ICIS POWER FORESIGHT: EU achieves a 6.6% cut in Winter 22/23 power demand

In this free-to-read update, ICIS analysts Luca Urbanucci and Robbie Jackson-Stroud review demand over Winter-22 and utilise the soon to be launched Power Foresight model to provide an outlook for the rest of the year.

If you are interested in a trial of the new Power Foresight model, contact our Lead Analyst Matthew Jones (matthew.jones@icis.com).

EU countries achieved a 6.6% cut in power demand over the winter, short of the non-binding 10% cut that was set last year, our analysis of TSO data shows. The EU also fell short of the binding 5% cut for peak demand, achieving a 3.6% decline over the period, though seven member states met this target.

In this analyst update we break down the demand data across the EU and GB markets, quantifying the role of temperatures in the demand cut, analysing the reasons for significant differences across countries, and assessing the hourly patterns for demand reduction. We also look forward to what to expect from European demand in the coming months as prices continue to fall.

Country results

• As shown in the chart below, EU countries did not manage to achieve the 10% reduction target set by the commission this winter, with an average 6.6% reduction in power consumption.
• EU-27 demand averaged 327GW over this winter, 24GW lower than the 5 years average.

[Sources: ENTSO-E]

• Nationally, no country was able to reduce demand enough to meet the EU voluntary targets.
• Great Britain was not included in the EU-27 targets but is shown to have reduced demand nearly 9% vs the 5-year average.
• Of the higher consumption European countries, Poland and Switzerland appear to have not reduced power demand at all in winter-22.
• November and January showed the highest reduction in demand on average across countries; however, November also had the largest variance in demand reduction.

• Countries were also only observed to reduce the peak demand hours by 3.6%, short of the 5% EU target.
  • Peak demand refers to the EU target definition, the greatest 10% of hours over winter based on demand.
• Whilst collectively the target wasn’t met, our analysis shows that seven EU countries did meet the 5% peak demand reduction target. GB also met this target.

Adjusting demand data: the embedded generation in the Netherlands

• ENTSO-E load data, which has been leveraged to assess the demand changes of EU countries, includes embedded generation. That means that actual consumption demand is generally higher, as part of it is met locally (i.e., with behind-the-meter residential-scale solar panels).
• Yet, the impact of embedded generation is negligible for almost all countries (lower than 3% of the annual demand), but the Netherlands are a notable exception as solar embedded generation represented around 17% of the demand in 2022.
• Consequently, Dutch data had to be adjusted to remove the impact of additional solar generation from the demand reduction.
  • Around 3.6 TWh of embedded generation was recorded during last winter, compared to an average of about 1.3 TWh in the 5 previous winters.
• As a result, Dutch adjusted-demand reduction against 5-previous-year average is 6.0%, while a 9.5% figure would have been assessed without removing the impact of embedded generation.
• Indeed, the Netherlands have witnessed a significant increase in solar capacity over the past years (from 2.1 GW in 2017 to 14.3 GW in 2022), with the national daily demand profile increasingly shaping as a ‘duck curve’.

 Analysis of results

• January and November stood out as months showing the greatest demand reduction across countries. There are some considerations for this:
  • November was one of the warmest on record for many parts of Europe. This will have aided the lowering of electric heating demand.
  • Similarly, Europe has experienced prolonged cold spells in January several times in recent years. This will play a part in raising the 5 yr average and making a reduction in W-22 more obtainable.
  • This influence is noticeable for countries like France, where heating demand is largely electrified.

• Price point is another major consideration, as it is very effective way of reducing demand, as covered in our story last year.
  • Tracking the forward curve, prices were highest in November, so this would be another reason for the month having the highest demand reduction, as industry avoids high running costs. Spikes in March and August in 2022 forced energy intensive industry to scale back, such as aluminium production.
• The correlation between price and winter demand reduction does not seem linear however, suggesting a longer-term structural change in demand.
  • This will be influenced by the lag between realise costs for consumers and wholesale prices. Many UK and European fixed tariffs will have been hedged using seasonal and annual contracts, meaning fixed costs for the winter.
• The table below shows the reduction in average demand over each hour of the day. The reduction in demand is mostly evenly distributed across the day; however, there are clear differences between countries.
• As covered above, Dutch demand data has been heavily influenced by the growing solar capacity in recent years, however for countries with a lower recent buildout in solar capacity the profile is not affected.

• Germany has shown a consistent net reduction this winter, with little to no change in the daily demand profile. This likely indicates an overall reduction in industrial and economic activity, rather than being a behavioural or sectoral specific reduction.

• Italy is an example of shifting the demand profile rather than purely absolute demand reduction. The hourly profile has reduced most during the evening ‘super peak’, showing a shift in when power is being used. This could be related to behavioural changes with businesses and individuals changing time of use of appliances or work hours.

The impact of temperature

• Last winter in Europe was characterized by mild weather, with temperatures above normal in most countries.
• To assess the impact of temperature on heating demand, heating degree days (HDDs) over the November-March period were calculated for the main European countries and compared against the 5-previous-year average.
  • HHD is an indicator of heating demand over a certain period. A base temperature of 15.5 °C was considered.
• In most countries, HDDs last winter were below historical averages, thus clearly highlighting how the heating demand was reduced compared to a typical winter.
  • On average, the HDDs were 6% lower than the average of the past 5 years.
  • As mentioned, November was especially warm (HDDs 18% lower than normal), followed by March (-12%) and January (-7%); on the other hand, December and February were in line with historical average (HDDs 3% and 2% higher than normal, respectively).
• As discussed in the previous update, the impact of weather on power demand may vary significantly across the different countries, depending on the share of heat pumps and resistive electric heating on the overall demand.
  • France and Sweden use up to about 25% of their annual power consumption for space heating, while, at the other end, in Germany, Italy, Denmark, and Austria, heating accounts for around 10% of the power demand.

