Firms and Households during the Pandemic: What Do We Learn from Their Electricity Consumption?

AuthorBover, Olympia

    In this paper we investigate how electricity consumption patterns have changed during the COVID-19 pandemic. Since electricity consumption has a strong correlation with economic growth, it has traditionally been used as an indicator of economic activity (Kraft and Kraft, 1978; Henderson et al., 2012; Stern, 2018). However, as we show in this paper, changes in work and life habits triggered by the lockdown measures have implied a structural break in the relationship between electricity consumption and economic activity. In particular, we provide detailed evidence of a strong reduction in the amount of electricity consumed by firms, which was partly offset by an increase in the amount of electricity consumed by households. Therefore, to the extent that economic activity is better captured by firms' electricity consumption, using total electricity consumption would under-estimate the severity of the economic impacts of the pandemic.

    We focus on the Spanish economy, which has been hardly hit by the COVID-19 crisis (IMF, 2020; de Espana, 2020). (1) Nevertheless, the changes in the electricity consumption patterns that we document should also be illustrative of the effects in other countries that have implemented similar lockdown measures, including travel restrictions, social distancing and shutdowns of non-essential businesses, schools and public offices. Indeed, the electricity demand reductions in other developed countries are similar in scale with those reported in this paper. (2)

    Understanding the link between electricity consumption and economic activity has proven to be particularly relevant during the pandemic; for instance, to keep track of the state of the economy in real-time, to assess the trade-offs between health and economic issues when designing the lockdown and deescalation measures, or to assess the need to provide public support to firms and businesses to avoid closures and layoffs. However, there are good reasons to suspect that the link between electricity consumption and economic activity might now be weaker than previously thought. The link between the two had already become less strong before the pandemic--arguably, due to an improvement in energy efficiency and an increased weight of the service sector in the economy (Hirsh and Koomey, 2015; Buera and Kaboski, 2012; Metcalf, 2008)--but the pandemic might have weakened it more by changing the patterns of electricity consumption by firms and households. In the future, the likely increase in remote work (Dingel and Neiman, 2020), the deployment of on-site rooftop solar installations and electric vehicles, or the improvement in energy efficiency (Davis, 2017; Stern, 2018) might further confound the link between metered electricity consumption and economic activity. These trends do not cast doubt on the usefulness of electricity data to measure economic activity, but rather call for revisiting the link between the two.

    In this paper we argue that a key step to do so is to decompose electricity consumption by firms and by households. Whereas the former is mostly unambiguously correlated with economic activity, the latter might be either negatively or positively correlated with it. For example, households' consumption might increase due to unemployment (at least in the short-run) or due to an increase in remote work. However, the data needed to decompose total electricity consumption across consumer types is typically not available close to real time, which is when the use of electricity data as a proxy for economic activity is more valuable relative to other indicators.

    We exploit one institutional feature of the Spanish electricity market in order to decompose total electricity consumption into consumption by firms and by households. In particular, different consumers, depending on their peak consumption and voltage, face different choices of types of electricity tariffs. For instance, only households (and SMEs) have the right to be supplied at last resort rates, while only firms have the right to access the wholesale electricity market directly. In this study, we use publicly available information provided by the Spanish System Operator on hourly electricity consumed under the various tariffs, allowing us to estimate the consumption by firms and households. (3) To check the validity of our proposed decomposition, we compute the correlation of our estimated series with other data sources that should correlate positively with firms' and households' actual electricity consumption data. In particular, we show that our series for households' consumption data has a strong correlation with TV News audience and with Google's Residential Community Mobility Index, which measures the time people spend at home (Google, 2020). Moreover, we show that quarterly GDP year on year growth rates (INE, 2020) are strongly correlated with our series of firms' consumption (and not so with households' consumption), particularly so during the pandemic.

    This study measures the effects of the pandemic on electricity consumption patterns. For that purpose, our empirical analysis captures the departure of (daily or hourly) electricity consumption from what one would predict using previous years' data, while controlling for temperature and seasonality. The daily analysis allows highlighting the diverging trends of firms' and households' electricity consumption during the pandemic. In turn, the hourly analysis allows uncovering changes in electricity consumption patterns of firms and households across the day and across the week, depending on the stringency of the lockdown measures.

    We find that total electricity consumption fell substantially during the first wave of the pandemic, reaching declines of 18.2% under the total lockdown. Yet, the reduction in firms' demand was much stronger, 29.1% below its normal levels, which was partly offset by the increase in households' electricity demand, 9.0% above its normal levels. During the second wave (from mid August 2020 until the end of the year), the reductions in electricity demand have been milder, which is explained by the less stringent lockdown measures in place. Yet, the 2% reduction in total electricity consumption masks a 4.0% decline in firms' electricity consumption, given that households' electricity consumption was still 2.1% above its normal level. It is unclear whether this is due to a slower rate of economic activity and/or due to an increase in remote work. Nevertheless, this asymmetry is reflective of the change in electricity consumption patterns during the pandemic.

    In the Spanish case, the evidence shows a strong correlation between the growth rate of firms' electricity consumption and quarterly GDP. However, in general it is not possible to perfectly map the change in electricity consumption by firms with the decline in economic activity. The reason is that some of the economic activity that used to take place at the workplace has now shifted to the household. Hence, our proposed demand decomposition can be seen as providing bounds to the fall in economic activity: the decline in firms' demand provides an upper bound (as if no activities had shifted to the households) and the decline in total electricity consumption provides a lower bound (as if all the increase in households' demand were due to remote work). Nevertheless, it is important to point out that not all activities that can move into remote work are equally energy intensive. Furthermore, the amount of energy consumed at the workplace and at home for the same amount of work need not be the same. Hence, it is unlikely to find a one-to-one correspondence between the reduction in firms' consumption and the increase in households' consumption.

    We also provide evidence of substantial changes in the hourly patterns of electricity consumption, which again differ across firms and households. In particular, we observe large declines in electricity consumption by firms during working times, which are paralleled by simultaneous increases in households' electricity consumption. During the first half of the year, we find a morning effect in households' demand patterns, i.e., a decline from 8am to 9am. This morning effect is no longer present when schools re-opened after the summer, even if a vast majority of people were still under remote work. Yet, this is paralleled by a late evening effect, with people reducing their electricity consumption at 11pm, probably because they were going to bed earlier. This indicates that the lockdown measures may have led to longer sleeping times, in line with what is reported for other countries (Lee et al., 2020), with potential positive impacts on health outcomes and labour productivity (Marco Hafner and Stolk, 2016). We also find an increase in electricity consumption at lunch and dinner times in Spain (i.e., from 2pm to 3pm and from 8pm to 9pm), which is consistent with people cooking at home rather than eating out. Likewise, we also find that the pandemic affected people's holidays, as they seem to spend more time at home during the summer months than in previous years, an effect that is not explained by remote work. A similar effect can also be seen on Sundays during the total lockdown period.

    Related Literature The impact of the pandemic on the power sector has attracted the attention of several institutions and scholars worldwide (Benatia, 2020, 2022; Cheshmehzangi, 2020; Steve, 2020; Fabra et al., 2020; Fezzi and Fanghella, 2020; Leach et al., 2020; Ghiani et al., 2020; Ruan et al., 2020). These studies focus on measuring the declines of electricity consumption and the consequences for the performance of electricity markets in various countries. However, they all look at aggregate consumption figures without decomposing demand by firms and households.

    As far as we are aware of, there are only two other papers focusing on the impacts of the pandemic on households' electricity consumption...

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT