Pratfall effect as a methodology principle
Anyone who honestly names their limits gains more trust than someone who makes perfect promises. That is the Pratfall principle, formulated by social psychologist Elliot Aronson in the 1960s.
In our methodology, Pratfall is a deliberate design decision. Instead of marketing our calculator as a perfect tool, we explicitly document what it cannot do. Six concrete limitations, each with its own explanation. We believe this transparency is more valuable to journalists, consumer advocates and critical users than any marketing statement.
Concretely this means: where we do not know, we say we do not know. Where our model simplifies, we document the simplification. Where a statement comes with a range, we show the range explicitly as ±20%. This methodology is not only more honest, it is also scientifically sounder. Studies on the perception of data tools (Stanford, MIT, IDEO) consistently show that tools with transparent limits generate more adoption and trust than tools with perfect claims.
In the following six sections, we explain each limitation individually: what it means, why it exists, and what consequences it has for the interpretation of our calculations. Anyone who has read all six can place our numbers in serious context and draw their own conclusions.
Where the model still calculates is described in the pass-through effect.
Limitation 1: No forecast for individual households
We calculate additional costs for a typical household of your size and country. What we do not know: your specific car, your heating model, your electricity provider, your contract type, your commute on a given day. All these factors influence the real pass-through for your household.
Concrete example: a four-person household in Hamburg with a modern heat pump, solar array and an electric car experiences a much lower pass-through than a same-size household in Münsterland with oil heating, an old diesel car and a standard electricity tariff. Our model returns similar values for both because it only knows the household size and country.
The ±20% range reflects exactly this spread: 30% of households of your configuration experience significantly more, 30% significantly less, 40% lie in the middle. This is not a weakness of our model but an honest representation of what a statistical model can deliver. Anyone who wants an individual forecast should feed their own consumption data into a detailed energy audit tool that calculates per device.
Limitation 2: No exchange-rate prediction
Brent is traded in USD; your energy bill arrives in EUR or CHF. An increase in the Brent price therefore has a double impact: through the USD rise itself and through any exchange-rate movement. We use the spot exchange rate at the time of calculation but do not model an exchange-rate forecast.
In stable market phases this is harmless, because USD movements are small in the short term. In crisis phases, however, the exchange rate can swing significantly. Example: if Brent rises by 10 USD while USD appreciates by 5% against EUR at the same time, the effective pass-through for EUR households is about one third higher than our model shows.
Anyone who wants to model a crisis scenario with an exchange-rate component can manually enter a higher Brent price in the Wizard to compensate for the exchange-rate movement. Example: with an expected USD appreciation of 5% against EUR, multiply the Brent shock value by 1.05. That is a pragmatic approximation, not an exchange-rate model.
Limitation 3: No speculation premium
In crisis scenarios, traders add a risk premium to the Brent spot price. This premium reflects geopolitical risk, uncertainty about future deliveries, and speculative expectations. It can range between 5 and 25 USD per barrel and is not included in our pass-through coefficients.
Example: during a Hormuz escalation, the Brent spot price might show 95 USD, but the market consensus for the next 3 months is at 110 USD because of the risk premium. Our model calculates with the current spot, not with the forward price. Anyone who wants to see the forward effect should enter a correspondingly higher Brent value manually in the Wizard to model the expected premium.
The risk premium is methodologically hard to forecast because it depends on political escalation, market sentiment and trader behavior. IEA and EIA document historical risk-premium movements, but no serious source forecasts them in the short term. We therefore deliberately do not include them in our model.
Limitation 4: Constant consumption over time
Our model holds user consumption constant over the 12-month horizon. In reality, households respond to price shocks: when fuel prices are high, some drive less, others switch to public transport, some buy an electric car. When heating costs are high, more insulation is added or a heat pump is installed.
These behavioral changes can significantly reduce the real pass-through, sometimes by 15 to 30%. Our model does not model them because they are highly individual and difficult to estimate in aggregate statistics. ACEEE studies show that behavioral changes during energy price shocks take about 6 to 12 months to become measurable.
Anyone who wants to assess their own adaptation options should look at the saving tips in our tip area. Each tip shows the theoretical and the realistic (30% capture) saving effect. Anyone who consistently implements multiple tips can push the actually experienced pass-through significantly below our model estimate.
