Great article! As an energy economist, do you have a preference between wholesale market centric designs like ERCOT and Australia, or wholesale market + capacity payments like PJM?
You are a better economist than I am but, I would argue that renewables bidding into the market and receiving the last generators bid is not a marginal price. The bid price doesn't cover their significantly higher CAPEX and they rely on the higher marginal fuel cost of the gas generators. When that gas usage decreases the market will quickly get out of equilibrium. Maybe I am missing something?
Hi Kay! Thanks for your comment. Without resorting to the calculus proof, here's the logic of the general theoretical result, which we also generally see borne out empirically in a range of applications.
Consider a setting in which there are N sellers, and each producer incurs a fixed cost F and a marginal cost c of producing one unit of the good (i.e., seller i's cost function is C_i = F_i + c_i*q_i). Note that that subscript means that each seller has a different fixed cost and different marginal cost of producing the same good.
Now suppose that the institutional framework is a uniform price auction as I described. Analytically, what we're looking for is whether the auction institution (1) induces the "right" sellers to produce, meaning the lower cost sellers produce before the higher cost sellers, and (2) enables the participants to create the most possible surplus, the gains from trade. What's going to ensure that those two things happen? An institution that gives buyers incentives to bid based on their true values and not misrepresent their preferences, and that gives sellers incentives to offer based on their true marginal cost.
Why does a uniform price auction do that? The logic here is essentially proof by contradiction. Suppose a seller submits an offer greater than its marginal cost, trying to increase their revenue. There are two reasons that may not actually be profit maximizing for them: (1) they risk not selling, because if they offer at a false c that ends up being higher than the market clearing price, they don't sell, earn 0 revenue, and still incur their fixed cost; (2) strategically they are competing with other sellers, so they can think the other sellers also have an incentive to overstate their c ... but each of them then also has an incentive to undercut the competition, to make sure that they get to produce. Now suppose instead that the seller submits an offer less than its marginal cost. If the market clearing price is greater than their offer but lower than their true c, then they are not even covering their variable costs, let alone any of their fixed costs. Thus from both directions, each seller has an incentive to offer at their true c; this is what it means when you see the statement that offering your true marginal cost is a Nash equilibrium -- no seller has any incentive to deviate from offering their true cost.
The same logic also applies to buyers and bidding their true marginal value in a double auction.
Because the uniform price auction induces truthful bids and offers, it means that the resulting market supply curve and demand curve reflect accurate marginal benefits and marginal costs, which means further in an electricity setting that it induces efficient dispatch, dispatching the lowest c units first etc. etc. That means that the last unit dispatched, the one that *in conjunction with the last unit's buyer's preferences* determines the market clearing price in that period, is the one with the highest marginal cost that is still compatible with buyer preferences. All of the other units with lower c are what's called inframarginal units, with lower c.
A uniform price auction recognizes the essential economic insight that (1) those inframarginal units earning producer surplus enables those units to earn revenue to help cover fixed costs, and (2) dynamically, that producer surplus communicates into the market incentives to invest in lower-cost production methods that enable that producer surplus. Notice that this is how most exchange works, including most consumer products that we purchase through a posted-price take it or leave it institution (e.g. the supermarket). This is how market processes induce producers to wring costs out of their production processes.
If a seller's total costs are higher than the revenue that they can earn, that fact communicates important information to them, that they should exit the industry and do something more valuable with their resources. THIS is where subsidies like the PTC/ITC create distortions because they interfere with the clear communication of that value information ... but they also are meant to inject (in a coarse and clunky way) information about the environmental value of those resources that is missing in current market institutions.
Sorry that was long, but this is the essential logic of all markets and of the uniform price auction in particular. I'm not sure I understand what you mean by "marginal price"; when determined by the interaction of the marginal value of the last unit and marginal cost of the last unit, by definition the market clearing price is a marginal concept. The inframarginal producers receiving that price doesn't change the fact that it's marginal.
One more thought on your specific point: I think with more low/zero marginal cost resources in the mix, in more periods the market clearing price will be set by demand's marginal value. If that's occurring at a quantity at which the sellers can't make money they'll exit, prices will go up, buyers will change their behavior in ways that reduce demand in an effort to reduce their expenses. That's the negative feedback effect that makes markets self-correcting and therefore more flexible and adaptable to changing exogenous factors like weather.
I can’t wait for the next installment!
