Source - IVL Konsumentperspektiv Efterfrågeflexibilitet (2023)

IVL Svenska Miljöinstitutet, “Konsumentperspektiv på efterfrågeflexibilitet” (Consumer perspectives on demand flexibility). Commissioned by Energimarknadsinspektionen (Ei). Principal investigator: Magnus Hennlock (magnus.hennlock@ivl.se). Published 2023. ~167 pages + appendices.

Survey fielded January 2023, immediately after the highest-ever December electricity prices in Sweden.

Methods

Mixed-method design across four components:

1. Qualitative — market actors: 14-actor reference group; deep interviews and workshop
2. Qualitative — households: Deep interviews and two focus groups (flexible and non-flexible households)
3. Quantitative — market actor survey: Survey to 1,189 contacts at electricity suppliers, DSOs, certified solar installers, and Svensk Solenergi members; 143 responses (12% response rate)
4. Quantitative — household survey + choice experiments: 10,328+ household survey respondents; discrete choice experiments with 2,679 households that did not have hourly pricing (timprisavtal)

Key findings

The 2022 activation event

75.6% of households that had started managing their electricity use began in 2022; 54.4% of those started specifically in autumn 2022. The December 2022 electricity prices were the highest ever recorded for a December month. The report was fielded in January 2023, at peak awareness.

Penetration (January 2023)

• 34.2% of all households reported actively steering electricity use over the day
• Of those, 75.6% used manual behavioral change (shifting when dishwasher/washing machine runs)
• 28.3% used timers
• 23.6% used smart control via app/device
• Of all households: 25.8% manually shift appliances; 9.7% use timers; 8.0% have any form of smart control
• Conclusion: only 1 in 5 of those who steer have any automatic control — manual behavior dominates

Why households become flexible

Ranked motivations for those who steer:

1. Reduce electricity trading costs via hourly pricing (timprisavtal) — #1 for both lägenhet and villa households
2. Reduce grid costs (effekttariff/tidstariff)
3. Contribute to grid relief
4. Contribute to environmental/societal benefit
5. Tech interest (last)

Historical shift: before 2022, early adopters were tech enthusiasts with an energy efficiency interest. From 2022, economic incentives became the primary driver as price volatility increased sharply.

Why households don’t become flexible

Primary reasons (non-flexible households):

• #1: Don’t believe savings would be large enough (25% lägenhet, 22% villa)
• Knowledge barriers: complexity of hourly pricing, contract types, technical solutions
• Comfort/convenience conflicts: showering, cooking, family schedules not negotiable
• Fixed contracts: many locked into favourable fixed-price contracts from before the price spike
• Flat-dwellers: often no individual electricity contract (BRF manages collectively)
• Old technology that cannot be remotely controlled

Contract confusion: approximately 9% of non-hourly-pricing households believe they already have hourly pricing — they confuse monthly-average “rörligt pris” with timpris. Supporting evidence: 1 in 5 households citing “reduce hourly pricing costs” as their motivation actually have monthly average or fixed pricing.

Choice experiments — quantitative results

Three discrete choice experiments testing adoption probability for smart control + hourly pricing vs. current contract:

Experiment 1: Smart heating control, n=1,405 villa households with electric heating

Key attribute effects:

• “Prioritize renewable energy”: +12.9 pp adoption (p=0.000)
• Each additional 5% savings: +1 pp (p=0.006)
• Each 1°C lower minimum indoor temperature: −5.4 pp per degree (p=0.031)
• “Prioritize reduced grid load”: no significant effect

Background variable effects:

• EV access via car-sharing service: +38.9 pp (p=0.000)
• Groundwater heat pump: +16.8 pp (p=0.000)
• Each adult child (>18) in household: +12.7 pp (p=0.000)
• Each degree higher winter indoor temperature: +9.3 pp (p=0.000)
• Being married/cohabiting: −16.7 pp (p=0.000)
• Each year of age: −0.3 pp (p=0.000; equivalent to −6 pp at age 50 vs. 30)
• Outside the three largest cities: −0.67 pp (p=0.015)
• Gender, income, education: not significant

Experiment 2: Smart EV charging control, n=572 households with EV/PHEV

Interpretation: EV-owning households are more experienced with electricity and energy costs; uncertainty has little effect. 14.3% already had automatic timer-based charging, 9.7% with load balancing, 13.1% already steering manually to hourly price.

Higher income increases adoption (+30% per 10,000 SEK/month in certain-savings group, only +9% under uncertainty). Men more likely than women (+5.7%, p=0.042).

Experiment 3: Behavioral change — shower/appliance timing, n=737 households

Key findings:

• “Prioritize renewable energy”: +2.3 pp (p=0.037) — significant, unlike experiments 1 and 2
• “Reduce grid load”: +1.6 pp (p=0.089) — significant
• Moving showers after 22:00: −3.3 to −4.1 pp (p=0.012)
• Moving washing/dishwasher: not significant (savings too small)

Women more likely than men (+4.0%, p=0.057). Married/cohabiting under uncertainty: −10.9 pp (p=0.007).

The moral attributes finding: environmental/grid-benefit arguments have no significant effect on smart-control decisions but have small, statistically significant effects on behavioral-change decisions. Explanation: manual action feels more directly connected to making a real difference than being a node in an automated system. A DSO that appeals to customers to reduce load when the grid is near capacity can get more immediate behavioral response than persuading them to install smart controls.

The uncertainty finding: the dominant experimental result. Uncertain savings reduced heating-control adoption by 23.7 pp — nearly 1 in 3 potential adopters was lost. This was much larger than for behavioral change (−9.7 pp) and negligible for EV charging (−1.6 pp). The pattern is explained by EV households’ greater familiarity and by behavioral change having no upfront investment to justify.

