The real life example here is electric utilities. The way they’re regulated they charge a kWh price roughly equal to the average total cost (let’s say about 12 cents). The proper way to price would be at the marginal cost (at around 4 cents). The fact that marginal costs are below average total costs are what makes them a natural monopoly.
The somewhat obvious better solution would be to charge marginal cost for each kWh and then have some other method to collect the massive fixed costs. But for whatever historic reasons, we don’t do that and most (all?) utilities price each kWh at about the average total cost. This means that as a society our quantity demanded kWh is way below where economic theory says it should be.
However, there is probably a fairly substantial pollution/CO2 externality to producing electricity. Without some analysis it isn’t obvious whether we’re producing too much electricity or too little.
I did try once to look at estimates of the size of the externality to see if it made up for the pricing way above marginal cost issue and the preliminary results were that the externality was smaller (meaning, global warming considered, we’re still not using enough electricity). However, there were a couple of points I’d need to get into deeper.
1) The pricing above marginal cost issue is greatest for residential rates and smallest for industrial rates. I was looking at residential rates. Using the same cursory analysis on industrial rates would mean that we’re over using electricity in industrial sectors.
2) The carbon externality number I used from the EPA seemed to be derived by figuring out how high the price of electricity would need to be to get usage down to the level they wanted. Under correctly priced utility rates (i.e., priced at the marginal costs), their analysis may have had a much higher $ / kWh externality number. But at the same time, I’m a little suspect of that method of calculating the externality as it would indicate if the cost of production halved it wouldn’t be optimal for society to produce more. So I’d need to do some more research to make sure I’m using good pollution/CO2 numbers.
I haven’t seen this issue discussed by people like Mankiw when they talk about the Pigou Club and I think it probably should be. If there’s interest I could probably write this up a bit more formally and make it a post.
There exists much better work than that on power production externalities. ExternE, for starters. Which mostly prove that the amount of pollution has remarkably little to do with how much power you produce, and a heck of a lot to do with which technologies you use to produce them. Major takeaway if you do not care to read that “Coal is not a good idea, even ignoring the carbon.”.
This doesn’t apply in all countries. In UK for instance, it is common to have a standing charge (flat fee per day) as well as a usage charge (fee per kWh). Or some utilities charge a high price for the first few kWh, and then a lower price for subsequent kWh, which has a similar effect. See here for some details.
Even where there is a single price (a price per kWh) it is not true that the “correct” market price is just the marginal cost. Suppliers do need to recover their costs of capital, and fixed costs, or they will go out of business. Imagine a market with a huge numbers of suppliers, where the price drops to marginal cost. They will all be losing money, but some will go broke quicker than others. As suppliers exit the market, the remaining suppliers find they can increase their price, and equilibrium is established when some marginal supplier is just hanging on in the market (barely making enough revenue to cover total costs). If the number of suppliers drops below this equilibrium, then they all start making large profits, but this situation should attract a new entrant into the market, so restoring equilibrium. That’s how the market theory works of course: real life situations create both barriers to entry (overregulation, obstruction of access to wholesale supplies, or to the distribution network) and barriers to exit (loss-making firms are propped up for years by subsidies, bailouts etc).
This doesn’t apply in all countries. In UK for instance, it is common to have a standing charge (flat fee per day) as well as a usage charge (fee per kWh). Or some utilities charge a high price for the first few kWh, and then a lower price for subsequent kWh, which has a similar effect. See here for some details.
My perspective is US-centric, but from what I’m aware the per kWh price in most countries for most people is well above the marginal costs. Many places do have a daily or monthly charge but that tends to be $10 or less—not even close to high enough to recover all the fixed costs associated with a customer. Looking through some of the Scottish Power rates that you linked to, the daily charge doesn’t get much higher than 30p. That helps mitigate the issue a bit, but it’s still there. In that case, retail kWh prices—after the standing charge—is still over 10p / kWh. Wholesale rates look like they’re 4.5p in the UK (which should be a good proxy for short run marginal costs) so there’s still a big gap.
Even where there is a single price (a price per kWh) it is not true that the “correct” market price is just the marginal cost...
As far as I’m aware, economic theory says the “correct” price for electric utilities is lower than where the actual price is. It’s probably easier to visualize on a graph like this one. (I’m saying the difference between Pf and Pr, at least in some cases, may be higher than the externality, which is a real-life example of what the op is talking about). If that’s not standard economic theory though let me know as it’s an area of interest to me.
