the socially most valuable use; i.e., that transaction which generates the most social surplus? The latter value we propose to measure by the difference between buyers' and sellers' respective reservation prices23. In the case of (limit) orders submitted to the exchange-based trading system, we assume that market participants accurately reveal their valuations via the prices at which they are willing to buy or sell. In the case of requests for explicit capacity, we assume that the requesting market party's willingness to pay for capacity is an acceptable proxy for the social surplus created by the corresponding energy transaction and that they, too, correctly state their valuation.
To answer the question, we consider three scenarios24:
a) There is no congestion (all feasible trades have been executed resp. all requests for explicit capacity have been met and additional capacity is still available)
b) There is congestion (there are feasible trades that have not been executed resp. there is unmet demand for explicit capacity, but no additional capacity is available)
22
Under implicit allocation, priority is determined by the surplus that a trade could generate and not by the timestamp. However, the fact remains that first-come-first-served is used: If capacity has been assigned to a transaction, this is not revoked if a little later another bid is put into the order book, acceptance of which would create greater surplus. Market parties would, of course, be free to engage in arbitrage.
23
The assumption in this is that these accurately reflect the social gain and cost, respectively, of energy consumed resp. produced.
24
c) Starting from a situation of congestion, additional capacity becomes available In situation a) additional demand for capacity could arise both because additional bids in the exchange-based trading system mean that additional trades become feasible and / or because there is additional demand for explicit capacity. Note that in the present section we assume that all demand for explicit capacity is genuine; the potential problem of hoarding is considered in sub-section 5.2.4.
The efficiency criterion can be translated into the question "If and when capacity is allocated to a transaction - be it implicitly or explicitly - is this transaction the one the generates the highest possible surplus at the moment at which the capacity is allocated?" In other words, given that there seems to be a consensus in favor of continuous, first-come-first-served trading it is of no importance whether three hours or three minutes or even three seconds later the capacity could have been allocated to a different transaction that would have created more surplus. All that matters is the situation at the time at which the capacity is given out. Given this definition of the efficiency criterion and assuming that some kind of tie-breaking rule is in use that ensures that additional requests for capacity always arrive sequentially, it is clear that a setup with OTC trade is almost equivalent to a set up without OTC trade. Starting from a situation of "no congestion" explicit capacity might be allocated to a trade that generates less surplus than another trade that would have been possible a little later. In the same way, implicit capacity might be allocated to a trade that generates less surplus than another trade that would have been possible a little later. The equivalence is not complete because we conjecture that a trade that turns out to be inefficient ex post would be more likely to be reversed via arbitrage if all trades were executed on the exchange-based trading system (e.g., because a market party also trading over-the-counter might not follow the exchange-based trading system as carefully). Therefore, assuming that this conjecture is well-founded and that we are starting from a situation of "no congestion", the exchange-based trading system might have a slight advantage.
A very similar reasoning applies to situation b). Because there is congestion by assumption, no capacity is allocated and no trades take place. However, consider those trades that were concluded for a certain hour. Assume that a market party has obtained power at a price of 35 which it values at 40. Assume further that this market party observes that there are additional offers to buy at 50. Clearly it would be beneficial to sell on the power. Again, we would expect this type of arbitrage to be somewhat more likely to occur if all trades were handled by an exchange-based trading system. If this expectation is justified, the exchange-based trading system would once again have a slight advantage.
Situation c) is more complicated because there are more cases to distinguish. By assumption we start from a situation of congestion so there are feasible trades in the exchange-based trading system that are not executed resp. requests for explicit capacity that are not met because of a lack of capacity. As the additional capacity becomes available, the question is whether it is given to those trades generating the most surplus.
that at the moment at which the capacity is allocated, it is allocated to those transactions generating the highest surplus.
If there is only demand for implicit capacity (i.e., for transactions on the exchange-based trading system), our understanding is that automatic matching of limit orders submitted to the order book occurs. Those buyers offering the highest price are automatically matched with those sellers asking for the lowest price (subject to overall surplus being maximised; this means that a slightly higher sale price or a slightly lower purchase price might be accepted if the additional volume is such that this maximises surplus25). Even though the price of capacity is zero, this is clearly an efficient arrangement (regardless of the price at which the transaction is settled which could be any price between the bid and the ask).
Note that in principle it is also possible to directly accept ("click on") a limit order that has been submitted to the order book and is displayed on the exchange-based trading system. In the ELBAS system this would only be possible for transactions for which the required capacity is available; other orders are not displayed. However, if the required capacity is available then by definition there is no congestion. From an efficiency point of view, a potential problem with direct acceptance of limit orders is that the valuation of the market party accepting the transaction is not known (in practice the trading system seems to generate a counter-bid at a price exactly equal to that specified in the limit order that is being accepted). It is therefore not easily possible to rank resp. compare this transaction against other transactions. However, since it ought to be possible to directly accept a limit order only when there is no congestion, this is not a problem in practical terms.
