Michael Giberson
Most of the integrated power markets in the United States release market participant bid and offer data after a six month delay, with identities of the participants masked. Recently ISO-New England proposed reducing the amount of time the data release is delayed. A recent overview of data release policies, conducted by William Dunn for the American Public Power Association, discusses the “most discussed issue” when information release policy is addressed — collusion — and noted something interesting about comments filed with FERC concerning the ISO-NE proposal:
The most frequently discussed concern associated with data release in electricity markets is whether the rapid release of data is more likely to facilitate competition or collusion. The most interesting aspect of this discussion is that market participants seem to be taking positions that are counterintuitive. Those who support continued confidentiality or delayed and masked release of data seem primarily to be the generation resource owners, who could be expected to benefit from data release if it would truly facilitate collusion. Conversely, it is the LSEs who generally advocate faster and resource-specific release of data, and they are the market participants who would be harmed by any collusion such data release facilitated.
In his report, Dunn recommends:
- Unmasked electricity market offer and bid data should be released on the day after the operating day;
- The unmasked physical operating characteristics of generation resources (i.e., minimum run/down times, ramp rates, etc.) should be publicly available;
- Transmission system conditions should be available to those with secure access to the market systems; and
- All market operators should post this information in a consistent and easily accessible manner.
As Dunn’s overview points out, the standard six-month delay is not the only approach possible. Australia’s National Electricity Market releases unmasked offer data on the day after the operating date. Perhaps there are market characteristics that would favor a long delay in some cases and little or no delay in other cases. In the Australian case, Dunn reports that market participants favored rapid disclosure so they could monitor NEM for favoritism shown to government-owned generation at the expense of private operators.
FERC has approved ISO-NE’s request to move to a three-month delay beginning on March 1, 2008, but I suspect that this is not the end of discussions on the topic in the United States.
Until 2004 PJM included a two-digit alphanumeric code for each bidder in the published bid data. Looking at the results, it was pretty easy to figure out which company mapped to each code, and thus to fuigure out who was bidding how. In 2004, they stopped using the ID code, and left the “bidder” field blank in the published data. I’ve always assumed this was because of complaints from bidding firms.
I would hazard a wild guess that power is priced competitively at about 95% of the nodes in PJM-NY-NE, and monopolistically at the other 5%. My guess is that a quicker data release would cause better competition at the competitive nodes and more market power at the constrained nodes, but that would take actual analysis to validate.
In the report Dunn says essentially the same thing, that just about anyone who cares can figure out who is who in the masked data, though Dunn thinks that small generators may not have the necessary resources to figure things out. (But being small, they would tend to maximize their profits by bidding their marginal costs in the markets anyway.)
In the report Dunn says essentially the same thing, that just about anyone who cares can figure out who is who in the masked data, though Dunn thinks that small generators may not have the necessary resources to figure things out. (But being small, they would tend to maximize their profits by bidding their marginal costs in the markets anyway.)
Yes the major New England LSEs did support reducing the time lag for bid data release from 6 months to 3 months, while only generation owners opposed. However, the same LSE’s voted down proposals to shorten the bid release further or to completely eliminate the time lag (as Dunn now proposes). One concern among the LSEs was that eliminating the lag completely would create short-term market power opportunities.
Yes the major New England LSEs did support reducing the time lag for bid data release from 6 months to 3 months, while only generation owners opposed. However, the same LSE’s voted down proposals to shorten the bid release further or to completely eliminate the time lag (as Dunn now proposes). One concern among the LSEs was that eliminating the lag completely would create short-term market power opportunities.
Yes the major New England LSEs did support reducing the time lag for bid data release from 6 months to 3 months, while only generation owners opposed. However, the same LSE’s voted down proposals to shorten the bid release further or to completely eliminate the time lag (as Dunn now proposes). One concern among the LSEs was that eliminating the lag completely would create short-term market power opportunities.