Andrew D. Smith has a question

There are a lot of companies out there trying to help consumers reduce energy consumption, theoretically to save money AND the environment. The NYT had a long story on a smart thermostat company today:

Can you discuss whether this can possibly work? As I understand the power industry, such a high percentage of the costs are upfront (with nuclear plants  in particular, but with carbon burning plants as well) and the marginal price of producing energy (up to plant capacity) is so low, that falling demand would  mostly cause plants to cut prices until they were again operating at capacity.

So “saving” energy at the consumer level won’t really reduce total energy consumption or gas emission.

Is that right?


Perhaps true in the short-term. However, this will lead to less building of new power plants.

Is he really arguing that lower demand will not lead to lower supply? My mind is LITERALLY boggled.

Vertical supply curve. Changing demand will only decrease price, not quantity supplied. In the long run, supply is more elastic. How elastic, however, depends on how long the long run is.

The life of a power plant is measured in high decades. All operating nuke plants in the US are between 30 and 60 years old. Wiki says that the oldest coal plants in the US were built in the '40s and late '30s, making them 60-70 years old. Given that, how long does the long run have to be to make the supply curve elastic?

Vertical supply curve *in steps*.

A plant is either feeding the grid, or it's not. Average plant size is probably 300-550MW (per unit). Less demand, you shut units down (highest cost first).

Ok, but "there are about 5,700 operational power plants in the United States with a nameplate generation capacity of at least one MegaWatt." Even with a long lifespan of each power plant, there must be dozens of decisions made each year about whether to keep existing plants open, and about whether to build new ones. The level of demand for electricity affects all of these decisions.

Lots more data here if people want to analyze further:

dan1111 - plants remain open for two reasons. capacity and energy. in pjm a plant never has to run during a year and it can still earn 90k per mw year (depending on what zone it is in).

I understand that, but lower energy consumption also reduces the peak capacity needed.

Depending on the source, the variable costs of power production are high.

See for some numbers. Note that these numbers are highly sensitive to fuel prices. The DOE reports give the heat rate and fuel price assumptions.

It could even lead to higher prices in the short term as companies need to repay their existing building costs debts from a smaller quantity of electricity sold.

Exactly. I seriously doubt that an aggregate reduction in consumption will ever equal a concomitant aggregate reduction in prices charged. Not from a utility company.

They'll simply raise the per unit cost.

LIPA had an interesting line item in an annual report from (I think) 2008. It was asking the PUC for an increase in prices due to a decreas in the amount of power consumed (and therefore less rev to cover fixed costs).

As far as I understand the device it won't increase or decrease energy use based on price. Intelligent thermostats are designed to reduce energy usage. The cost saving is a side effect. I don't see how reduced consumption would not inevitably lead to decreased production.

No. For one thing, very few consumers heat their homes with electricity. The marginal cost of natural gas used is the marginal cost of the gas. But I don't see a big future for the smart thermostat. You'd need not only the thermostat, but also motion detectors and electronically controlled dampers for all your ducts all wired to a controller (as well as the contractor cost to install it all). So we're no longer talking about a $250 thermostat, but at least several times that much. And heating a room can't be easily switched on and off as you move through your house -- when you walked into a room, it'll be cold and wouldn't be heated up until you'd already left.

In the US south east especially, I believe a common, even majority, form of heating/cooling is the air sourced heat pump?

'zoned' heating systems are recommended best practice in the UK (air con is still almost unknown and most heating systems are hot water rads or in-floor). That would be similar to a 'smart' thermostat.

Specified from the beginning I suspect not too expensive. $500 on a new home is neither here nor there. refurb that's a bit different.

If you want to save money below the mason-dixon line, turn off your heat pump and throw on a sweater. With decent insulation, you house isn't going to freeze.

That sounds like a quick way to get divorced or to become a one term president.

Other ways to save money:
Honey, you don't need a car, you can bike to where you need to go, and frankly you could stand to lose a few pounds.
Kids, yep the pizza tastes good, but the plain pasta with oatmeal and waters is only pennies per meal.
We don't need a TV or the internet, books and magazines are 'free' at the library.

"‘zoned’ heating systems are recommended best practice in the UK": pretty widespread, I think. Otherwise you'd be heating your bedrooms to living room temperatures, which would be daft, and heating everything to the same extent whatever the time of day or day of the week. It's also common to pay different prices for your electricity according to the time of day.

