CBS introduces new energy price measurement method

Checking gas and electricity consumption by means of a smartphone app.
© ANP / Rob Engelaar

Erratum:

Because of an inaccuracy, the figures in the graph on inflation differed from the corresponding customized table. The figures in the customized table were correct, the figures in the graph have now been corrected.
From June 2023, Statistics Netherlands (CBS) will employ a new method to measure the energy prices paid by consumers. The aim is to derive a more accurate inflation rate. This article explains how and why this new method is being adopted.
Each month, CBS publishes the inflation rate for consumers in the Netherlands to show how much the prices of goods and services have changed. This rate is calculated by comparing a particular month with the same month one year earlier. This January, for example, inflation was 7.6 percent, which means that, on average, goods and services cost 7.6 percent more in January 2023 than they did in January 2022. Inflation is determined using the consumer price index (CPI). To compile this index, millions of prices are tracked on a monthly basis, covering all kinds of goods and services, from groceries and subscriptions to petrol, rent and energy.

How high energy prices affect inflation

For years, CBS has measured the price of gas and electricity using the monthly tariffs recorded by the Netherlands Authority for Consumers and Markets (ACM). Energy companies are obliged to report their tariffs to the ACM.
The tariffs ACM obtain from energy suppliers are those offered to customers when they sign a new contract. However, some households have a fixed contract with their energy supplier and pay pre-agreed tariffs over a longer period. In many cases, even households that do not have a fixed contract but stay with the same energy company do not pay the same price as new customers. When energy prices develop at a steady pace, the type of tariff used when calculating the inflation rate does not make a substantial difference. After all, under stable conditions, the change in tariffs for new contracts more or less keeps pace with the change in the tariffs that households pay on an ongoing basis.

In 2021, however, global energy prices began to rise, a development that was soon reflected in the tariffs for new contracts in the Netherlands. At this stage, people with fixed energy contracts remained unaffected by the rising prices. In the summer of 2021, this was true of approximately half of all households in the Netherlands. As a result, the gas and electricity price trends as measured by CBS became less representative of the price trends that households on average saw reflected in their energy bills. The change in tariffs for new contracts was no longer in step with the change in the tariffs actually being paid by many households.

Consumers with fixed energy contracts
jaarmaandElectricity ( %)Gas ( %)
2020January55.1257.36
2020February54.6156.70
2020March54.0756.70
2020April53.3355.28
2020May52.7354.66
2020June52.4854.33
2020July52.6754.02
2020August52.2753.62
2020September51.8953.23
2020October51.2852.62
2020November50.6151.96
2020December50.1051.48
2021January49.4850.83
2021February49.4750.83
2021March49.2150.59
2021April48.8050.15
2021May48.4149.73
2021June47.8849.21
2021July47.1048.75
2021August46.8948.51
2021September46.5948.19
2021October46.2847.84
2021November45.9547.47
2021December44.9246.39
2022January43.0544.43
2022February40.7942.01
2022March39.3440.44
2022April38.2939.41
2022May36.8737.96
2022June35.4836.61
2022July34.2335.37
2022August33.1434.30
2022September32.1733.32
2022October31.0532.17
2022November29.2430.38
2022December28.3229.49

In early 2022, CBS launched a study (Dutch only) into the impact of the measurement method used and the data and method required to ensure that the CPI takes in price rises that better reflect the price trends as experienced by households on average.

Results

On 31 October 2022, CBS published an article on its research into a new method for tracking and processing energy prices. This involved using transaction data provided by energy suppliers, so that the tariffs paid under long-standing energy contracts could be taken into account. The article compared the outcomes of various calculation methods, but these initial outcomes were adversely affected by substantial uncertainty about the availability and usability of data from energy suppliers. The resulting figures, generated by the various methods based on actual prices paid, still spanned a broad bandwidth in terms of possible outcomes.

Since then, considerable progress has been made: the most appropriate method has been identified and there is greater certainty about the data. The bandwidth for the outcomes of the new method is therefore considerably narrower and only relates to the remaining uncertainty about the final availability of data from the energy suppliers with whom this process is being coordinated.
The graphs below show the indices and resulting inflation rates, comparing the figures published to date with the research results using the new method based on transaction data. The new method made it possible to calculate a research-based index series from January 2020, and corresponding research-based inflation rates from January 2021.

