Has time varying pricing changed consumer behavior in Spain?

Yesterday we had the last class of the outstanding “Energy and Environmental Markets” (EEM) course taught by Prof. Severin Borenstein. As we were doing a quick review of the section on the different types of electric tariffs (eg. flat rates, time-of-use rates), one of my classmates came up with an excellent question: would time-varying pricing significantly increase the efficiency of the decisions that consumers are making? That is, would consumers that currently pay flat rates behave very differently if they faced volatile hourly varying prices? If they wouldn’t, that is, if demand is very inelastic, then flat rates are not really creating any deadweight loss.

As far as I know, Spain is the biggest electricity market were time-varying pricing has been implemented. So I thought I might share some insights on how the experiment is going so far.

Time-varying pricing has been in place for almost a year now, since the 1st of July 2015. Specifically, the tariff reflects the wholesale day-ahead market price, which clears at 5pm each day, setting a price for each of the 24 hours of the next day. So, even if this pricing does not perfectly reflect short-run marginal cost, it is pretty close.

As we’ve learned in EEM, short-run marginal cost is always the efficient price to set. But in Spain, it wasn’t implemented to increase efficiency… Before switching to this tariff, electricity was priced at a fixed rate that was updated yearly at first, and quarterly later. The rate was set through an auction mechanism that tried to emulate a futures market. The government was concerned that the incumbent companies were exercising market power in this auction and thought that the wholesale spot market was less vulnerable to this type of manipulation. After an auction in which the alarm was especially intense, they decided to go ahead with time-varying pricing.

It’s kind of funny, because the Minister of Energy made a significant effort to create confusion around the volatile nature of the new tariff. If he were trying to incentivize efficient decision making, he should have probably highlighted the variability to encourage people shift loads and save money. Instead, he chose a name that explained very little about the nature of the tariff (“Voluntary price for the small consumer”) and insisted that everybody would be better off no matter what they did. He was probably worried that people were not going to like volatility at all…and I’d say that he was right. Some of the pictures that circulated on the social media under the tag #PVPC, which is the acronym for tariff, illustrate people’s anxiety with unknown volatile prices.

ane1

I would say that PVPC also increased public sensitivity to spikes in the wholesale price, to the point where a price of 90 €/MWh in a peak hour in December 2015 appeared in the front page of most newspapers and triggered a three month investigation of market power. Usual peak prices were around 60-70 €/MWh in that period.

I’ve talked about why we implemented time-varying pricing and the most visible political and public reactions. But the interesting issue is: has the tariff actually changed the behavior of consumers? Obviously, answering this question requires a complex and rigorous analysis that isolates the impact of price from other factors that might have also affected consumer behavior in the last few months. The most important one I can think of is that Spain has experienced a significant recovery from the very long economic crisis that we have been through. Therefore, I cannot answer the question with any accuracy or certainty. But I wanted to share some data on the distance between average and peak hourly demand, with and without time-varying pricing. Hopefully it will get the discussion going.

ane2

A percentage of 131.5% in January means that peak demand of January was 31.5% higher than average hourly demand on that same month. As shown in the table, during the period in which we have had time-varying pricing, peak demand has generally been lower relative to average demand than when we had fixed rates. However, the difference doesn’t seem very significant when compared to the variability of the metric across different years.