"air con is still almost unknown": and likely to remain so in houses. It makes little sense in our climate(s).

Electricity is priced nodally at the wholesale level. Dispatching heat to alleviate transmission congestion would be a very simple solution to some grid issues. For instance, right now (~1030 ept on 10/28) if you were living in eastern Iowa around Dubuque, you could be buying electricity from the grid for a negative $200/MWh (ie the grid would be paying you to take the electricity). Dispatchable thermostats could save hundreds of millions in congestion costs annually (which are completely independent of energy savings).

Thermostats also control cooling which is mostly electric.

I think you're right for nuclear, geothermal and renewable generation. That being said, in many markets the marginal generators are flexible thermal gas fired plants. For these plants, the marginal costs are quite high relative to the fixed cost. Also, increasing the output of a thermal plant increases the rate of depreciation and can decrease the efficiency at the plant so the marginal costs are increasing in the output, and the most profitable output at a given plant will be dependent on the price.
This is kind of like how your car will depreciate more quickly if you are always accelerating as quickly as you can after every stop. You have to make the tradeoff between potentially getting where you want to be a bit quicker, or driving more efficiently, using less gas and requiring less maintenance.
So you do get fairly elastic supply in most electricity markets, even in the short run.

good analysis.

It is also worth stating that in many markets the marginal generators are higher contributors of greenhouse gases, and therefore the environmental impact of marginal declines in demand can be significant. Two reasons:

1) The types of plants with the highest fixed to variable cost ratios are the least carbon-intensive - nuclear and hydro
2) If usage declines, and so does price, there is a price at which the most-expensive plant cannot produce profitably. The most-expensive plant is likely to be the least efficient, and therefore the oldest and dirtiest of the coal or gas generators.

Hydro can usually flex output (but not in peak runoff or dry seasons).

Marginal power:

- peak power is usually open cycle gas turbines on coal fired stations on 'spinning reserve' - fuel price does indeed matter
- Mid merit is a mixture of combined cycle gas turbines and coal (depends which utilty, where, when)
- baseload is hydro + nuclear

The oldest plant is, oddly, the cheapest. Because of the nature of accounting depreciation-- those plants have been written off (that is more or less what has happened to the nuclear industry). They price at marginal cost of production (which can be very low). EPA 'New Source Rules' have been distorted so that many of the coal fired plants in the US are 40+ years old (and so meet lower EPA standards).

In cash flow terms that is nonsense, but nonetheless many utilities are profit regulated and most are investor owned, so they earnings maximize.

So in cash flow terms the oldest plant is likely to be the least efficient, but it can also be your most profitable.

Not sure it is true that most utility companies manage for GAAP earnings rather than cash flow, but this is veering very far off topic....

But if their cash flow is driven off their GAAP earnings (regulated situation) then it can pay to run older, less efficient plants?

I don't trust the numbers. The article claims each degree cooler you set your thermostat results in 5% savings in heating season. That seems wildly inflated.

Well, Fourier's law states that heat transfer is directly proportional to the negative gradient of temperature. Basically this means that heating power required by a house is directly
proportional to the different of indoors and outdoors temperature. If the difference
is around 20 degrees Celsius, then setting the thermostat to one degree cooler indeed saves 5% energy. If the different is smaller, the (relative) saving is even greater.

Given that the article is in the NY Times, it's likely that the scale is Fahrenheit and therefore your theoretical savings estimate is 3% at your 36° difference. You also need to integrate that equation over the entire heating season.

Alternatively one could look at the many studies of programmable setback thermostats that show little to no savings when used (or misused) by most homeowners. In many homes it takes a significant amount of time for the heating system to recover from a setback of more than a few degrees. Any smart system that results in users frequently walking into uncomfortably cold rooms and waiting tens of minutes to get comfortable is going to be manually overridden a lot.

This is wrong even in the short-term. Power supply is made up of baseload, which is the steady supply that is difficult to quickly spin up or turn off, and operating reserve, which is supply that can be dispatched fairly quickly to meet fluctuations in demand. (This is an oversimplification, but good enough for present purposes.) High cooling load (i.e. high outdoor temperatures) are generally coincident with peak demand, so lowering your thermostat lowers the need to spin up operating reserve.

This is the principle that underlies demand response, a fledgling industry that provided about 7 gigawatts of energy efficiency this past summer. 7 gigawatts is a small portion of the U.S. power supply, but it's not inconsequential, particularly given that this is all marginal capacity supplied when the grid is brushing up against capacity constraints. And there is much more in the way of savings like this to be had.

demand response is standard in industry where contracts often specify time of day power pricing.

