An analysis of electricity demand data from three sets of customers by Powerconsumer Inc. shows that the imposition of a Provincial Emergency Order in Ontario on March 17 has significantly reduced power consumption by industrial, commercial, and a sample of 200 multi-residential customers.

COMMERCIAL CUSTOMERS AFFECTED MOST

The data for our analysis comes from a group of commercial customers in the Toronto downtown core comprising more than 1.4 million square feet of mixed shopping centre, street-level retail, and high-rise office space. The graphic shows how demand has fallen since March 17 against forecast: 24 percent lower in March, 32 percent lower in April and 30 percent lower in May.

INDUSTRY GOING BACK TO WORK

Our sample of industrial customers includes 10  small and mid-sized manufacturing companies in Newmarket, Ontario, including automotive, food and beverage, electronics, packaging, and advanced manufacturing. Some customers immediately were affected by the emergency order. Other customers have continued to operate.

Overall, electricity demand by the group is lower by 22 percent compared to forecast. We can see in the plot, however, that hourly demand already is picking up towards the end of May as the Government of Ontario eased restrictions on which businesses could operate.

CONFOUNDING RESIDENTIAL DEMAND

Our sample of residential electricity demand data comes from 200 suites within multi-residential apartment buildings in Toronto. All 200 suites are heated with natural gas.

The chart shows a mixed result in demand in response to the emergency order, with higher on-peak and lower off-peak, for these customers. The 24-hour moving average trendlines reveal the days where actual demand varied most from forecast.

In order to control for other factors, to analyze the effect of the emergency in isolation, we performed a series of regression analysis to find models which best fit the data, and to use those models to estimate the effect—in absolute and percentage terms—that the imposition of the emergency order had on demand.

The data is time series, in which the imposition of the emergency order occurred at a discrete moment in time, so we are testing for statistically significant differences in patterns of hourly energy consumption before and after March 17, when the order was made.

In March, April and May, weather tends on average to be milder—in the transition between winter and summer, so other factors like hour of day might be expected to be relatively more significant. In our analyses, we controlled for temperature and a range of hourly periods so as to get at the change related to the emergency order on an hour by hour basis. The results are summarized in the chart below.

The interpretation of these results—the estimated coefficient on the indicator variable for the period under emergency order—is to say that in all hours (reading from the bottom), the effect of the emergency order is that hourly demand is 98 percent of normal, down by 2 percent.

In the 9 to 5 period overall, the emergency order is estimated to have increased demand by 7 percent above normal, 108 percent of normal between 4 and 5 pm. The increase in demand during the 10 to 8 period also is 7 percent.

Outside of these working hours, however, hourly demand is down, 6 percent down outside of 9 to 5, and 18 percent lower between 6 and 7 am.

Overall, and perhaps counter-intuitively, our analysis of these 200 customers reveals that electricity demand is down on average since the emergency order. We analyzed actual hourly electricity consumption to isolate the effect of the emergency order, controlling for weather, and day of week. Our findings confirm a statistically significant reduction in demand under emergency order overall.

A NEW NORMAL

Our finding that electricity demand is significantly lower for commercial and industrial customers comes as no surprise and is in line with numbers published by the IESO.[1]

We did not expect, however, to find that—for our sample of 200 multi-residential customers—residential demand also would be lower overall.

Industrial demand is affected most by production processes and patterns, and those in turn—in a just-in-time supply chain economy—are most vulnerable to work stoppages, whether upstream or, in the case of COVID 19 and Ontario’s automotive industry, as a result of decisions taken by the assemblers downstream to cease production in the face of rapidly declining demand.

Commercial and residential patterns of electricity demand, by comparison, are mostly related to human needs for comfort and convenience: lighting, heating and air conditioning, hot water, cooking, electric appliances, and electronic devices.

The shutdown resulting from the emergency order has caused losses in economic production, as measured by Gross Domestic Production or labour employment, but it has not reduced basic human needs for Ontario’s population.

By showing that Ontarians basic electricity needs are being met with lower consumption, not only at work but also at home, our analysis points to a new perhaps more sustainable normal post-pandemic.

Anticipating new codes and guidelines for social distancing and public health hygiene, many businesses are revising their thinking on what office space they have, how much they need, and how it is to be used. Some already have announced office closures.

Daily commutes downtown are less important, more discretionary, and significantly more expensive than an alternative where people work more at home. Employers will be asking themselves how much redundant office accommodation they need to pay for, not only during a prolonged period of social distancing, but after, whenever that may be.

[1] http://www.ieso.ca/en/Corporate-IESO/Media/COVID-19-Related-Updates