Press Release - Measuring inflation: the stakes and the state of play
Article published in the Economic Review, December 2013
A major overhaul of the national consumer prices index (NCPI), upon which the health index is based, is currently in the works. The new index will come into force in January 2014. In addition to a complete update of the index’s weighting system and the addition and removal of certain products in order to better reflect households’ new consumption patterns, the reform – which happens every eight years – is also an opportunity to adjust the methodology to make inflation measurements more accurate and ensure that the index is representative.
Furthermore, at the government’s behest, two changes to the national index went into effect ahead of schedule, in January 2013. First of all, the effect of seasonal sales is now taken into account when compiling prices, as was already the case for the harmonised index of consumer prices (HICP, whose methodology is set by Eurostat), with the difference that the impact is spread proportionally over the entire year as opposed to just the sale months. Second, a “payment” approach has been adopted for heating oil contracts (trailing 12-month moving average).
These reforms, which allow the NCPI to reflect consumption habits more quickly and more accurately, should be encouraged, notably to avoid overestimating consumer price inflation. The reform of the national index in 2014 offers an opportunity to converge the two methodologies by moving into line with the HICP. Thus, one important reform recommended for the NCPI is the switch to a chain index, which allows for weightings to be adjusted more frequently (annually) in response to household consumption trends. Furthermore, taking better account of quality changes and substitution effects (when quantities consumed change in response to relative price movements) would also help measure inflation trends more precisely.
Other changes are to be recommended both for the NCPI and the HICP. They include switching to scanner data for products sold in supermarkets and using these data not just to partly replace traditional price surveys, but also to better reflect changes in the market shares of different kinds of retailers and different types of products, even if this can only be done implicitly. In addition, the method used for gas and electricity prices could be applied to other products, such as mobile phone contracts, here again to better reflect market share movements and the substitution effect. Lastly, pending a decision on whether or not to include owner-occupied housing costs in the HICP, some improvements can already be made to the current index of real rents in order to more accurately reflect rental market trends.
Some of the changes being considered for the NCPI were made in 2013 to the HICP (geometric mean, better measurement of quality adjustments). According to our estimates, these changes would likely cause inflation as measured by the NCPI to drop slightly if they were adopted in 2014. This downward impact would come on top of the one the NCPI experienced in 2013 when the seasonal sales effect was introduced. Furthermore, it could be strengthened by the use of scanner data (as suggested by the Netherlands’ experience) and applying the new method to telephone contracts. However, other methodology changes could offset some of the decline, notably better measurement of rent increases when tenants sign a new lease. Lastly, it is important to point out that the impact of these methodology changes on inflation will also depend on how the new index is introduced. As in 2006, the Index Commission will have to set the conversion coefficient enabling the transition from the old to the new index, which will influence the year-on-year change in the latter in 2014.