All consumer electronic technology is allowing us to do is roll that existing stuff out to homes.

There, much depends on regulation. In a fully regulated environment, reduced consumption in the short term might lead to increased cost per units. But there would be long term savings in power station and transmission line construction-- lower capital cost, lower rate base, lower utility bills.

In a deregulated environment the response could be much quicker. Lower demand means lower prices and hence producer exit. For the peak load providers (gas turbines) fuel is a very significant part of Operating Cost.

Isn't this a straightforward question about elasticity?

More than one elasticity.

Price elasticity of supply is very high. Because above the baseload (for sake of argument the 4.30am demand is about 40% of 4.30pm demand so that's lo to hi) there are high variable fuel costs in much of the generation mix (coal and gas, the latter particularly).

Price elasticity of demand is short term quite low, because consumers face fixed retail rates, most rates are not time of day, and most consumers don't have a close awareness of how much an individual activity spends (eg that c. 10% of your consumption could be the 'instant on' features of appliances, plus routers, set top boxes etc.).

Longer term one suspects price elasticity of electricity demand is quite high. Not above 1, but probably something 0.5 to 1.0.

Note income elasticity of supply is also a strong factor.

Income rises, propensity to have more electronic gadgets and bigger homes with more lights (and higher AC loads) increases as well.

I was happy to hear about this, as a first step -- my smart thermostat was really like going back to the 80s, but it did save about half the gas bill, compared to the circular mercury-switch kind, in an apartment with two forgetful people. Something nicely designed and easy to use would probably get us to a better program, and perhaps get adopted by more people.

Of course for gas and oil heat you burn the same fuel whatever the time of day. (Although perhaps it could be smart about knowing how efficient the boiler is at various throttle settings?)

For electricity, yea I can't see how a single-room control like this would be able to take advantage of the variable cost of power. But perhaps, further down the road, non-awful controls could make thermal storage systems more attractive...

Boilers don't necessarily have different settings (the more sophisticated ones can 'phase' up and down).

But in industry sophisticated boiler controls that 'learn' about daily usage patterns, and monitor outside temperature to optimize boiler function, are the norm. You can get them for homes, but they would work best if you have a quite predictable pattern of use (and your home stays fairly warm if you are out: ie well insulated).

Is Jevons paradox a factor here?

Not really. Or at least not directly.

Just because it costs you less to heat your home does not mean you heat it more *if* you already heat it to the desired temperature.

Depends what else you do with the money. If you buy a 3rd tv and leave it on instant on, yes.

While you would not overheat a room, there may be less frequently used rooms you currently keep cooler that you now longer keep cooler than normal.

If you're already heating the house, then surely the 3rd TV is running at 100% efficiency. That is, any extra energy that the TV uses while it's not producing pictures or sound is converted into heat, and so the boiler has less work to do. Assuming you have electric heating, this should exactly cancel out.

Yes although *where* the heat source is matters.

eg electric lights are heaters, but they heat the *ceiling* (and the rooms above it, if any).

TV waste heat is waste heat, but it's over in that other corner, some will radiate into wall, ceiling, not all gets convected around the room.

I doubt many people in the US heat with electric bar heating in any case? Heat pumps or gas, primarily? (or oil, wood, propane if off gas network).

Well, heating is just one example of power use. And my guess is that its a pretty small part of the use of power from the nuclear plants mentioned.

So i'm still not sure about how much Jevons paradox comes into this. For electrical cars i would guess it is quite a large factor, but that is still a small part of the energy use. Household appliances... Would we use the owen or dishwasher more if would use less power, probably not.

Lighting, perhaps a bigger factor. It should be interesting to see the effects on the ban of the old type of lightbulbs (in favour of new energy efficient ones) in Sweden. Will the energy consumtion really be lowered with the difference between the power use of the old lightbulbs and the new one. Or will people use more light, or not be as particular with turning it off when they leave a room.

Even if total production was constant, the electricity could be used for more valuable applications. It seems like an exercise in picking your price: cost, plants, usage, application, etc.

It's not like you can assume that because they install a smart timer that therefore the price would be so low that they wouldn't install the smart timer. So, the electricity would be used for something more building more houses in Phoenix.

Just a general point about generation.

US power breakdown is c.

