---
id: "VOLT-HOME-WP-046"
title: "When wholesale prices and network time-of-use charges disagree"
slug: "when-wholesale-prices-and-network-time-of-use-charges-disagree"
description: "An opposing-charge sensitivity tests how an illustrative network time-of-use signal can dilute a wholesale heating-flexibility signal."
published: "2026-08-30"
cluster: "Heat pumps and thermal storage"
status: "measured"
evidence: "/research-data/home-papers/when-wholesale-prices-and-network-time-of-use-charges-disagree.json"
figure: "/research-media/home-papers/when-wholesale-prices-and-network-time-of-use-charges-disagree.webp"
figure_alt: "Chart for When wholesale prices and network time-of-use charges disagree: wholesale flexibility signal remaining under a 60 EUR/MWh opposing network charge, shown as +0, +30, +60, +90."
source_ids:
  - "energy-buildings-heat-pump"
  - "nordic-cost-comfort"
  - "acer-retail-2025"
  - "ec-retail-flexibility-2026"
  - "dynamic-tariff-viability"
peer_reviewed: false
---

# When wholesale prices and network time-of-use charges disagree

## Abstract

A heat-pump controller that follows wholesale prices alone can choose the wrong interval when network time-of-use charges point in the opposite direction. This working paper applies four illustrative opposing charges—0, 30, 60, and 90 EUR/MWh—to the observed mean daily day-ahead price range in a frozen 2,367-observation temperature-price panel. The resulting remaining wholesale signals are 350.516569, 320.516569, 290.516569, and 260.516569 EUR/MWh-equivalent. The registered primary result is 290.5165694972539 EUR/MWh-equivalent under a 60 EUR/MWh opposing charge. All four tested cases remain positive because the pooled mean price range is unusually large; a reversal does not occur inside this grid. The exercise nevertheless demonstrates the accounting rule: tariff components must be combined before selecting an interval. The charge values are illustrative and are not a quote for any country, network, retailer, or household. No uncertainty interval is reported for this estimand. The reduced-order synthetic scenarios are not building simulations, customer telemetry, savings claims, or retail bills.

## Plain-language answer

Add the signals before scheduling. If the cheap wholesale interval carries a 60 EUR/MWh higher network charge than the alternative, subtract 60 from the wholesale advantage. In this pooled scenario, 350.52 EUR/MWh-equivalent becomes 290.52. At opposing charges of 30 and 90, 320.52 and 260.52 remain.

That does not mean an actual household faces any of those numbers. The scenario is designed to show direction and scale. A real tariff may apply a network charge per kWh, per kW peak, by time block, by season, or through a fixed subscription. The [Energy and Buildings study](https://doi.org/10.1016/j.enbuild.2023.113257) provides assigned context on tariff design and thermal flexibility. [ACER and CEER](https://www.ceer.eu/wp-content/uploads/2025/11/ACER-CEER-2025-Retail-monitoring.pdf) provide assigned context on retail contract choice. The only operational rule supported here is to optimise the complete marginal tariff rather than one component.

## Research question

The registered question asks what happens when wholesale prices and network time-of-use charges disagree. We operationalise disagreement as an illustrative charge differential that opposes the observed wholesale daily range. The primary scenario subtracts 60 EUR/MWh; the sensitivity grid uses 0, 30, 60, and 90 EUR/MWh.

The study does not model a named tariff. It does not compare countries, distribution system operators, or retailers. It also does not simulate a household bill or capacity tariff. Its purpose is to make tariff stacking explicit within the registered reduced-order heat-pump method family and to quantify how much of the wholesale signal remains under each declared opposing charge.

## Data and provenance

The evidence is frozen at publication cutoff 2026-08-30T00:00:00Z. Its declared daily price window is 2021-01-01 through 2026-08-29; its detailed interval window is 2025-10-01 through 2026-08-29; and its long-history window is 2015-01-01 through 2026-08-29. The source-table contract contains `day_ahead_prices`, `zone_temp_weighted`, `zone_load`, `generation_mix`, and `forecasts`.

The operative panel has 2,367 daily price summaries paired with non-null population-weighted zone temperature. For each row, the market input is maximum minus minimum day-ahead price. The arithmetic mean of that range is 350.51656949725395 EUR/MWh. The network charges are not observed data. They are explicit 0/30/60/90 EUR/MWh scenario inputs.

