By Isaac Dinner – Kenan-Flagler Business School, University of North Carolina
Importance of the Problem
Schneider and Sunstein’s paper tackles a central challenge of the extremely important – and very idiosyncratic – electricity market. Unlike most commodities, electricity cannot be easily stored, the cost of production increases convexly, and consumption is extremely inelastic. In addition, this is an extremely large marketplace where residential consumers account for less than a third of the market, while commercial and industrial applications are nearly as large. Unfortunately, these characteristics imply that short bursts in demand will lead to extremely high costs for the consumer, as described by the authors. In fact, the data collected by the authors in Austin, TX is worth restating: “just under half of the total energy costs were incurred during 20% of total hours, and just 2% of hours were responsible for over 20% of total energy costs.” This problem was also reiterated in a recent WSJ article which states that “California’s solar farms create so much power during daylight hours that they often drive real-time wholesale prices in the state to zero. Meanwhile, the need for electricity can spike after sunset, sometimes sending real-time prices as high as $1,000 a megawatt-hour.” Thus, any method that smooths energy consumption will lead to substantial consumer savings and place far less stress on the entire system.
A straightforward solution to curbing consumption is to institute time-varying pricing (TVP) so that only consumers with a high willingness to pay for electricity will remain active during peak periods. This pricing strategy has been used in many commercial and industrial markets but is still uncommon in residential markets. As noted by the US Energy Information Administration (EIA), only 3% of households use a form of time-varying pricing, either due to a lack of available options or due to a choice to remain with a fixed price. A lay argument would suggest that TVP should be used by residential consumers for the same reason that it is effective in containing peaks within commercial and industrial markets. However, Schneider and Sunstein’s paper analytically demonstrates that behavioral biases and transaction costs imply that a clear strategy like TVP may not be the most efficient in the existence of this type of friction. Specifically, due to behavioral biases in consumption, residential consumers may not optimally allocate energy consumption according to market prices and, therefore, end up paying more under TVP. For example, on a particularly cold (or hot) day, a residential consumer may not be able to shift their hours at home to take advantage of temperature differences. In contrast, commercial and industrial sectors are less susceptible to these biases and can take a more economical route by simply shifting hours of production.
Tempering peak electricity demand is a much more important issue than decreasing average electricity consumption. Further, this problem is only growing bigger. Therefore, the goal of pricing and behavioral nudges is not to use price discrimination methods such as TVP to decrease total consumption, but to shift consumption to other time periods and consequently decrease peak consumption. However, a challenge with shifting consumption is that these behavioral biases and transaction costs are non-trivial.
Interestingly, Schneider and Sunstein’s results apply beyond the electricity market. In fact, their findings provide insights into other markets where demand is rather inelastic or where storage is difficult, such as in the ride-sharing, or oil and gas markets.
The ride-sharing market also follows a similar structure in which consumer prices are low unless there is a surge in consumer demand well above the supply of drivers (e.g. Uber surge pricing). This research suggests that a TVP mechanism may lead to increased consumer costs but not solve the real problem of smoothing consumption.
The oil and gas market is another example of a market where storage can be very expensive and demand is rather inelastic. For example, in 2016 a leaking pipeline in Alabama caused gas shortages and price spikes throughout the southeast United States However, locations closest to reserves on the coastline were less impacted by price swings.
Of Schneider and Sunstein’s proposed solutions to this problem, the suggestion of automated devices to manage consumption truly stands out. This is a noteworthy solution given the increasing growth of personal power supplies. For example, a product like Tesla’s Powerpack allows a consumer to not only have a residential battery backup, but also to smooth electricity consumption throughout the day by charging a battery when the market price is low and discharging when the price is high. This type of solution is precisely in line with Schneider and Sunstein’s comments because it is a way to circumvent biases related to consumption and attention by making the product consumable at alternative time periods. However, a product like this faces another challenge in that it has an extremely high upfront cost.
Read the full article by Ian Schneider and Cast R Sunstein “Behavioral considerations for effective time-varying electricity prices” in the second issue of Behavioural Public Policy for free here.