Introduction of new method for the clothing price index

CBS is constantly improving its methodology price observations. One example is the automatic registration of prices in webshops using the so-called web scraping method. CBS has introduced this method in the measurement of clothing prices over the past few years. After conducting small-scale experiments in 2015 and 2016, web scraping was implemented on a larger scale in 2017 and it became the sole method used in the measurement of clothing prices as of December 2017.

This method offers several advantages. Prices on websites can be observed integrally and several times per month. This allows for close monitoring of developments in the supply of clothing. Observation is also much less costly. The method previously used by price observers took a much smaller representative sample of prices and observers had to indicate in advance which items needed observation in which month.

As a consequence, replacement of items under observation in the samples took place in a particular month (winter by summer collection items, for example). In many cases, the last observed price for winter clothing was a sales price, resulting in a sharply rising price index for clothing from one month to the next. With the new method, prices of summer clothes are observed straight from the point of introduction while winter items stay under observation as well for a longer amount of time. The resulting price index has a comparatively flat curve during changes in seasonal fashion collections.

Therefore, the results of the old and the new method show a different seasonal pattern. As long as only one method is used consistently, this will produce correct year-on-year price index changes. Replacement of one method by another takes place at a year transition, in accordance with Eurostat guidelines. In the first year, this may lead to some distortion in the year-on-year changes which are published monthly. In the month of March, measurements under the old method always showed a much sharper increase in the clothing price index compared to the new method; therefore in this month, there will be a decline in year-on-year changes as a result.