The market correction mechanism you described works for most industries but electric utilities are typically treated as natural monopolies, the optimal number of suppliers is one. But even if that ’s not true (i.e. it’s not actually optimal), in many places regulation only allows one supplier so the market forces described couldn’t work. The result is that the average /kWh price customers pay is higher than the average marginal costs (optimal society price) and it continues indefinitely because new firms can’t come into the market. There isn’t large profits made though because they’re pricing at the regulated price (at the average total cost) and not at the monopoly price (again, easier to visualize on the graph linked to above).
It’s probably easier to visualize on a graph like this one. (I’m saying the difference between Pf and Pr, at least in some cases, may be higher than the externality, which is a real-life example of what the op is talking about).
I think that, strictly, Stuart was arguing that the difference between Pm and Pf exceeds the externality cost, which may well be true. However, politically it is of course much easier to force a polluting monopoly to lower its price (to Pf) than to subsidise said monopoly still further. It is also economically more efficient (there are better things to do with public money).
You may also be right that the externality cost exceeds the difference between Pf and Pr—referring to the UK numbers, does the externality actually work out at less than about 5p per kWh? Even if it does, I’d argue that it is unrealistic to expect the price to drop to Pr and stay there indefinitely (while the suppliers go broke).
The “Correct” price for electricity is one price to be connected to the grid and several more relating to one’s power used, power factor, peak demand, and the like.
The average price paid is lower than the “correct” price, because charging the “correct” price adds lots of measuring and billing costs. It’s better to allow some subsidizing to be happening than to spend more just to make sure it isn’t.
An easy way to do it would be to charge the “correct” marginal cost for all kWh and have a separate fixed fee. My water bill is something like $50 fixed and then a small amount for the water I use after it; the electric bill could work the same. Ronald Coase argued that here
Commercial meters have priced kW for a long time and I think the reason residential didn’t was more along the lines of they’re more homogeneous than the meter costs. But either way, it seems everyone is getting smart meters now and you could match it up to theory exactly if it were politically feasible.
The political problem is that some people would be charged more; regression to the mean suggests that those people would be the ones who currently pay the least. The people who pay the least are the people who use the least power.
In a proper cost-sharing setup, dividing the fixed cost of maintaining each portion of the grid among the people served by that portion of the grid, pricing would be fiendishly complicated and appear unfair: Consider ten rural houses roughly in a line sharing a single branch line from the distribution station: each of these houses would be charged equally for their use of the larger distribution network, but the first house would be charged with 1⁄10 the cost of maintaining the first segment of their shared circuit, the second house would be charged that plus 1⁄9 the cost of the second segment, and so forth.
Or is it: If someone builds an 11th house at the end of that line, and the added load requires that the first segment be upgraded (to a line with higher maintenance costs) to handle the additional load, how is that cost fairly distributed? (What if the 11th house is added in the middle of the line?)
A $X+$y/kWh system makes more sense, but there is no system which is perfectly fair and appears to be fair to most people.
The political problem is that some people would be charged more; regression to the mean suggests that those people would be the ones who currently pay the least. The people who pay the least are the people who use the least power.
Everything you say is true, but your implied argument is flawed (you are implicitly making an “all A are B, all C are B, therefore all A are C” argument). If we had a fixed fee, and were discussing the possibility of eliminating it, your argument would apply just as well.
Sorry- that made much more sense as a lead-in to the self-redacted segment where I pointed out that the higher-spending (and presumed higher-income) users were subsidizing the poor, and suggested that might be a feature.
I get the difference in country perspective, and the difference between a regulated local monopoly (which I believe you have in US) vs a market structure with a regulated distribution network, but a large choice of retail suppliers.
Incidentally, there are 6 large UK suppliers, and multiple smaller ones, so this is far from perfect competition, but also far from a monopoly. According to this report retail margins are about 7 per cent, and the previous margin was only about 3 per cent. Even in a “good year”, the big six suppliers are making an average profit of around £100 per year on an average bill of around £1400 per year. So if the retail price dropped to 9 pence per kWh as opposed to 10 pence per kWh, the firms would probably be making a net retail loss.
While the current profit margin probably does encourage entry, I can’t see any way that the retail price could drop to 4.5 pence per kWh and still allow a supplier to make a profit. It seems quite possible that a new supplier could start up offering 5 pence per kWh, and would rapidly grab market share: the fact that no-one in UK is trying that suggests that it just doesn’t stack up as NPV positive.
Even where there is a single price (a price per kWh) it is not true that the “correct” market price is just the marginal cost.
In a competitive market, the correct market price is the marginal cost. If the market price is higher, the firm will profit by producing more; if it’s lower, it will profit by producing less.