- 20% nuclear - so production does not flex with demand (negative pricing is even possible at low periods ie pay the grid to take the stuff away)
- 8% hydro - highly flexible with demand, but not during spring runoff
- 45-50% coal - fuel is a significant cost. Individual units do not flex with demand but every generation company retains a schedule (low cost to high cost) of plants that are brought online (from spinning reserve) to meet demand spikes. So the load curve moves in 'steps'. Usually individual units are c. 550mw (an optimal scale achieved in the 1960s and not changed much since, I don't think. A 3000 MW station is just 6 500mw units) so those are your 'steps'
- balance is mostly gas where fuel cost is high and the key is the 'spark spread' between price and cost. They will almost certainly respond to demand (and very quickly)

Absolute peak supply is Open Cycle Gas Turbines (most gas stations are Combined Cycle, with efficiencies of 50-57%, OCGT is c. 30%, ie a jet engine on its side). Very sensitive to fuel cost and hence electricity demand given low efficiencies. (your alternative is to keep coal fired stations on spinning reserve, which is polluting but widely used)

Your posts have a lot of great information. Thanks.

It's more accurate to talk about the US as three power grids -- the Eastern, Western, and Texas Interconnections -- rather than one, as there are very limited power transfers between those. The Western Interconnect has quite different characteristics than the other two. The West includes about 23% of the US population, but only consumes about 9% of the electricity. Sources are quite different as well. The West is (in round numbers) about 30% gas, 30% coal, 23% hydro, 9% nuclear, plus bits and pieces of other things. The non-West is 50% coal, 20% gas, 20% nuclear, plus the other bits. High-quality undeveloped renewable resources tend to be concentrated in the West, as well.

Quite different situations, probably requiring quite different policy approaches.

I know something about the electricity market and there is a lot confusion here. Adam has the best answer so far.

The first confusion here is that the quantity of electricity generation supplied depends somehow on the economics of the generator. It does not. No matter what, the generation has to balance the load. If there is 5 MW less demand because users are more energy efficient there will be 5 MW less generation dispatched. After the generation is dispatched, the prices will be calculated using the generator's bid curve. That price could go very low, even negative. It does not matter. The dispatch is the dispatch and the generators are obliged to follow it.

The marginal cost of fossil fuel units is not all that low. It varies from unit to unit but coal units and combined cycles have a marginal cost of roughly $30/MWh to $40/MWh in most places in the US right now. So if a generator is asked to back down, the marginal cost will determine the local marginal price (LMP) for electricity, and this turns out to be about $30/MWh to $40/MWh in most places. It goes much higher during the peak four of so hours of load during the day.

The fixed costs of a generator are only relevant when a company thinks about entering a new generator into the market. The company has to see if the LMPs and the capacity payments are high enough on average to cover the fixed costs and operating costs of a new plant. Older plants are retiring all the time so there are opportunities for newer plants, with better technology and perhaps a better fuel to come in a make a profit.

"That price could go very low, even negative. It does not matter. The dispatch is the dispatch and the generators are obliged to follow it."

I'm not sure how your market works, but in a competitive market with centralised trading it would not make much sense for generators to produce at below their marginal cost. In a monopolistic environment, the regulator might be able to force an "obligation". In a competitive market who would decide which share of demand should be produced by each producer?

because of fixed costs of startup, many generators, in the right circumstances, run below their short-run marginal costs, because shutting down and starting up again is too expensive. There's currently a number of battles over this, as wind-powered generators drive prices below 0 at night. (it's worse in Spain, where wind is a higher fraction of total generation but getting worse here.) They get a production tax credit, so they make money (net) but nuclear plants and some coal plants can't afford to shutdown and start up again, so they bid below their own short-run marginal costs to stay on line (at minimum load levels, which are 20 percent or so of total output.)

You are correct that in general, everyone who is operating is operating at a cost that is less than or equal to the final price (there could be some operating out of merit but they will be compensated separately). However, the electricity market does not work on the principle of suppliers seeing a price and then deciding how much they want to produce. It works like an auction (actually, it is an auction): people bid and based on demand (load) and the bid curves, the dispatch software determines the lowest cost generators and afterwards the software determines the prices that would make sense given the bid curves (which are the marginal cost curves of the generators, in theory).

The final price would equal the marginal cost of the last generator to be scheduled either up or down. The prices could go negative if all you have in the market are generators who are incapable of ramping down (generating less electricity) any further in the short term and wind turbines that prefer to run even when the prices are negative. Then the marginal unit is a wind turbine that might agree to pay to generate. Of course, the only reason why the wind turbine would pay to generate is because the government, in its infinite wisdom, is paying the generator to operate in all circumstances, including times when their output is of literally less than zero value.

