Lessons learned from the electricity tracker

This blog is partly based on an article published in the COVID edition of the IAEE Energy Forum.

Our tool, updated daily, tracks electricity consumption in 2020 as a percentage of the corresponding day in 2019, in most European countries and a select few outside Europe. We are now launching a new interactive feature allowing for more user-friendly visualisation of consumption by country or region (Figure 2 on the tracker page).

High-frequency and internationally comparable indicators of economic activity are helpful to track developments in a fast-moving, international crisis such as that triggered by the COVID-19 pandemic. Changes in daily electricity consumption are a strong proxy for industrial and/or economic activity. Data on industrial activity for March and April strongly correlates with observed changes in our measure for electricity consumption (Figure 1). Regression analysis in 44 countries found a positive but variable relationship between our electricity consumption data and a measure of social distancing.

Figure 1: Average monthly electricity consumption vs industrial production (January to April 2020)

Caption: The figure represents average monthly electricity consumption relative to 2019 levels against the European Commission industrial production index. This is plotted from January to April 2020. The colours represent the different countries, whilst the symbols show the different months plotted.

Shifts in electricity consumption are strongly affected by the lockdowns put in place to respond to the COVID-19 pandemic. However, there are some caveats: many other factors are at play, making it challenging to isolate the causal effect of COVID-19 on electricity demand. To reduce the room for error we have used historical data to correct for the daily temperature variable. As national holidays also reduce demand, we have removed them from the sample. Our tracker has nonetheless often shown depressed electricity consumption in the days immediately following such holidays. For example, in Japan some of the biggest drops in electricity consumption appeared immediately after the ‘Golden Week’ holidays.

Figure 2: Japanese consumption: screenshot from electricity tracker page (figure 2)

Such noise in the data is more problematic when interpreting daily ratios. Weekly average ratios tend to be more robust. Nonetheless, interesting observations can be extracted by using both.

What have we learned?

Electricity consumption immediately dropped when governments effectively shut down economies. A transition through a period of depressed electricity demand followed in most countries. Over the past few weeks, as lockdowns have been cautiously lifted, consumption has begun to gradually pick up. But it has not yet returned to 2019 levels in most countries, indicating that their economies are not yet functioning at full capacity.

Speed of lockdown measures coming into effect

In March, electricity consumption showed clear differences in policy approaches to lockdown between countries that gradually imposed more stringent measures (European countries including Spain and Poland, and US states such as New York) and those who imposed restrictive measures literally overnight, such as India and Russia (Figure 3).

Severity and length of lockdown

Future revisions to GDP growth will likely be linked to the length and extent of depressed electricity consumption. Figure 4 shows the average change in electricity consumption for working days between 16 March and 12 June. Within Europe, countries including Italy, Slovenia and Spain have all been hard hit by the crisis while others, particularly in the Scandinavian region, have suffered small if any reductions. Outside Europe, India has seen a notably large drop.

Timing of recovery

Electricity consumption is beginning to bounce back as countries move out of lockdown.

The key question now is whether recoveries will be V-shaped or U-shaped: will economic activity quickly return to its pre-COVID 19 levels (perhaps even overshooting last years’ values to compensate for lost output)? Or will aggregate demand remain depressed for the foreseeable future? Additionally, to what extent will damage to the supply side prevent production levels from returning to their pre-crisis levels? Our tracker’s daily updates will use electricity consumption as a proxy to answer these questions by monitoring how long consumption remains depressed in individual countries.

The pace of recovery so far can be gauged by comparing the reductions in electricity demand during the strictest lockdown periods (the first 10 weeks) with those during the past three weeks that saw an easing of lockdowns (Figure 5).

In France and Spain, economic activity is clearly beginning to pick up as electricity consumption has increased. On the other hand, demand in Austria, Germany and Poland appears to be recovering very slowly. Clearly, the depth of the initial demand drops plays a role. In Australia and Japan there is little difference because demand was not significantly affected in the first place. Aggregate activity in the US appears to be returning to levels comparable with 2019, while the European average remains notably reduced.

Recommended citation

Mcwilliams, B and G. Zachmann (2020) 'Lessons learned from the electricity tracker', Bruegel Blog, 22 June, available at https://bruegel.org/2020/06/lessons-learned-from-the-electricity-tracker/