In the situation you describe (price = marginal cost), it may be impossible for the firm to make a profit; all it can do is minimise its losses, so that it “only” loses its fixed costs. However, since suppliers are making losses, the number of suppliers will contract i.e. this is not an equilibrium situation.
True, there are idealized models of “perfectly competitive markets” where all fixed costs, costs of entry, costs of capital etc are negligible. In that case, price will stay at marginal cost in equilibrium. But “competitive markets” are a bit more general than “perfectly competitive markets”.
A slightly more detailed analysis looks at the “capacity” of the suppliers. Suppliers tend to have a marginal cost curve which is fairly flat over a certain range, but then climbs like a sheer cliff when they hit capacity. If every supplier is producing at or near to capacity, no-one makes extra profit from producing more; price is then determined by the demand curve (I.e. at what price will customers demand the current capacity?) Further, at equilibrium, supplier projects which aim to increase capacity will show up as NPV negative (because they are not expected to recover fixed costs, capital costs etc.) So these projects won’t happen, and the equilibrium is maintained.
It’s assuming the fixed costs can be recuperated. Large fixed costs don’t mess up the equilibrium, if a firm can exit the industry and sell its initial investment at a comparable price to what they bought it for. It’s large fixed investments that can’t be liquidated that cause the problems.
In the situation you describe (price = marginal cost), it may be impossible for the firm to make a profit; all it can do is minimise its losses, so that it “only” loses its fixed costs.
The price=marginal cost is a consequence of homogeneous products (technically, of the fact that each individual firm faces a flat demand curve). The rate of profit is connected with ease of entry and exit. These are two separate things.
It’s assuming the fixed costs can be recuperated… if a firm can exit the industry and sell its initial investment at a comparable price to what they bought it for
That sounds like assuming no depreciation and no cost of capital, right?
Otherwise, imagine a firm considers increasing its capacity via investing in a productive asset, for an upfront cost A. There is a depreciation rate d, and a cost of capital (interest, dividends etc) of c. If the firm sells the asset after r years, then its sale value in year r will be something like A x (1 - d)^r, and the present terminal value will be something like A x ((1 -d)/(1+c))^r.
So in the business pan, the firm should assume a fixed cost not strictly of A, but rather of A x (1 - ((1-d)/(1+c))^r). If (d+c)r is roughly 1 or more then this will be about A; only if (d+c)r is much less than 1 can this cost be ignored.
As | understand it, it assumes no or low depreciation, but says nothing about the cost of capital. The definition of “profit” that I’ve been using is “extra profit to the company after all investors have been paid at the rate the market would demand, and employee and entrepreneurs have been reimbursed for their time and effort”.
What about fixed operating expenses, such as the insurance on the capital investment? (Or equivalently, the cost of the risk associated with owning it)
The real life example here is electric utilities. The way they’re regulated they charge a kWh price roughly equal to the average total cost (let’s say about 12 cents). The proper way to price would be at the marginal cost (at around 4 cents). The fact that marginal costs are below average total costs are what makes them a natural monopoly.
The somewhat obvious better solution would be to charge marginal cost for each kWh and then have some other method to collect the massive fixed costs. But for whatever historic reasons, we don’t do that and most (all?) utilities price each kWh at about the average total cost. This means that as a society our quantity demanded kWh is way below where economic theory says it should be.
However, there is probably a fairly substantial pollution/CO2 externality to producing electricity. Without some analysis it isn’t obvious whether we’re producing too much electricity or too little.
I did try once to look at estimates of the size of the externality to see if it made up for the pricing way above marginal cost issue and the preliminary results were that the externality was smaller (meaning, global warming considered, we’re still not using enough electricity). However, there were a couple of points I’d need to get into deeper.
1) The pricing above marginal cost issue is greatest for residential rates and smallest for industrial rates. I was looking at residential rates. Using the same cursory analysis on industrial rates would mean that we’re over using electricity in industrial sectors.
2) The carbon externality number I used from the EPA seemed to be derived by figuring out how high the price of electricity would need to be to get usage down to the level they wanted. Under correctly priced utility rates (i.e., priced at the marginal costs), their analysis may have had a much higher $ / kWh externality number. But at the same time, I’m a little suspect of that method of calculating the externality as it would indicate if the cost of production halved it wouldn’t be optimal for society to produce more. So I’d need to do some more research to make sure I’m using good pollution/CO2 numbers.
I haven’t seen this issue discussed by people like Mankiw when they talk about the Pigou Club and I think it probably should be. If there’s interest I could probably write this up a bit more formally and make it a post.