Law of the Second Best.

Since we don't carbon tax the carbon emitting producers, we subsidize the renewable ones. UK is going to a feed in tariff (a system of tradeable permits (ROCs) arising from selling clean electricity has not worked in triggering entry too volatile to make capex plans).

As another participant in the energy markets, I'd like to second Michael's responses. Prices routinely go negative at night in Texas when there is a lot of wind power in the mix; the fossil fuel plants essentially bid against each other to determine who has the lowest costs to turn their systems off and on.

Aside from misunderstanding electricity markets, Andrew Smith's logic is flawed with regard to the consumer's response to changes in the electricity price. American residential electricity demand is both inelastic and insensitive to market signals in much the same way that the gasoline market is, but to a greater extent. It's not just that nobody makes a decision to watch TV based on the price of power, nobody even knows how much watching their TV affects their energy bill and so they can't even start down a path towards making that decision.

Actually gasoline is ST price inelastic (c. 0.1) but long term fairly price elastic (estimates 0.3 to 0.8).

People can, and do, take into account gasoline prices when they buy their next car. But they only buy a car every few years. The average auto lasts 13 years and so it takes a long time for the fleet fuel economy to rise.

And consumers do know how much their drive costs them: gasoline is one of the most salient costs (in fact *the* most salient cost) in a shopping basket-- studies show consumer confidence is directly inversely correlated with oil price spikes.

(gasoline consumption rises over time due to income effects, which are of course positive, few goods have a negative income elasticity)

Over to electricity. Agree much less sensitive, in fact virtually insensitive in the short term, *except* places like California where time of day pricing and usage pricing has come in.

Electricity is normally a lot smaller part of the monthly bill than gasoline, so income effects overwhelm.

Studies of energy efficiency suggest it is not low income users who are inefficient in their use of energy, it is middle and high income households-- where the income effect overwhelms any price substitution effect.

OK though our Gencos rank their stations and 'bid them in' to the Pool, as I understand it (UK practice).

In addition most electrricity distributors here own generation assets-- 6 vertically integrated companies own 90% of generation. So they would rigorously price against cost of production.

We no longer have capacity payments.

Nest's proposition is that a lot of the energy we use is simply wasted and so we can maintain present or greater levels of comfort without proportional increases in energy use. If gas and electricity tomorrow dropped to zero, I would not suddenly start heating my house to 300 degrees or cooling it -70. The marginal value of electricity drops precipitously once I've consumed my fill of complementary goods. In the long run, I could buy a larger house but whether this increases pollution will depend on whether the afforded increase in house size will, with the new efficiency, require more energy than my former abode. That doesn't seem likely.

Buying more consumer goods that require large amounts of electricity is possible, but even fairly wealthy households avoid additional appliances; I don't need three dryers and I'd have difficulty filling (or accommodating) some new refrigerators. What seems more likely is that I'd invest more heavily in electronics, but because they eat through the savings faster than they add to energy use, my energy consumption still falls.

Business, however, could use the cheaper electricity more immediately, but this seems like a net win. There are two channels I see: expanded production, increasing reliance on central generation. In the former case pollution per dollar falls, in the latter case pollution falls absolutely because of the greater pollution efficiency of central generation.

It's all about aggregate demand. One person using less electricity doesn't reduce the amount of fuel burned in a power plant. This is because peak load plants (powered almost exclusively by fossil fuels), which are the "first responders" to changes in demand, will continue to burn despite a drop in demand (within a certain range). Remember the physics of electricity: supply and demand must be in perfect equilibrium. This equilibrium is maintained by regulating how much electricity generated by the spinning turbines is "connected" to the grid (think of a car engine that's always idling, and only put in gear when a light is turned on). If aggregate demand drops below a certain threshold, then a fossil plant may not power up as often, and less fuel is burned.

Price, however, may not change at all, because price is driven more by the cost of fuel and operations than by electricity demand (our national power system is a holy mess of regulated and unregulated businesses). Lower aggregate demand will NOT necessarily lead to lower electricity prices; it will, however, lead to fewer externalities produced by burning fossil fuels.

So, the short answer is: if aggregate demand drops significantly and permanently, then fewer power plants will be "on" and the environment will be at less risk. The effect on price is unknown, due to the way price is determined today.