Great article! As an energy economist, do you have a preference between wholesale market centric designs like ERCOT and Australia, or wholesale market + capacity payments like PJM?
Thanks! I have a long record of being very critical of capacity mechanisms. More on that later this week ...
You are a better economist than I am but, I would argue that renewables bidding into the market and receiving the last generators bid is not a marginal price. The bid price doesn't cover their significantly higher CAPEX and they rely on the higher marginal fuel cost of the gas generators. When that gas usage decreases the market will quickly get out of equilibrium. Maybe I am missing something?
Hi Kay! Thanks for your comment. Without resorting to the calculus proof, here's the logic of the general theoretical result, which we also generally see borne out empirically in a range of applications.
Consider a setting in which there are N sellers, and each producer incurs a fixed cost F and a marginal cost c of producing one unit of the good (i.e., seller i's cost function is C_i = F_i + c_i*q_i). Note that that subscript means that each seller has a different fixed cost and different marginal cost of producing the same good.
Now suppose that the institutional framework is a uniform price auction as I described. Analytically, what we're looking for is whether the auction institution (1) induces the "right" sellers to produce, meaning the lower cost sellers produce before the higher cost sellers, and (2) enables the participants to create the most possible surplus, the gains from trade. What's going to ensure that those two things happen? An institution that gives buyers incentives to bid based on their true values and not misrepresent their preferences, and that gives sellers incentives to offer based on their true marginal cost.
Why does a uniform price auction do that? The logic here is essentially proof by contradiction. Suppose a seller submits an offer greater than its marginal cost, trying to increase their revenue. There are two reasons that may not actually be profit maximizing for them: (1) they risk not selling, because if they offer at a false c that ends up being higher than the market clearing price, they don't sell, earn 0 revenue, and still incur their fixed cost; (2) strategically they are competing with other sellers, so they can think the other sellers also have an incentive to overstate their c ... but each of them then also has an incentive to undercut the competition, to make sure that they get to produce. Now suppose instead that the seller submits an offer less than its marginal cost. If the market clearing price is greater than their offer but lower than their true c, then they are not even covering their variable costs, let alone any of their fixed costs. Thus from both directions, each seller has an incentive to offer at their true c; this is what it means when you see the statement that offering your true marginal cost is a Nash equilibrium -- no seller has any incentive to deviate from offering their true cost.
The same logic also applies to buyers and bidding their true marginal value in a double auction.
Because the uniform price auction induces truthful bids and offers, it means that the resulting market supply curve and demand curve reflect accurate marginal benefits and marginal costs, which means further in an electricity setting that it induces efficient dispatch, dispatching the lowest c units first etc. etc. That means that the last unit dispatched, the one that *in conjunction with the last unit's buyer's preferences* determines the market clearing price in that period, is the one with the highest marginal cost that is still compatible with buyer preferences. All of the other units with lower c are what's called inframarginal units, with lower c.
A uniform price auction recognizes the essential economic insight that (1) those inframarginal units earning producer surplus enables those units to earn revenue to help cover fixed costs, and (2) dynamically, that producer surplus communicates into the market incentives to invest in lower-cost production methods that enable that producer surplus. Notice that this is how most exchange works, including most consumer products that we purchase through a posted-price take it or leave it institution (e.g. the supermarket). This is how market processes induce producers to wring costs out of their production processes.
If a seller's total costs are higher than the revenue that they can earn, that fact communicates important information to them, that they should exit the industry and do something more valuable with their resources. THIS is where subsidies like the PTC/ITC create distortions because they interfere with the clear communication of that value information ... but they also are meant to inject (in a coarse and clunky way) information about the environmental value of those resources that is missing in current market institutions.
Sorry that was long, but this is the essential logic of all markets and of the uniform price auction in particular. I'm not sure I understand what you mean by "marginal price"; when determined by the interaction of the marginal value of the last unit and marginal cost of the last unit, by definition the market clearing price is a marginal concept. The inframarginal producers receiving that price doesn't change the fact that it's marginal.
Does that help?
One more thought on your specific point: I think with more low/zero marginal cost resources in the mix, in more periods the market clearing price will be set by demand's marginal value. If that's occurring at a quantity at which the sellers can't make money they'll exit, prices will go up, buyers will change their behavior in ways that reduce demand in an effort to reduce their expenses. That's the negative feedback effect that makes markets self-correcting and therefore more flexible and adaptable to changing exogenous factors like weather.