There exists much better work than that on power production externalities. ExternE, for starters. Which mostly prove that the amount of pollution has remarkably little to do with how much power you produce, and a heck of a lot to do with which technologies you use to produce them. Major takeaway if you do not care to read that “Coal is not a good idea, even ignoring the carbon.”.
A couple of points.
This doesn’t apply in all countries. In UK for instance, it is common to have a standing charge (flat fee per day) as well as a usage charge (fee per kWh). Or some utilities charge a high price for the first few kWh, and then a lower price for subsequent kWh, which has a similar effect. See here for some details.
Even where there is a single price (a price per kWh) it is not true that the “correct” market price is just the marginal cost. Suppliers do need to recover their costs of capital, and fixed costs, or they will go out of business. Imagine a market with a huge numbers of suppliers, where the price drops to marginal cost. They will all be losing money, but some will go broke quicker than others. As suppliers exit the market, the remaining suppliers find they can increase their price, and equilibrium is established when some marginal supplier is just hanging on in the market (barely making enough revenue to cover total costs). If the number of suppliers drops below this equilibrium, then they all start making large profits, but this situation should attract a new entrant into the market, so restoring equilibrium. That’s how the market theory works of course: real life situations create both barriers to entry (overregulation, obstruction of access to wholesale supplies, or to the distribution network) and barriers to exit (loss-making firms are propped up for years by subsidies, bailouts etc).
My perspective is US-centric, but from what I’m aware the per kWh price in most countries for most people is well above the marginal costs. Many places do have a daily or monthly charge but that tends to be $10 or less—not even close to high enough to recover all the fixed costs associated with a customer. Looking through some of the Scottish Power rates that you linked to, the daily charge doesn’t get much higher than 30p. That helps mitigate the issue a bit, but it’s still there. In that case, retail kWh prices—after the standing charge—is still over 10p / kWh. Wholesale rates look like they’re 4.5p in the UK (which should be a good proxy for short run marginal costs) so there’s still a big gap.
As far as I’m aware, economic theory says the “correct” price for electric utilities is lower than where the actual price is. It’s probably easier to visualize on a graph like this one. (I’m saying the difference between Pf and Pr, at least in some cases, may be higher than the externality, which is a real-life example of what the op is talking about). If that’s not standard economic theory though let me know as it’s an area of interest to me.
The market correction mechanism you described works for most industries but electric utilities are typically treated as natural monopolies, the optimal number of suppliers is one. But even if that ’s not true (i.e. it’s not actually optimal), in many places regulation only allows one supplier so the market forces described couldn’t work. The result is that the average /kWh price customers pay is higher than the average marginal costs (optimal society price) and it continues indefinitely because new firms can’t come into the market. There isn’t large profits made though because they’re pricing at the regulated price (at the average total cost) and not at the monopoly price (again, easier to visualize on the graph linked to above).
I think that, strictly, Stuart was arguing that the difference between Pm and Pf exceeds the externality cost, which may well be true. However, politically it is of course much easier to force a polluting monopoly to lower its price (to Pf) than to subsidise said monopoly still further. It is also economically more efficient (there are better things to do with public money).
You may also be right that the externality cost exceeds the difference between Pf and Pr—referring to the UK numbers, does the externality actually work out at less than about 5p per kWh? Even if it does, I’d argue that it is unrealistic to expect the price to drop to Pr and stay there indefinitely (while the suppliers go broke).
The “Correct” price for electricity is one price to be connected to the grid and several more relating to one’s power used, power factor, peak demand, and the like.
The average price paid is lower than the “correct” price, because charging the “correct” price adds lots of measuring and billing costs. It’s better to allow some subsidizing to be happening than to spend more just to make sure it isn’t.
An easy way to do it would be to charge the “correct” marginal cost for all kWh and have a separate fixed fee. My water bill is something like $50 fixed and then a small amount for the water I use after it; the electric bill could work the same. Ronald Coase argued that here
Commercial meters have priced kW for a long time and I think the reason residential didn’t was more along the lines of they’re more homogeneous than the meter costs. But either way, it seems everyone is getting smart meters now and you could match it up to theory exactly if it were politically feasible.
The political problem is that some people would be charged more; regression to the mean suggests that those people would be the ones who currently pay the least. The people who pay the least are the people who use the least power.
In a proper cost-sharing setup, dividing the fixed cost of maintaining each portion of the grid among the people served by that portion of the grid, pricing would be fiendishly complicated and appear unfair: Consider ten rural houses roughly in a line sharing a single branch line from the distribution station: each of these houses would be charged equally for their use of the larger distribution network, but the first house would be charged with 1⁄10 the cost of maintaining the first segment of their shared circuit, the second house would be charged that plus 1⁄9 the cost of the second segment, and so forth.