In a world with just solar and wind resources marginal costs would be zero. In that case saving energy would'nt make sense, at least in the short term. As long as there are fossil fuel plants I agree with the other comments that costs at the margin are quite high and that energy savings result in less energy consumption.

Without energy storage that world is impossible at least in the sense of lights on at night, reliable electric power etc.

And if you have energy storage, presumably there would be a cost to it. And a reason to respond to short term price signals (because the energy might be more valuable in 4 hours).

Obama's EPA is shutting down scores of old coal plants, like Stateline Generating in Indiana (on the border of Chicago). We need massive energy efficiency increases to absorb the loss of this capacity (and, to be fair to Obama, Stateline is 60 years old and dirty relative to new technology, not to mention energy efficiency).

gas mmbtus are now cheaper than Central Appalachian coal mmbtus...this is a big reason that a lot of the older coal plants are shutting down. New gas cc's have picked up the slack in total generation capacity...we don't have to have new efficiency increases although we will always.

The reason those old coal fired plants are still around is 'New Source Rules' ie they didn't have to meet the emission standards of new plants, so the utilities just kept them running-- a regulatory loophole.

That and in regulated markets the reluctance of regulators to pay for new utility capacity. Easier for a utility just to run what it has.

Programmable thermostats (like the Nest) and energy efficiency (EE) are probably two different things with two different effects:

tldr: Programmable thermostats shift load so no reduction in pollution but could save you a bunch of money (on a Time-of-Use rate). Energy Efficiency probably raises P and lowers Q thus reducing pollution, unless Jevons paradox causes D to move right which doesn't seem likely.

In Phoenix, we have a lot of programmable thermostats and they don't really reduce load, they shift it. This saves quite a bit of money (going from onpeak to offpeak prices). But because the load just shifted from afternoon to morning, probably not much of a reduction in pollution. Maybe the Nest will actually use less electricity because of even smarter use, but I'm skeptical that it would be enough to show up in the data. A thermostat that was trying to save the most money would precool the house during off peak periods so that the AC didn't run during onpeak. Currently, people aren't overcooling in the morning such that the house is colder than what they wanted in the afternoon, they're undercooling and the AC runs during the onpeak period anyways. So a smart, cool, fun thermostat that figured this out and did a better job precooling would save a bunch of money, but insofar as it only shifts load, won't reduce pollution (much at least, the peak generator that gets run less is probably a bit older and dirtier than the intermediate one that gets run more).

On the energy efficiency side, keep in mind that we're not looking at perfect competition, we're in a natural monopoly where marginal cost doesn't really have anything to do with the price. The utility prices at the average total cost (which is always downward sloping and 2-3x higher than MC). If the demand curve shifts to the left due to EE, price goes up and quantity sold goes down (recall all the talk about decoupling). Almost guaranteed, the decrease in Q comes from a natural gas plant, so pollution is going down.

This is assuming the demand curve does shift left due to energy efficiency. Jevons paradox is almost certainly at play here. (To the person with EE, electricity appears cheaper so they may end up using more of it than before). I haven't seen a good analysis of this, energy efficiency advocates ignore it completely assuming that if the EE AC uses half the power then the customer will use half the power (instead of cooling their house more because it is cheaper). Because electricity is so inelastic, it would surprise me if EE actually caused increase load, but my best guess is that it mitigates the decrease some. Either way, I'd really like to see some good actual data on it before the money is all spent.

"...falling demand would mostly cause plants to cut prices until they were again operating at capacity."

Thinking of the electricity price in a given area as a "market clearing" rate in a perfect market (all electricity is identical, etc), that's not quite how it works. For a given quantity of electricity demanded, you stack up the producers from lowest marginal cost to highest (usually something like: solar, nuclear, high efficiency coal, low efficiency coal, gas, fuel oil) and the price is equal to the marginal cost of the marginal producer. If demand falls, price falls and marginal producers turn off until quantity supplied = quantity demanded again. Of course, there's a countervailing effect where quantity demanded rises a little w/ lower prices, but consumer demand is relatively inelastic unlike supply.
Notice that with marginal plants shutting in, you burn less gas or coal, lowering emissions and lowering the price of those commodities.

Demand grows over time (w/ GDP, population) so new plants have to be built. It's an economic win if you can delay building new plants, transmission lines, etc. Assuming those plants would burn fuel (gas, coal, uranium), you need less of those resources to supply electricity, lowering marginal costs for the existing producers ceteris paribus.

