Electricity consumption and autonomous control: technology and prices to devices for decentralized coordination

Lynne Kiesling

One of my favorite things about our house renovation is the spiffy, brandy-new Kitchenaid refrigerator that you can see in the corner of this kitchen photo:


It’s well-designed, it’s energy-efficient, it keeps our food and drink beautifully … and, because I am particular about mineral and biological tastes in water, I love its built-in water filter for both water and ice cubes.

Yesterday morning as I was pouring water for tea, the “filter” button on the door display turned orange, and the display now shows the message “order filter”. Without my having to do anything, the designers of the refrigerator programmed it to communicate this information to me when the remaining filter capacity fell to 20%.

In my Pavlovian way, what do you think I did? I naturally looked up the part number and did an online search for the filter. I found a filter store with prices that are 5% off of retail, I don’t have to pay sales tax (which is meaningful in Cook County, which has the highest sales tax in the country), and I got an email about two hours later that the filter had already shipped. In fact, I bought two, so I won’t have to think about this again for another year. DONE! And I was happy that my refrigerator was intelligent enough to communicate this information to me.

But think about it … this capability is just the tip of the iceberg (pun intended!). What if, instead of just an orange light on the display, the filter depletion triggered one of two things, depending on how I set it up?

1. An email to me with links to the correct part, so I don’t have to dredge out the owner’s manual, and links to filter vendors.

2. Automatic filter ordering from the online filter vendor of my choice; the refrigerator could arrive pre-programmed for a specific Kitchenaid-approved vendor, but through a remote web interface with my refrigerator using my house’s wires or wireless I can choose a different one if I prefer, and I can choose to order multiples at the same time, so my refrigerator interface can keep track of my filter inventory.

These are the transactions of the future. This is the beginnings of embedded intelligence in appliances, and the individual consumer’s autonomous control that such intelligence enables. It enables autonomous control to lead to overall decentralized coordination in the electric power system. Think about it …

Notice that in this future I can choose how much human involvement I want to have in the filter replacement transaction; I can either receive the information and do it manually, or I can program the appliance, “set it and forget it”, and a filter will arrive without my having to go out of my way to deal with it.

All that’s required to make these transactions and this functionality possible is ethernet capabilities in the appliance, and some programming and thoughtful customer service-oriented interface design on the part of the appliance vendor. More simply put, the fridge just needs an IP address and a spiffy, user-friendly web interface through which I can see it on my house’s network and adjust its settings to suit my preferences (and changes in my preferences over time). Adding the technology to add that functionality would not increase the cost of the appliance substantially, and they could start by doing it in the top-end of the product line and then proliferating it through the broader mass-market products.

Now take the next step: imagine and dream well beyond the autonomous filter replacement transaction. Once the refrigerator has an IP address and is on my home network, I can control it remotely with the right network access and user interface. Now dream big … I can program it to cycle its cooling in response to changes in electricity prices.

Remember that in aggregate in the residential economy, refrigerators consume more electricity than space heating, water heating, or lighting; how big a share of electricity consumption your refrigerator accounts for depends on its features (especially age) and the other uses in your home, but in any case, the fridge is a pretty serious chunk of your typical house’s consumption. That means it’s a substantial share of your electricity bill, and that if you can reduce its power use without harming your food, you can be just as well off but save yourself money. Due to the physics of thermal mass, you can cycle off refrigeration for a while without damaging food, especially if the doors stay closed.

Imagine a hot August afternoon, and you have an IP-enabled refrigerator on your home network. Moreover (and this is the crucial part), you have a contract with a retail energy service provider (either your utility or a competing retailer) under which you are charged real-time prices for the electricity commodity portion of your service. This contract also means that your home has a digital meter that can communicate this price signal into your home and its devices and appliances; the meter is the communication gateway.

This arrangement opens up a wide range of possibilities for you. You can program your refrigerator’s compressor to turn down or off if the electricity price goes above a certain set of trigger prices; notice that this functionality is continuous, it’s not just an on-off thing. Say you’re a working mom, and you want to know when your kids get home from school; you can program the refrigerator to send you an email or an SMS if the door opens during the hours you are at work (because what’s the first thing kids do when they get home? Eat!). If you are particularly concerned about your budget, you can set an electricity budget for your refrigerator, and have it modify its settings and behavior to manage itself to your budget target. Or suppose that you care deeply that the power you consume is only generated from renewable energy; then you can set your fridge to cycle off if all renewable sources on your network are maxed out and your next kilowatt will come from a fossil fuel generator. And if you are having an important dinner party and have delicate, expensive food in the fridge, you can override all of this stuff.

And you can do it autonomously. Or you can adjust it through a remote access web interface.

Now imagine that quite a few of us have this setup in our homes. What is the aggregate outcome? Individually, we each use the technology and the retail price contract to adapt our behavior to suit our personal, individual best outcomes. We use the technology and the pricing to save money while still getting the functions we want from the appliance, and from our electric service. In aggregate, these autonomous choices change demand patterns in ways that reduce demand in expensive periods, which will induce wholesale electricity market prices to do a better job of reflecting both consumer preferences and the actual costs of supplying power at a given time. Those changing demand patterns, and the ability to change behavior autonomously, can also lead to reduced energy use, particularly if it means less use of less energy-efficient (higher heat rate) generators to meet peak demand.

Imagine how technology and retail choice enable us to enjoy that kind of autonomy and achieve these benefits through decentralized coordination. I look forward to that future enthusiastically.

The consequences of falling oil prices, and the benefits of high oil prices

Lynne Kiesling

At Econbrowser, Jim Hamilton has a very thorough post on the economic consequences of falling oil prices. Consumer confidence, inflation, GDP, Fed policy, … and his conclusion is not very optimistic in the short run.

In fact, here’s a question for you: will the short-run and long-run adaptations of consumers and producers actually be beneficial enough to make high oil prices net beneficial? And is it possible for that adaptation to include the political will to streamline energy policy away from targeted subsidies/pork?

The wonderful economist Robert Pindyck takes on some of these questions in a recent interview in which he comments on the energy policy proposals of the two dominant presumptive Presidential nominees.

Pindyck — an expert in microeconomics and industrial organization, the behavior of resource and commodity markets, capital investment decisions, and econometric modeling — recently examined the energy plans put forth by senators Barack Obama and John McCain as posted on their web sites. He was not impressed.

But, as Pindyck acknowledges, being honest about what it would take to wean Americans off oil could be political suicide.

Pindyck’s policy proposal: eliminate alternative energy subsidies and/or implement carbon pricing and a gasoline tax:

Look, what are going to be needed ultimately is a tax on carbon and a tax on gasoline — a large one. Another way to have a tax on carbon is to have a cap-and-trade system so you only allow a certain amount of carbon dioxide to be emitted. That will raise the cost of carbon. A gasoline tax would greatly reduce gasoline use. It would create the incentives we need for other energy sources, including conservation.

The theory behind Pindyck’s proposals is straightforward microeconomics, including its absorption of the Austrian concept of the subjectivity of preferences. Targeted subsidies presume that the party choosing the targets (i.e., the Federal government) knows what the “optimal” alternatives are, which means both that they know the relative costs and benefits across all alternative energy technologies AND that they know the preferences and evaluation of opportunity costs of all of the affected consumers. Similarly, regulatory responses are prey to the same fallacy of centralized control critique.

Only government is arrogant enough to presume such knowledge, or to be indifferent to the costs imposed by their failure to aggregate such knowledge.

Thanks to The Economist’s Free Exchange blog and to Mark Thoma at Economist’s View for the link.

UPDATE: see also posts from Greg Mankiw (touting the “Pigou Club” interpretation of Pindyck’s remarks, naturally) and from John at Environmental Economics on the Pindyck interview.