The read-only transaction used a 180-second statement timeout. Provenance is fixed by snapshot SHA-256 `7e97489fc8528c8cc8c38830e05b48d949ce1f67b98575dff26f5d7c321e4c67`, analysis-code SHA-256 `57c57de79cdab2b5b6d6c54c485cb5162598c5ba0b0bfe995da40d75e6c52ba9`, protocol SHA-256 `adb36bf6b447af9f96339249b8becaefc20422499cca1977242866347a97bd4b`, registry SHA-256 `7bcb91d7476d0a69fe9fa75a5c7782f8117e0153f82f9112b7e1d307d3943717`, and source-registry SHA-256 `07949550ac443ff673fda5c0209b99f137544f3ffecf6775f109bb9d09663bd6`. The evidence and figure hashes are `c633d3bc3fe23545c522128204fc27f9c898b90a5a774081bf009ccddfc6d4ec` and `b81e967dc45b0e01d03d0b0ea78e5c6837d773360ba00b59d554514be57742f8`.

The registered [European Commission communication](https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A52026DC0115) concerns retail prices and flexibility. The [dynamic-tariff viability study](https://doi.org/10.1016/j.adapen.2024.100174) provides economic context for household tariff choices and enabling costs. The [Nordic cost–comfort paper](https://doi.org/10.3390/en18215568) supplies assigned context for tariff and comfort trade-offs. No external number is inserted into the scenario.

## Method

Let the pooled mean daily wholesale price range be \(R\). Let \(C\) be an illustrative network time-of-use differential that makes the wholesale-cheap interval more expensive relative to the alternative. The remaining signal is defined as `max(R - C, 0)`. The non-negative floor prevents the displayed “remaining” quantity from becoming negative; a zero would indicate that the opposing charge fully consumed the wholesale envelope.

The primary scenario sets \(C=60\) EUR/MWh. The sensitivity grid sets \(C\) to 0, 30, 60, and 90. Because \(R=350.51656949725395\), none of the tested charges reaches the zero floor. A charge differential larger than the observed mean range would reverse the raw combined preference, but such a value was not included in the frozen grid and no empirical reversal rate was measured.

The preserved assumptions are synthetic heat demand, no customer telemetry, COP sensitivity declared per paper, wholesale energy only, and “illustrative opposing network charges; not a country tariff quote.” No COP is applied, no heating-energy amount is specified, and no tax or supplier component is added. The result is an EUR/MWh-equivalent signal, not euros on a bill.

The regenerated evidence reports no uncertainty interval and names the interval method as “not reported for this estimand.” No stale spread-resampling output is a scenario-specific interval for the primary retained signal. Within-family Holm control applies to inferential claims; this descriptive sensitivity makes no unadjusted significance claim.

## Results

With no opposing charge, the remaining signal is 350.51656949725395 EUR/MWh-equivalent. At +30 EUR/MWh it is 320.51656949725395. At +60 it is 290.51656949725395, reported in the evidence as 290.5165694972539. At +90 it is 260.51656949725395.

The 60 EUR/MWh charge removes 60 EUR/MWh-equivalent from the declared wholesale envelope, leaving about 82.9% of the starting value. That proportion is derived from the frozen values, not observed household retention. The sample size underlying the market input is 2,367 matched zone-days.

The evidence reports no interval for the underlying daily price-range input or the 60 EUR/MWh scenario output. The result does not show a reversal within the tested grid. It does show that omitting a charge differential produces a mechanically overstated signal. The general evidence interpretation says a network time-of-use signal can offset the wholesale signal and reverse the preferred interval; this paper narrows that statement by noting that the registered 0-to-90 grid offsets but does not itself reach reversal at the pooled mean.

## Robustness and placebo checks

The four-charge grid is the main robustness check. It shows the result at zero charge and three increasing opposing values, making the one-for-one subtraction visible. The zero-charge arm reproduces the underlying mean wholesale range and acts as an implementation baseline.

A same-direction charge would increase rather than reduce the combined preference, but that arm is not in the evidence. A charge equal to the mean range would reach zero under the formula, and a larger raw differential would reverse preference before the non-negative display floor. Those are algebraic implications, not measured placebo results.