Limitation 5: No regional subsidies
In crisis scenarios, governments often intervene with subsidies and tax cuts. Examples: Klimabonus AT (annual payment to households), heating cost subsidy DE (one-time payment to low-income households), MaPrimeRenov FR (subsidy for heating modernization), Bono Social ES (social electricity tariff). These measures can significantly reduce the real pass-through.
Our model does not model them because they are politically unpredictable and change at short notice. What was available in 2024 may be abolished in 2026, as the Klimabonus AT was abolished in 2024. We do not want to give the impression that these measures provide reliable compensation, because they are politically fragile.
Anyone who wants to know which subsidies are currently available in their country should consult our country detail page for their country. There we document the currently active energy subsidies per country with source links. This is updated semi-annually.
Limitation 6: No political forecast
OPEC cuts, sanctions tightening, Hormuz escalations, political changes in producer countries: all these events influence the Brent price, often dramatically. Our model does not forecast political events, it only shows what would happen in a concrete Brent scenario that you choose yourself.
The choice of Brent scenario is therefore a political assumption, not a forecast. When you set the Brent price in the Wizard to 120 USD, that is your hypothesis, not our forecast. Our model calculates the consequences of this hypothesis. That is methodologically cleaner than a forecast because we are not an investment advisor and because serious political forecasts have rarely worked in energy market history.
Anyone who wants scenario suggestions can consult our country detail pages. There we document the most likely crisis scenarios per country with historical comparison values. The choice of which scenario you want to model remains with you.
Confidence interval ±20% explained
When our model shows 30 Euro, what do the ±20% mean concretely?
The range ±20% reflects empirical dispersion in our data sources. It is not derived from a theoretical confidence interval but from real pass-through studies on concrete households. Concretely this means: when our model shows 30 Euro additional costs per month for a configuration, about 30% of households experience over 36 Euro, about 30% under 24 Euro, and about 40% lie in between.
This distribution is not symmetric around the model value. In fact there is a slight asymmetry: households with inefficient energy infrastructure (old oil heating, high fuel consumption) tend more often toward values above the mean. Households with modern infrastructure tend toward values below it. For simplicity we show a symmetric range but document here that the real distribution has a right-skewed tendency.
The range is defined in code as the RANGE_BAND constant (wizard.dev.js line 2149). It is fixed at 0.20 (20%), based on the aggregated dispersion in IEA, ACEEE and USDA studies on pass-through distributions in real households. With larger data sets the range would tend to be a bit narrower; with smaller ones a bit wider.
The source basis behind the calculation values is documented in our 9 data sources.
Frequently asked questions about our limitations
The questions we are asked most often about our limits.
Why do you not build a forecasting function?
A real forecast requires modeling geopolitics, exchange rates, OPEC decisions and speculative behavior. These factors cannot be predicted methodologically with seriousness. We have consciously decided against a forecasting function because we do not want to be an investment advisor and because a poor forecast does more harm than good. Instead we show what would happen in a concrete Brent scenario that you choose yourself.
What does the ±20% range mean concretely?
When our model shows 30 Euro additional costs per month for a four-person household in Germany, the ±20% range means: about 30% of households experience more than 36 Euro, about 30% less than 24 Euro, about 40% lie between 24 and 36 Euro. This distribution is empirically calibrated from IEA and ACEEE data, not derived from a theoretical confidence interval. It reflects real dispersion in the data.
Which limits apply particularly in crisis scenarios?
In crisis scenarios (Hormuz escalation, OPEC cut, sanctions tightening), several of our limits apply particularly strongly: we do not model a speculation premium (can add 5 to 25 USD per barrel), no political emergency measures (tax cuts, subsidies), and no behavioral changes (users switch to e-bikes or reduce heating). In a serious crisis scenario, the real pass-through for households can be 30 to 50% lower than our model shows because government interventions are activated.
How do you handle uncertain sources?
When a source has methodological weaknesses (e.g., small sample, unclear methodology, missing peer review), we do not use it. Instead, we search for studies from the nine established institutions (IEA, EIA, ACEEE, ADAC, TCS, BFE, USDA, FAO, Eurostat) that are methodologically clean. In case of conflicts between our sources (e.g., when IEA and EIA show different values), we document the discrepancy in a code comment and choose the value backed by the larger sample or better documented methodology.
Pass-through effect: full methodology with coefficients →
9 sources of our model →
Ready to calculate?
You now know the methodology behind this sub-page. Try the calculator with your own numbers, or return to the full methodology overview.
Calculate my number now