Based on the laws of supply and demand, it is not surprising that peak prices for electricity occur at peak periods of demand. But in the power business, there are several dynamics that make peak prices not just a little higher, but 10x higher, and more in certain situations. For starters, the legacy grid cannot store electricity, so the utility must constantly guess how much power will be needed and then must deliver exactly that much to avoid over or under voltage conditions. Utilities are very good at this matching process but they can't and don't always get it right.
The current strategy for matching supply with demand is to build 
a certain percentage of generating capacity using gas turbines that can quickly be turned up or down as needed. These plants are quite expensive given that they are not needed very often. The classic situation is the utility that spends $250 million for a plant that is used ten days a year for a hour or two each day. You can guess that a “peaker plant” such as this is very expensive per kilowatt-hour. When we amortize the plant's expense over 10-100 hours of operation per year, we are producing very expensive electricity.
Because almost all utilities in the US charge flat rates, those extremely expensive kWh's get averaged in with the low-cost baseload kWh's and consumers never experience the true cost of peak power. If that peak power is essential, then it is worth every dollar. However, when it is used to raise idle hot water tanks from 110 to 120 degrees on hot sunny afternoons, then it is a bad deal for consumers. If enough demand on those few superpeak days can be shifted to a few hours later in the day, then the entire cost of a $250M power plant can be avoided. Everybody wins with a more intelligent system.
Peaker plants usually make up part of the utility's “spinning reserve.” Just as banks are required to have a certain percentage of cash on hand to cover unexpectedly high withdrawals, utilities are required by law to have a certain percentage of spinning reserve capacity. The plant has to be “spinning” so the generators can be switched on instantly to provide power in case of unexpectedly high demand. Spinning reserve is even worse than an idle plant because it is burning fuel and emitting CO2. Think of having your car on the driveway with its engine running for several hours per day just in case you need a quick get-away. Luckily cars start up fast enough that we don't have to go through what utilities go through every afternoon.
Large capital investments with low utilization rarely make economic sense in a free market. GTM Research (2009) estimates that the capital cost necessary to build 1 MW of demand response capacity is roughly $240,000. This is a bargain compared to the approximately $400,000 (prorated) cost to build 1 MW of a natural gas plant. Analysts have estimated that reducing peak demand can save approximately 40GW of electricity and $3 billion dollars annually.
More intelligent management of peak demand requires both technology improvements to support demand response and regulatory change so that utilities can pass through their actual cost of generation based on time of day. This will allow the market function as it should by letting consumers decide whether they want to buy power during those periods when prices go to superpeak. 
Carbon-Pros Analyst Blog