No bootstrap interval is reported for day-level variability in observed price ranges or variation in actual network tariffs. No country tariff, billing rule, season, or household is resampled. There is no causal test showing that a network charge changed customer behaviour. Within-family multiplicity rules remain, while this paper makes no p-value claim. A complete robustness programme would apply actual tariff calendars to interval-level load and prices; it has not been performed here.

## Limitations

The scenario treats the network signal as a simple EUR/MWh differential. Real network tariffs may have fixed, volumetric, time-block, peak-demand, seasonal, threshold, or subscription components. Capacity tariffs cannot generally be represented by subtracting one energy rate from a daily wholesale range.

The pooled daily maximum-to-minimum range is an opportunity envelope. It does not guarantee that a heat pump can move energy between those intervals. Daily temperature and price summaries cannot reproduce a building-specific thermal state. COP, comfort, device power, rebound, hot-water needs, and thermal losses are not modelled.

The high mean range and the fact that all tested charges remain positive should not be generalised to a local market. A household may face much smaller wholesale differences and therefore reach a reversal under a modest network differential. Conversely, fixed charges may not change the marginal interval choice at all.

The reduced-order synthetic scenarios are not building simulations, customer telemetry, savings claims, or retail bills. The charge grid is not a tariff quote. Taxes, VAT, supplier margin, levies, metering, and fixed terms are absent, so no annual cost or household recommendation can be calculated.

## Practical implication

A controller should construct one interval-level marginal price containing every applicable variable component before it optimises. It should preserve units, local tariff time zones, seasonal calendars, VAT treatment, and capacity logic. Displaying wholesale and network components separately helps users understand why the selected interval changed.

If tariff data are missing or stale, the controller should fail safely rather than assume zero network charge. The assigned ACER/CEER and European Commission sources make retail choice central to flexibility; implementation should therefore treat the household’s actual contract as a required input, not as a footnote to wholesale prices.

## Reproducibility

The public evidence is `/research-data/home-papers/when-wholesale-prices-and-network-time-of-use-charges-disagree.json`, and the matching figure is `/research-media/home-papers/when-wholesale-prices-and-network-time-of-use-charges-disagree.webp`. The JSON contains the declared charge limitation, primary value, sample size, explicit null interval, source tables, coverage windows, assumptions, interpretation, and provenance hashes.

To reproduce the statistic, compute each matched zone-day’s maximum-minus-minimum day-ahead price, then take the arithmetic mean over 2,367 rows. For each charge in 0, 30, 60, and 90 EUR/MWh, compute the maximum of mean range minus charge and zero. The 60 arm is the primary result. Preserve the frozen cutoff and snapshot. Applying an actual tariff or interval-level schedule creates a new analysis.

Licensing and attribution terms are recorded at `/legal/data-licensing` and [Voltcast data licensing and redistribution](https://github.com/ossedk/voltcast/blob/main/docs/voltcast/LICENSING.md). Reusers must retain the statement that the charge is illustrative and not a country tariff quote.

## Disclosure

Analysis and drafting were model-assisted; sources, code, assumptions, and evidence hashes are disclosed. This paper is not peer reviewed. It makes no trading claim, is not trading advice, and does not authorize live control. Volt has no live traders or live capital. The tariff examples are analytical inputs, not offers, quotes, or financial advice.

## References

- Energy and Buildings. “Assessment of the thermal energy flexibility of residential buildings with heat pumps under various electric tariff designs.” https://doi.org/10.1016/j.enbuild.2023.113257
- Energies. “Exploring Cost–Comfort Trade-Off in Implicit Demand Response for Fully Electric Solar-Powered Nordic Households.” https://doi.org/10.3390/en18215568
- ACER and CEER. “Rewarding flexibility: How retail contract choice can help unlock consumer flexibility.” https://www.ceer.eu/wp-content/uploads/2025/11/ACER-CEER-2025-Retail-monitoring.pdf
- European Commission. “Communication from the Commission on the Citizens Energy Package.” https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A52026DC0115
- Advances in Applied Energy. “Assessing the conditions for economic viability of dynamic electricity retail tariffs for households.” https://doi.org/10.1016/j.adapen.2024.100174