Chapter 1 Inflation Indices

Introduction

Inflation-indexed products (“linkers”) promise investors a certain real return, including the effect of inflation. The cash-flows of these products are linked to a price index (usually a consumer price index [CPI]), which captures the changes in the prices of goods and services of the country where the product is based. CPIs form the backbone of inflation markets and in this chapter we present the key characteristics and components of the major euro area and US price indices and discuss measurement issues as well as metrics for assessing underlying inflation. The inflation markets are constantly changing and thus metrics for size, flow and product popularity change with time. We have supplied dates for figures where relevant; otherwise the numbers are those extant at the end of 2021/spring 2022.

Euro Area – HICPx

Since its inception in 1998 the European linker market has evolved into a mature and liquid market. In the euro area, the benchmark index is the non-seasonally adjusted HICP ex tobacco (Harmonised Index of Consumer Prices [HICPx]). Tobacco prices tend to distort the index and thus the version of the index that omits them is more widely used.

German, French, Spanish and Italian inflation-linked bonds (ILBs) amount to €620.7bn or about 9.8 percent of government bonds outstanding for these issuers. Not all linkers are alike; France, for example, has linkers outstanding tied to its domestic inflation index (i.e., though they trade in euros they reference the French inflation index not the HICPx) and Italy has about €77.5bn in BTP Italia, which are linked to domestic inflation and target domestic retail investors. See Figure 1.1 for detail.

Figure 1.1: Euro Area Linker Volumes.Source: Bloomberg, Commerzbank Research.

Inflation swaps referenced to HICPx are actively traded across maturities and inflation options also trade on a regular basis.

HCPIx Versus CPI

As CPIs can vary on a national level, for example due to different calculation methods or coverage of households, HICPx provides a consistent framework across euro area countries.

Key differences between HICPx and national CPIs concern the treatment of owner-occupied housing (see also below for details), subsidized healthcare and education costs as well as the geographical coverage of households. The HICP is published on a monthly basis by Eurostat, typically two weeks after the end of the month. It can be found on Bloomberg via the CPTFEMU index. The current reference year for the index is 2015 (index = 100 at this point in time).

The following charts in Figure 1.2 illustrate the major constituents and country weights of HICPx. Note that household final consumption expenditure is the underlying concept for the product as well as country weights that can vary on an annual basis.

Figure 1.2: HICPx Composition.Source: Eurostat, Commerzbank Research.

HICPx is obtained by aggregating the respective country indices. There are pronounced differences in the item weights across countries, however, arising from the differences in domestic consumption patterns, as shown in Figure 1.3.

Figure 1.3: HICPx weights per country.Source: Eurostat, Commerzbank Research.

France – FRCPI

The French CPI (Indice des prix à la consommation [IPC]) is the other major inflation index in the euro area. Outstanding French government bonds linked to FRCPIx amounted to €66.7bn in 2021. The index is published by INSEE (Institut national de la statistique et des études économiques) at about the 22nd of the subsequent month. The current reference year for the index is 2015. It can be found on Bloomberg via the FRCPXTOB index.

The following chart in Figure 1.4 highlights the major constituents:

Figure 1.4: French CPI weights.Source: INSEE, Commerzbank Research.

HICPx Index Events and Their Market Implications

In February 2019, the euro inflation market underwent a set of adjustments following Eurostat’s revisions of HICPx series and the first official release of the new series. This had complex knock-on effects that it is worth understanding. The new treatment of packaged holidays (and other items) in the German HICP was at the core of these changes. To mitigate the sharp price swings in this item, they were to be spread throughout the calendar year from then on and no longer recorded in the month they fall due, which was to be achieved via constant rather than variable country weights (say for trips to Greece). These adjustments also resulted in significant adjustments of up to 0.3 percentage points for the euro area HICP.

However, they were only calculated back to January 2015, that is. affecting the base year (2015) and thus required a rescaling of the index to ensure that the average index level for 2015 continued to equal 100. Thus, HICPx levels after 2015 were recalculated and rescaled and those prior to 2015 only rescaled. Eurostat also provided the corresponding “implicit rebasing keys” to force the averages of the revised indices in 2015 back to 100, with 0.99837 being the key for HCIPx.

The Impact and Frictions

Although the year-on-year profile of HICPx was not affected, the seasonal profile changed significantly with the month-on-month changes in November and April particularly affected – and thus also the characteristics of bonds linked to HICPx (inflation-linked bonds or ILBs).

The major sticking point was the rescaling, however. The new HICPx series was significantly above the old (unrevised) series before 2015 by a constant factor and then exhibited sharp month-on-month swings due to the revisions (see Figure 1.5). Thus it was questioned whether Eurostat’s “rebasing key” should be applied to pre-2019 values of the unrevised index to ensure that CPI ratios between any two points in time are not affected by the methodological change.

Figure 1.5: HICPx change.Source: Eurostat, Commerzbank Research.

Don’t Call It Rebasing!

As euro sovereign debt management offices (DMOs) immediately made clear that base CPIs of their ILBs were not to be revised and future inflation HICPx is lower due to the rescaling, inflation accruals were thus reduced, constituting a windfall for DMOs of some €700mn (i.e., 16.3bp) at the expense of investors.

As derivative cash-flows linked to HICPx were affected in a similar fashion, the International Swaps and Derivatives Association (ISDA) assessed if the changes by Eurostat constituted a rebasing of the index or a material change of the index. If it had been rebased then the whole index would have to be rescaled. ISDA in consultation with market participants concluded that it was a rebasing and recommended that a rebasing key be applied to ensure the “economic integrity of the transactions.” One might argue that markets trade the first published index level, that is, that all this is irrelevant. However, rebasings of the first unrevised final published data do matter.

In addition, the documentation of German ILBs contains a very narrow definition of what constitutes a base year revision and the understanding at the time had been that Eurostat would have to officially declare (and calculate) such a rebasing – which seemed unlikely as Eurostat had quickly changed the official wording of its press release and spreadsheet to “back data linking factor,” no longer talking of a “implicit rebasing key,” presumably to underscore that it is not a base year revision.

Implications

A different treatment by the DMOs and ISDA would have resulted in a split between the cash and derivative market with the one trading on a rescaled series and the other on an unrevised one. Markets were clearly anticipating such a scenario judging by the aggressive moves in early 2019 (sharp swap outperformance versus cash).

In the end, such a split was avoided and the inflation market accepted the deflationary one-off event resulting from the methodological HICPx changes.

Does the HICPx Underestimate Inflation?

How strong is inflation in the euro area? According to official statistics, the HICPx inflation rate in 2020 was 0.3 percent. Excluding the always very volatile prices for energy, food and alcohol, the increase in the so-called core inflation rate was only modestly higher at 0.7 percent. However, it is widely believed that these figures underestimate the increase of the cost of living. Unlike many other countries outside the euro area and also unlike some national price statistics such as those in Germany, the Harmonised Index of Consumer Prices (HICPx) for the...