Good comments by Alfred and valuethinker. The marginal supplier in most regions is gas turbines, which sell pretty much at marginal cost already. A quick way to check this is by multiplying the price of natural gas by the heat rate of the marginal gas plant in that region, and comparing that to the wholesale price per MWh of electricity - they will usually be pretty close.

For what it's worth, U.S. industrial and commercial load is still below 2007 levels; energy efficiency standards from EPAct and EISA may be one factor, although obviously very difficult to parse out the effect of energy efficiency from other drivers like, say, the economy.

Difficult to see that short term conservation measures would have much impact compared to the downturn.

Exception would be transport fuels consumption, where businesses could see the P&L improvement very quickly of better scheduling, shipping full loads, more efficient modal choice etc.

AFAIK US utility bills are not up much so companies don't have the immediate P&L impact.

Most organisations probably can save 10-20% on their energy bill by being smarter about it (turning lights out, low energy bulbs, more efficient use of equipment etc.).

There are two government regulatory effects which will hinder transmission of price signals to consumers. First, most states regulate the price charged for electricity in some way, and shield consumers from short term price spikes while preventing consumers from realizing savings from short term price drops. There are also, in some areas (California, at least) incentives to electric utilities to encourage conservation. I'm not sure how those apply, or to whom, since the botched deregulation of 1996-2000.

Major value: avoided costs of building new generation. Mind you, this is not a *forever* thing--aggregate demand continues to increase, but per capita consumption stays steady or declines slightly. So lower demand (either in general or at the peak (i.e. marginal unit)) really means slower growth in demand. Delaying the building of power plants means less money spent right now and less money spent over a given period of time--which is a LOT of value.

GHG emissions: depends on your grid mix. A general rule is that less generation = less emissions, so avoiding building new plants can only help. That said, if your marginal unit is wind/solar/geothermal, then new generation doesn't mean more GHG emissions (aside from the upstream emissions from manufacture). Mind you, at least in California (where I'm familiar), the loading order -- the ranking of ways we make supply and demand meet -- specifies things like energy efficiency, demand response, and renewables before you get to building new natural gas plants. So it's possible you could raise the option of building a new gas plant, but presumably under CA law that's not going to happen anyway for some time. (However, CA does import from other dirtier states when it needs to, so lowering that would definitely cut emissions.)

Prices: depends on your market. Utilities are regulated monopolies, and by law (at least in CA, where I'm familiar), utility revenues are decoupled from kWh sales. In a really basic sense, prices are set through a negotiation between the utility and the regulator; prices are based on overall average costs to allow utilities to recover costs and have some profit. If demand were to fall in the VERY SHORT TERM only, this would cause problems with cost recovery. However, utilities project out demand at least a yeah ahead (usually 10 years out), and so energy efficiency measures like new thermostats would be taken into account -- or rather, the regulators and utilities would haggle over the expected demand reduction from such devices until a given number is set. So yes, it's possible that lower demand with existing resources means higher prices in the short to medium term...but then you have less need for new generation, and so those avoided costs of generation kick in at some point and keep prices lower than they would have been.

However, if you care about the rising price of energy, small increases from lowered demand are NOT what you should care about. Building out renewables is WAY more expensive. Dealing with volatile gas prices is WAY more expensive. Dealing with new T&D transmission cost it WAY more expensive.

Think I found the basic flaw in Andrew's initial economic logic: he's assuming that the consumers respond to changes in price as if the market is perfectly efficient (generating Jarvon's paradox) while assuming an inefficient power generation/supply market which allows energy to be sold at a price above the marginal cost of production. Reality shows almost exactly the reverse behavior: deregulated energy markets are reasonably efficient and clear close to the marginal cost of production (the price of gas, coal, or renewable energy depending on the market and time of day) and consumers respond very inefficiently to incentives in the realm of energy use.

Isn't people using more power for the same cost just as good as people cutting their power usage? It is not the case that our ultimate goal is to reduce power usage. The ultimate goals are to reduce pollution, and to reduce energy costs. Getting the same energy output at a lower price is a net win for the good guys.

From an engineering point of view, electrical generation is divided into two categories: base-load and peak. Base-load generation is provided by nuclear power and coal burning plants. These plants are expensive, slow to start up and shutdown. Peak demands are met by bringing gas-turbine plants on line. These plants do not require long start-up and shutdown times. They are literally turned off and on as demand rises and falls. Since air-conditioning is a part of peak demand, this device (if successful) could lead to greater idling of peak energy capacity.