[Sources: ENTSO-E, ECMWF]

• Nonetheless, weather-adjusted demand data by national TSOs show that mild temperatures played a secondary, albeit not irrelevant, role in last winter’s drop in demand.
  • In France, around one quarter of the seasonal demand saving was due to weather and calendar effects, while the majority of the reduction was driven by other factors.
  • Likewise, in Italy and Spain the above-normal temperatures contributed only 8% and 16% of the demand decline, respectively.
• Although an exact figure is not available, it is possible to assess the impact of mild weather on the overall demand reduction across Europe.
  • If we multiply the average share of space heating in the total demand (between 10 and 20%) by the HDD reduction (6%), we get an estimate of the demand reduction driven by weather (0.6-1.2%).
  • As a result, the mild weather should have accounted on average for between 10 and 20% of the overall demand reduction.

[Sources: RTE, Terna, Red Eléctrica] *year-on-year for Italy and Spain, vs 2014-2019 average for France

The impact of policy

• As a reminder, the EU targets were purely voluntary and non-binding, so the motivation of individual countries would arguably be no greater than without the policy.
• Demand will have been heavily influenced by Europe’s attitude towards spiralling prices and schemes used to either protect customers or curtail high end users. Protection measures would likely be a reason for demand targets not being reached, as it lessens the need for consumers to reduce or change their power usage in order to save money on bills.
  • Spain’s Iberian price cap is an example of this, where the cap on gas for power generation lead to an increase in gas usage as well as an increase in power exports to neighbouring countries.
• In our piece last year on Spanish demand change, we referenced a chart by Eurostat displaying changing electricity prices for consumer households as wholesale prices rose during H1 2022.
• In effect what this tell us is which countries consumers are most directly impacted by the recent surge in energy prices.
  • Looking at the table of peak hourly reduction, many of the countries reporting most demand reductions also had high exposure to rising wholesale market prices. Estonia, Greece, and Latvia are three examples of places where consumers have been hit the worst by rising power prices, indicating that behavioural changes by domestic users partially drove the demand reductions observed.
  • Italy is a country that has had high exposure for households and medium sized consumers, likely leading to the behavioural change that has impacted the evening peak demand this winter.
  • France saw government tax cuts and used at the time part owned utility EDF to sell power below the market rate. This protection likely suggests factors other than price also drove the demand reductions.
  • Poland saw little to no decrease in demand versus the last 5 years. In September last year, the Polish government froze energy bills for households, as well as capping prices for small and medium enterprises. This will have taken away a large incentive for domestic and commercial demand to reduce activity in W-22.

• The extent of industrial protection measures also varied between countries, but most provided some, reducing the impact of price driven demand destruction.
  • Germany introduced the Economic Defence Shield, which covered people and businesses with a package of 200 billion euros.
  • Norway introduced a scheme to help industry that spent more than 3% of their costs on electricity.
• Consumer awareness of energy usage was high last winter, not just due to rising costs but news and advertising campaigns. Great Britain is continuing to run several tv pledges to lower electricity use.

What to expect next?

• Looking forward, the decline in electricity prices recorded in the last few months across all Europe, mainly driven by bearish gas prices, is expected to correspond to a rebound in demand.
  • According to the newly launched ICIS Power Foresight model, French prices are forecasted to be 51%, 73% and 34% lower in Q2, Q3 and Q4 2023 compared to previous year, respectively.
  • Likewise, electricity power prices are expected to drop by 41%, 69% and 29% in the next Q2, Q3 and Q4 in Germany, compared to the same periods of 2022.
  • Yet, power prices are still forecasted to be more than twice as high as in 2017-2021 between now and the end of the year.

• If, on one hand, the price decline should support demand recovery, the delay between wholesale prices and consumer bills must be kept in mind, meaning that the effects of price decrease will impact most consumers with a time lag of a few months.
  • Moreover, the end or reduction of support schemes for households and businesses may partially offset the effect of price decline.
• Going forward, less news coverage and fewer awareness campaigns on energy costs are expected compared to last winter, which could result in partial restoration of old behaviours by residential consumers.
• As for the industrial demand, a specific analyst update will soon be published on the topic, focusing on the impact of prices on the different demand sectors.
  • However, the main question is whether industrial demand will show a rebound, as happened in 2021 after the pandemic crisis, or whether and to what extent in some subsectors the decline is becoming structural.

Conclusions

• As anticipated by our prior analysis, countries were not able to meet the voluntary demand reduction targets. Yet, the EU averaged a 6.6% decline in power consumption during the November-March period (below the 10% target), and peak demand was reduced by 3.6% (against a 5% target).
  • With gas and power prices much reduced this year, it is also unlikely a similar measure this winter will see equal or greater reductions, should such policies be put in place.
• Whilst last winter, particularly Q4-22, was characterised by anomalously warm weather, the impact of this on demand was limited. Only a 10-20% contribution to the consumption decreases is estimated, with heating degrees 6% lower than historical averages.
• Winter saw only small demand profile changes for most countries, but a generally even distribution of reductions points to economic downturn or flatter profiled industry driving demand reduction.
• If behavioural changes impact sufficient demand, we may see the timings of the evening ‘super peak’ brought forward/lower, as observed for Italy.
• Looking ahead, lower wholesale prices should trigger a recovery in demand.
  • Yet, the end of support schemes (e.g., the UK energy price guarantee) will impact countries commercial and residential sectors, thus partially offsetting the price reduction.
  • A lag between hedged energy tariffs and wholesale prices may mean a delay in demand rebound this year.
  • We are also mindful that wholesale markets are still greatly elevated vs 2021 prices, which will still impact economic demand in general.