Consultation with the Department of Treasury & Finance (DTF) and Centre for Market Design (CMD) was gratefully undertaken in research for this article, however any views, opinions and shortcomings are strictly those of the author and do not reflect either organisation or their personnel.
Part 1 highlighted a few pertinent pitfalls of the Coalition’s Emissions Reduction Fund (ERF), namely the auction design of their pay-as-you-bid sealed tender, in which firms place bids on the quantity and cost of their abatement, and the cheapest projects are selected until the ERF budget is depleted. The focus of this article will be to offer an alternative model and attempt to justify its viability. Part 3 will zoom out to look broadly at auction theory, and centre discussion on improving the outcomes of any chosen auction design.
Firstly, it is important to recognise that there is currently no closed-form solution or optimal design for a multiple-unit, multiple-bid auction, as is the case in the ERF. Many auction theorists point to Vickrey or Ausubel type auctions, however empirical evidence has not caught up with the theoretical predictions, and neither auction type has received discernible traction in their practical implementation. The unfamiliarity of these designs would most likely serve to be unpopular among policy makers, businesses and the community.
In light of this, an alternative design is being proposed here after consultation with CMD and the Market Design team at DTF, one that is suggested to better achieve the goals of the auction by maximising competition and incentivising bidders to reveal their true willingness-to-abate (the minimum cost of emissions reduction).
The proposed auction is characterised by the following properties:
- Open outcry, where bids, prices and the price ceiling are freely revealed
- Ascending clock:
- The bid price on the clock begins at a nominal low price (say $1 per tonne)
- Bids on quantity are accepted
- If the price times total quantity is less than the budget, then the clock price increases by a predetermined increment (say $0.25 per tonne)
- All previous bids are upheld and new bids on quantity are accepted
- The process continues until the predetermined budget is met or the clock price reaches the ceiling.
- Uniform clearing price, where the final price on the clock is paid to all winning bidders
- Large infrequent auction rounds
Reducing the frequency of auctions and setting a uniform price, will limit the incentives for firms to delay their bids or to bid higher than their minimum cost of abatement. The dominant strategy for all participants will be to bid at the earliest point in the auction when the price is equal to or greater than their willingness-to-abate. This will ensure that their project is favoured and selected, whilst also ensuring no surplus funding is left on the table.
If a firm’s bid is timely and the price rises, then they will receive the higher clearing price for their winning quantity, and still capture the excess available profits. On the other hand, if a firm delays its bids in order to bid-shade and let the clock price rise, the budget, which is confidential, may be depleted before the firm is able to claim its desired quantity. This is despite the fact that the clearing price may exceed their willingness-to-abate. Recall from Part 1 that ‘bid-shading’ is when a participant changes their bid in order to receive a more favourable price and increase their surplus value.
Furthermore, in order to satisfy economic efficiency, firms need to be ranked from lowest to highest in terms of cost-effectiveness, so that the most efficient firms can be selected. This can be achieved by designing an auction such that truthfully revealing the minimum cost of abatement in order to be ranked favourably is a (weakly) dominant strategy; one where no other strategy will lead to a strictly better outcome.
Cost minimisation is all about value-for-money and ensuring that the price for abatement is as low as possible. A uniform clearing price will inevitably result in some participants receiving a price that is clearly greater than their minimum cost of abatement. But will this ‘overpaying’ of low cost bidders lead to less abatement?
Well, no, and it actually may lead to more abatement. Here’s why.
In a pay-as-you-bid auction as explained in Part 1, the same low cost bidders have an incentive to bid-shade in order to capture excess surpluses. If they are all able to accurately bid-shade to the highest winning price, then neither pricing method will be more cost effective than the other. The below graphs illustrate this equivalence. Each point represents a firm and its respective minimum cost of abatement. Under uniform pricing, firms will bid truthfully, and receive the clearing price, resulting in expenditure equal to the enclosed box. Under pay-as-you-bid pricing and accurate bid-shading, low cost firms will bid up to the clearing price, indicated by the arrows, and the expenditure result will be the same.
However, due to confidential pricing and imperfect information, it is unlikely that the Government’s proposed auction will result in the lowest cost projects being selected. Uncertainty and the incentives to strategically bid-shade are likely to lead to a poor ranking of emission reduction projects and the pooling of bid prices. This may result in higher cost projects being inadvertently selected and higher prices being paid, culminating in lower levels of abatement performed by less efficient firms.
All in all, on the surface it appears that this proposed auction design would achieve economic efficiency and cost minimisation and consistently outperform the Government’s Green Paper proposed auction. Alas, unfortunately this auction may also be open to strategic gaming. Part 3 will look at how firms may attempt to maximise their gains from an ascending clock auction, how this may be countered, and general measures at improving the outcomes of both auction designs.
Given the process outlined today, how would you ‘beat’ this alternative auction design?
 For the auction design students: this is not dissimilar to the ‘Revenue Equivalence Theorem’ (RET), however, the RET has not been proven to hold under multiple-unit auctions.
Image: ‘Globe’ by Cynthia Chang, https://flic.kr/p/5XTyMK. Licence at https://creativecommons.org/licenses/by-nc/2.0/.