Married/cohabiting effect: under uncertainty, married/cohabiting households are consistently less likely to switch — for both heating and behavioral experiments. Interpretation: respondents making the decision alone (in the survey) cannot anchor with their partner, inducing extra caution. Implication: information campaigns should reach both partners in a household.

Market actor survey findings

63% of surveyed market actors offer flexibility-enabling products or services (as of survey date, ~late 2022).

Demand trend:

• 76% had seen increased consumer inquiries in the past year
• 65% had seen increased sales in the past year

Products/services offered (among those with flex offerings):

• 59%: app showing hourly consumption and/or cost
• 55%: advisory services on steering
• Direct load control and aggregation: minority

Contract landscape:

• 92% of responding electricity suppliers offer hourly pricing (timprisavtal)
• Only 30% of DSOs offer optional time-differentiated tariffs (valbar tidsdifferentierad tariff)
• 38% of DSOs planning to introduce mandatory demand charge tariffs (obligatorisk effekttariff)

Primary marketing argument used by market actors:

• Cost savings: 59%
• Environmental/societal benefit: 16%
• Sustainable electricity system: 11%

Consumer knowledge barriers cited as obstacles to sales:

• Lack of knowledge about societal/environmental benefit: 67%
• Lack of knowledge about the electricity market: 56%
• Lack of technical knowledge about smart control: 50%

Internal company barriers:

• Component shortages/long lead times: 40%
• Resource/time constraints: 38%
• Technical challenges: 37%

Regulatory and data infrastructure asks

From open-ended responses in the market actor survey:

• Spot price data: uncertainty about whether Nord Pool spot prices are freely available for service development on the Nordic market — needs clarification
• DSO grid tariff database: aggregators cannot optimize without machine-readable, accessible tariff parameters for all DSOs; called for explicitly
• BSP/BRP role split (occurring in 2024): seen as positive but delayed; before the split, the ~10–15 BRPs act as gatekeepers and “bromskloss” (drag) on aggregation
• Aggregator role: still undefined in Swedish law at time of survey (early 2023)
• Mandatory hourly metering: schablonmätning should be phased out to enable market transparency and proper incentives
• DSO active role: current rules prevent DSOs from offering services that enable efficient grid use; calls to allow DSOs to serve end customers more actively
• Central data hub: standardized meter data in a central hub would allow aggregators to optimize across DSO boundaries (“standardisera mätdata i centralhubb”)
• Green tech deductions (Grön Teknik): improve practical applicability; extend to storage
• Microresource category: open up a new resource category for plug-in-type assets that lets consumers participate without large fixed installations

Proposed KPIs (20, in 6 categories)

Contribution

1. Equilibrium-model estimate of maximum shiftable load (with price elasticity)
2. Peak/off-peak consumption ratio, normalized for seasonality (trend metric)

Incentives 3. Relative price of peak vs. off-peak hours (hourly spread ratio) × average price elasticity 4. Standardized savings estimate per consumer segment with typical flex solutions (requires API/hub access)

Barriers 5. Share of households perceiving uncertainty in economic incentives (from recurring survey) 6. Share rejecting flex solutions with hourly pricing in repeated choice experiment 7. Share lacking knowledge about demand flexibility (from recurring survey) 8. Average investment cost for smart control products on market 9. Share experiencing technical barriers (from recurring survey)

Market actor supply 10. Number of market actors offering hourly pricing and/or effective time-differentiated tariffs 11. Share of total electricity price per kWh that is variable (marginal incentive proxy) 12. Number of market actors offering energy optimization services (monitoring + control) 13. Number of market actors offering energy storage solutions

Market actor incentives 14. Share of market actors actively marketing flex products 15. Number of consumers using flex-enabling products/services 16. Share of market actor revenues from flex products 17. Number of market actors offering flex products (market breadth)

Market actor barriers 18. Share of aging household technology (older heat pumps, EVs) limiting third-party controllability 19. Size of hard-to-reach consumer groups (pensioners, parents on leave, non-native speakers, etc.) 20. Average time from order to operational installation of flex solutions

Relevance to wiki topics

• Demand Response — comprehensive empirical baseline for Swedish household demand flexibility: penetration, motivations, barriers, choice experiment results, behavioral vs. automated flexibility
• Aggregation — BSP/BRP barrier confirmed by market actors; aggregator role undefined at time of survey; data access as structural barrier; central hub as enabling infrastructure
• Flexibility Market — market actor supply profile, tariff landscape, consumer knowledge gaps, marketing strategies, regulatory asks
• Elmarknadshubb — market actors explicitly call for “standardisera mätdata i centralhubb” as enabling infrastructure for aggregation optimization
• Network Code on Demand Response — household-level moral attributes (renewable priority, grid load reduction) showed measurable adoption effects in behavioral experiments; relevant to NC DR’s demand for consumer engagement
• Distribution System Operator — tariff landscape (only 30% of DSOs have time-differentiated optional tariffs); 38% planning mandatory demand charge; calls to allow DSOs to serve end customers

Data gaps

• Post-2023 follow-up: has Ei commissioned a repeat of this study with 2024–2025 data? Penetration likely increased materially after NC DR transposition work and Tibber/smart-charging market growth
• Segmentation by electricity price zone (SE1–SE4): Southern Sweden had dramatically higher prices in 2022; northern households had weaker economic incentive — this geographic dimension is not broken out in the published report
• EV charging experiment sample (n=572) is relatively small; a larger follow-up study focused specifically on EV/V2G households would be valuable given the rapid EV penetration growth post-2023