Consumer price index
   Published (January 2020 = 100)Bandwidth of outcomes of the research series (January 2020 = 100)
2020January100100 - 100
2020February100.6100.6 - 100.6
2020March100.8100.9 - 100.9
2020April101.3101.4 - 101.4
2020May101.1101.2 - 101.3
2020June101.5101.6 - 101.6
2020July102.6102.7 - 102.7
2020August102102.1 - 102.2
2020September101.8101.9 - 101.9
2020October102.4102.4 - 102.4
2020November101.5101.6 - 101.6
2020December101.8101.8 - 101.8
2021January101.6101.5 - 101.5
2021February102.5102.3 - 102.4
2021March102.7102.6 - 102.6
2021April103.2103 - 103.1
2021May103.3103.1 - 103.1
2021June103.5103.2 - 103.3
2021July104103.6 - 103.7
2021August104.5104 - 104.1
2021September104.5103.9 - 104
2021October105.9104.8 - 104.8
2021November106.8104.8 - 104.9
2021December107.6105 - 105.1
2022January108.1104.2 - 104.7
2022February108.8105.5 - 105.9
2022March112.7106.6 - 106.9
2022April113.1108 - 108.4
2022May112.3108.6 - 109
2022June112.3109.3 - 109.8
2022July114.7110.5 - 110.8
2022August117111 - 111.4
2022September119.7111.7 - 112.2
2022October121113.6 - 114.9
2022November117.3115 - 117.1
2022December117.9115.4 - 117.8

Inflation
   Published (y-o-y % change of the CPI)Bandwidth of outcomes of the research series (y-o-y % change of the CPI)
2021January1.61.5 - 1.5
2021February1.81.7 - 1.7
2021March1.91.7 - 1.7
2021April1.91.6 - 1.6
2021May2.11.8 - 1.9
2021June21.7 - 1.7
2021July1.40.9 - 1
2021August2.41.9 - 1.9
2021September2.72.0 - 2.0
2021October3.42.3 - 2.3
2021November5.23.2 - 3.2
2021December5.73.2 - 3.2
2022January6.42.7 - 3.1
2022February6.23.1 - 3.4
2022March9.73.9 - 4.2
2022April9.64.8 - 5.2
2022May8.85.3 - 5.7
2022June8.65.9 - 6.3
2022July10.36.6 - 6.9
2022August126.7 - 7.0
2022September14.57.5 - 7.9
2022October14.38.4 - 9.6
2022November9.99.7 - 11.6
2022December9.69.9 - 12.1

In the course of 2021, a discrepancy arises between the published figures and the research results. From the end of 2021, this difference increases. With a growing number of households having to pay higher tariffs as their fixed contract expires, the prices observed using the new method increase rapidly, but less rapidly than the tariffs for new contracts. They peak in September-October and then begin to fall again, which leads to a drop in headline inflation. However, the research results based on the new method still show that many households are only just starting to pay the higher tariffs, so inflation continues to rise in this series of studies. By the end of 2022, this figure exceeds headline inflation for the first time. These results show that, regardless of whether energy prices rise or fall, the change is seen more quickly in the prices of new contracts as opposed to the prices being paid by consumers on an ongoing basis.

Where do we go from here?

Comparing the existing method with the new method of measuring and processing energy prices, we can see that the rise in energy prices is clearly visible in both. But by using the new method, this rise only occurs in the months when a large number of people are actually paying the higher prices. Using the existing method, a rapid rise or fall in energy prices is more likely to influence the CPI and the inflation rate. If the price trend is relatively stable, as was the case until the beginning of 2021, both methods give similar outcomes. But when prices are more volatile, the difference between the two methods can become more apparent.

Because the new method gives a more accurate picture of the actual increase in energy prices as experienced by consumers, CBS has decided to switch to this method. The new production process to facilitate this change is currently being set up. Starting with the inflation rate for June 2023 – a flash estimate of which will be published on the last day of that month – the new method will be incorporated into the published figures. The figures published until the end of May 2023, based on the existing method, will not be adjusted.

The implementation approach will be chosen in a way that, from June 2023, the CPI will be at a level it would have reached if the new method had been introduced in January 2021. Given that, in 2020, the CPI based on the existing method and the index in the research series are almost identical, the overall price development from 2020 to June 2023 and onwards based on the published figures will be almost identical to the price development measured over the same period using the new method.

The inflation rate is expressed as the percentage difference between the CPI in the reporting month and the CPI twelve months earlier. As figures which have already been published are not adjusted, inflation from June 2023 to May 2024 will be based on the CPI obtained using the new method compared to the CPI using the existing method. From June 2024, the inflation rate will be determined entirely according to the CPI compiled using the new method.