If there is both demand for explicit capacity as well as demand for implicit capacity, it is not clear how additional capacity becoming available would be divided up between the two. Since by assumption the price of capacity is zero and since it is not clear how much surplus would be generated by those transactions for which explicit capacity is requested, a plausible allocation rule would be to give priority to the orders submitted to the exchange-based trading system. If and when all of these feasible trades have been accepted and there is still capacity left we are back in situation a). That exchange-based transactions should have priority is only a suggestion; note that implementing this suggestion would not ensure an efficient allocation of capacity because it is not known how much surplus the competing OTC trades would generate.
Situation c) is the situation that arises each time the intraday market opens for trading.
25
Summary - Efficient utilisation of cross border capacity (5.2.1): Continued availability of OTC trading leads to a less efficient utilisation of cross-border capacity for two reasons. First, under a purely exchange-based trading system we would expect it to be more likely that intertemporal inefficiencies are reduced via arbitrage. In other words, if capacity allocated to one transaction later turns out to have higher value if allocated to another transaction, we believe that this correction would be more likely to occur if all trade were concentrated on an exchange-based trading system. However, this reasoning is based on a conjecture and subject to confirmation by empirical evidence. The second source of inefficiency arises in that no signals as to the social value of allocating capacity to any given OTC trade are available (assuming that capacity continues to be priced at zero). Therefore it is not possible to rank different requests for explicit capacity against each other; nor is it possible to rank them against competing requests for implicit capacity. This problem does not arise with limit orders submitted to the exchange-based trading system. However, to the extent that efficient pricing of intraday capacity is implemented in the future, the second source of inefficiency is reduced or even eliminated.
5.2.2 MARKET BASED ALLOCATION OF ADDITIONAL INTRADAY CAPACITY
This criterion considers the case where capacity is priced; i.e., the assumption of free capacity is relaxed. Is it easier to implement market-based pricing of capacity when all trading takes place on an exchange-based trading system? By market-based pricing we understand a system of pricing capacity that allocates capacity to those transactions that create the most social surplus subject to the maintained assumption that trading is continuous.
In order to answer the question, we need to reconsider scenario c) above (cf. 5.2.1) allowing, however, for the pricing of capacity.
When there is only exchange-based trading, the pricing of capacity is straightforward. Matching the lowest-price offers to sell with the highest-price bids to buy until the available capacity is exhausted creates maximum social surplus. Any capacity pricing rule that leads to these trades being realised is efficient. For example, a uniform price of zero would be a trivial pricing rule but it would certainly maximise surplus. One could also imagine setting a uniform price of capacity such that the available capacity is exactly used up. As another example, a perfectly price-discriminating monopoly would work just as well if we continue to assume that market parties accurately reveal their valuations (reservation prices) in their bids. Thus an exchange-based trading system makes market-based pricing of capacity easy.
Note, however, that in the above we have made the rather strong assumption that market parties reveal their true reservation prices even if they are pushed down to this reservation price by a perfectly price-discriminating monopoly. More generally, if a market party's (expected) profit depends on its stated reservation price, it is doubtful whether it would still have an incentive to reveal the true valuation.
hand, if the seller is matched (i.e., she does obtain capacity) she will make a strictly positive profit. The overall effect of the strategic misrepresentation is to increase the expected profit. We therefore need to ask whether the result - namely, that market-based pricing of capacity is straightforward under an exchange-based trading system - is critically dependent upon the assumption of truth-telling on the part of market parties. It is beyond the scope of this report to provide a rigorous, game-theoretical analysis of the problem. However, the following observations suggest that the result is robust: First, limit orders submitted to an exchange-based trading system do provide valuations of some sort even if the reservation prices are determined strategically. If market parties behave strategically in similar ways, it seems plausible that the pricing rule would still lead to the efficient transactions being scheduled. Second, note that market parties compete against each other. Intuitively, the more intense the competition the less scope there is for strategic behavior.
In the case where the only demand is for explicit capacity, allocating additional capacity in a market-based way requires that some kind of valuation be expressed. If market parties requesting capacity submit their willingness to pay along with their request, it would be easy to allocate explicit capacity in a market-based way (subject to the necessary IT preconditions being met).
The above caveat with respect to the strategic misrepresentation of valuations would apply with respect to explicit capacity, too, but in principle market-based pricing would not be a problem.
Note, however, that we have thus far maintained the assumption that the requesting market party's willingness to pay for capacity is an acceptable proxy for the social surplus created by the corresponding energy transaction. The discussion in section 2.5 suggests that this is not the case for the day-ahead timeframe. We would expect explicit intraday capacity allocation to be somewhat inefficient, too. The problem is likely less severe because trading occurs closer to real-time (which ought to lead to more accurate assessments) and unlike in the day-ahead time frame where there is only a single auction, there is scope for reversing inefficient transactions. All in all, we would expect explicit capacity allocation to be somewhat less efficient than implicit allocation even in the intraday timeframe (although we are not aware of empirical evidence to support this conjecture).