Or is it: If someone builds an 11th house at the end of that line, and the added load requires that the first segment be upgraded (to a line with higher maintenance costs) to handle the additional load, how is that cost fairly distributed? (What if the 11th house is added in the middle of the line?)
A $X+$y/kWh system makes more sense, but there is no system which is perfectly fair and appears to be fair to most people.
Everything you say is true, but your implied argument is flawed (you are implicitly making an “all A are B, all C are B, therefore all A are C” argument). If we had a fixed fee, and were discussing the possibility of eliminating it, your argument would apply just as well.
Sorry- that made much more sense as a lead-in to the self-redacted segment where I pointed out that the higher-spending (and presumed higher-income) users were subsidizing the poor, and suggested that might be a feature.
I get the difference in country perspective, and the difference between a regulated local monopoly (which I believe you have in US) vs a market structure with a regulated distribution network, but a large choice of retail suppliers.
Incidentally, there are 6 large UK suppliers, and multiple smaller ones, so this is far from perfect competition, but also far from a monopoly. According to this report retail margins are about 7 per cent, and the previous margin was only about 3 per cent. Even in a “good year”, the big six suppliers are making an average profit of around £100 per year on an average bill of around £1400 per year. So if the retail price dropped to 9 pence per kWh as opposed to 10 pence per kWh, the firms would probably be making a net retail loss.
While the current profit margin probably does encourage entry, I can’t see any way that the retail price could drop to 4.5 pence per kWh and still allow a supplier to make a profit. It seems quite possible that a new supplier could start up offering 5 pence per kWh, and would rapidly grab market share: the fact that no-one in UK is trying that suggests that it just doesn’t stack up as NPV positive.
In a competitive market, the correct market price is the marginal cost. If the market price is higher, the firm will profit by producing more; if it’s lower, it will profit by producing less.
That is ignoring the fixed costs.
In the situation you describe (price = marginal cost), it may be impossible for the firm to make a profit; all it can do is minimise its losses, so that it “only” loses its fixed costs. However, since suppliers are making losses, the number of suppliers will contract i.e. this is not an equilibrium situation.
True, there are idealized models of “perfectly competitive markets” where all fixed costs, costs of entry, costs of capital etc are negligible. In that case, price will stay at marginal cost in equilibrium. But “competitive markets” are a bit more general than “perfectly competitive markets”.
A slightly more detailed analysis looks at the “capacity” of the suppliers. Suppliers tend to have a marginal cost curve which is fairly flat over a certain range, but then climbs like a sheer cliff when they hit capacity. If every supplier is producing at or near to capacity, no-one makes extra profit from producing more; price is then determined by the demand curve (I.e. at what price will customers demand the current capacity?) Further, at equilibrium, supplier projects which aim to increase capacity will show up as NPV negative (because they are not expected to recover fixed costs, capital costs etc.) So these projects won’t happen, and the equilibrium is maintained.
It’s assuming the fixed costs can be recuperated. Large fixed costs don’t mess up the equilibrium, if a firm can exit the industry and sell its initial investment at a comparable price to what they bought it for. It’s large fixed investments that can’t be liquidated that cause the problems.
The price=marginal cost is a consequence of homogeneous products (technically, of the fact that each individual firm faces a flat demand curve). The rate of profit is connected with ease of entry and exit. These are two separate things.
That sounds like assuming no depreciation and no cost of capital, right?
Otherwise, imagine a firm considers increasing its capacity via investing in a productive asset, for an upfront cost A. There is a depreciation rate d, and a cost of capital (interest, dividends etc) of c. If the firm sells the asset after r years, then its sale value in year r will be something like A x (1 - d)^r, and the present terminal value will be something like A x ((1 -d)/(1+c))^r.
So in the business pan, the firm should assume a fixed cost not strictly of A, but rather of A x (1 - ((1-d)/(1+c))^r). If (d+c)r is roughly 1 or more then this will be about A; only if (d+c)r is much less than 1 can this cost be ignored.
As | understand it, it assumes no or low depreciation, but says nothing about the cost of capital. The definition of “profit” that I’ve been using is “extra profit to the company after all investors have been paid at the rate the market would demand, and employee and entrepreneurs have been reimbursed for their time and effort”.
Who would want to buy the equipment to get into an industry with zero profit opportunity?
What about fixed operating expenses, such as the insurance on the capital investment? (Or equivalently, the cost of the risk associated with owning it)
I’d like to see that post. The methodology alone would be very interesting!