Just to add, really you have:

- baseload - nuclear, hydro, whatever wind comes in, coal fired (depending on utility), some (little) CCGT

- mid merit - depending on utility primarily coal + CCGT

- peak - stored hydro, Open Cycle Gas Turbine (gas or fuel oil). You can also run coal on 'spinning reserve' and bring it on line quite quickly

Coal fired stations take hours to run up. Nuclear shutdowns are scheduled months in advance, so nukes are never as backup power.

Yes if you cut off peak demand (or shift it) you are getting rid of OCGTs running at 30% efficiency and replacing it with coal fired power running at 30-40% and CCGT at 50%+. In extremis, if there is nuclear or hydro around, you are replacing it with zero carbon emissions.

From the consumer POV, it doesn't matter if "total consumption" is affected, as long as you, the individual consumer, save money.

(Semi-relatedly, I approve of that Leaf thermostat mostly because it appears to be superior in usability and functionality. I don't give a fig for the other concerns at this point.

And I think Sohier's right, at least about how imperfectly consumers respond to marginal changes in utility prices.)

It isn't just reduction in demand saving the cost of the energy. A major expense for Utilities is the cost of additional capacity. The capital costs are enormous, so the 'avoided' costs are quite valuable. Think load leveling as another type solution which reduces cost without reducing consumption.

If the AVC for additional power production up to plant capacity is zero (i.e no impact on plant depreciation, maintenance costs etc.), then yeah, the plant will be operating at capacity regardless of usage... might as well just send excess charge to ground.

BUT that doesn't imply that power consumption is set in stone. Say you have a town with one power plant that's operating at capacity; all the power its producing is actually being used. Let's also say the town's Gini coefficient is very high, and that the rich people in the town are leaving the air conditioners on when they're not home, and that as a result poor people are getting priced out of the market to some extent. Instead of lighting their homes at night with electricity from the power plant, they're buying candles at the store.

If the rich people were to install the timers, and if we make a bunch of assumptions about the energy market in the town, then the plant's output stays at 100% of capacity but less energy is used in manufacturing candles, yielding a net reduction in power consumption. Ftw.

But cost of another unit of power is *not* zero. Because that unit comes either from burning gas, or burning coal (tiny amounts fuel oil).

If consumption falls then fuel cost falls across the whole system, which is reflected in lower prices (in a purely competitive situation) or a mix of that and higher profits (in an oligopoly).

It strikes me that most people don't base their electricity consumption on the price. They use what they're going to use, then pay whatever it costs. If energy-saving technologies became widely adopted to the point of actually decreasing usage, I would expect a utility company to respond by *raising* prices. They would have no new large up-front costs, since no new plants would be needed. However, they would need to respond in some fashion to the revenue hit a decrease in electricity consumption would represent.

In a municipality where there was already a strong upward trend in electricity usage (such as when the population is growing or incomes are rising at a strong clip) then widespread adoption of efficiency measures might simply counteract the upward trend, thereby eliminating the need to build new plants.

Utilities might seek to recover *fixed* costs arising from lower consumption (cost of generating station construction, distribution and line charges)


*variable* costs, no. If they save fuel with less demand (which they will) then they have lower costs. The regulator is not going to allow them to pass that on to a consumer, and in deregulated power markets if they tried, a competitor who does not do so could steal their market.

My assumption is that the cost to the consumer is a mix of fixed and variable, but that the pricing structure doesn't reflect that. That is to say the pricing is probably structured such that it's purely variable. You use more you pay more, albeit with progressive brackets based on usage. Since the fixed cost isn't built into their rate structure, a decrease in overall usage directly reduces their revenue. Most likely by a larger percentage than the actual decrease in usage, given the progressive rate brackets. They've got to recoup that money somehow. Either they introduce a fixed fee charged against all customers (regressive) and make their price structure accurate reflect the nature of their costs, or they raise their per-usage rates. Or possibly lower the cutoffs for each price bracket (which is an effective fee hike).

A drop in usage reduces their revenue

But it also reduces their costs-- by the cost of fuel. Which for peak power (Open Cycle Gas Turbine- gas or fuel oil at 30-35% efficiency) is considerable. It's not trivial even for coal at mid merit.

So yes, maybe, short term they have a cost recovery issue. But they do have lower generation costs as well.

Comments for this post are closed