# Do heterogeneity in consumption baskets lead to very different inflation rates for different consumers?

CPI is measured against a single consumption basket. In reality, consumers each have their own set of consumption goods, so each face their own "individualized CPI" rates.

But how different are those "individual CPI rates"? Has there been work on this?

Why I'm asking: economists frequently state that investors should invest in "inflation-adjusted products" to hedge against inflation. The problem is, if each individuals' consumption basket is sufficiently different from that used to compute CPI, holding inflation-linked securities may not be a hedge, but rather introduce more risk.

• Was my answer satisfactory? – luchonacho Aug 23 '17 at 12:37

This question is essentially about the representativeness of agents in economic models, and so the answer is yes and no.

Yes, there is considerable heterogeneity between investors/consumers. For example, a member of a suburban family with two children will buy a significantly different bundle of goods than a single pensioner. In addition, wealthier households will consume significantly more luxury goods than poorer ones. This means that inflation exposure is heterogeneous, with different investors experiencing different inflation risks due to the fact that they buy different bundles of goods.

But the answer is also no, consumption bundles may differ, but do they differ enough to justify not hedging inflation? Technically, hedging against a standardized inflation rate would only increase risk if the variation in interest rates over time is smaller than the variation between investors. This is an interesting empirical question, for which I did not find any published evidence after an (admittedly short) search.

However, I do know that the variation of inflation rates over time is considerable, between -0.10 and 8.50 monthly in the UK. I do not have any numbers on consumption bundle heterogeneity, but want to point out that there will be many commonalities underlying the price changes even in heterogenous consumption bundles (e.g. oil prices for transporting goods). Given these facts, it is likely that the commonalities outstrip the differences, and so hedging inflation is likely to reduce risk for the large majority of investors.

• Timo, in the short-term, inflation differs dramatically across households. See this paper. mpls.frb.org/research/wp/wp731.pdf. First two sentences of abstract: "We use scanner data to estimate inflation rates at the household level. Households' inflation rates are remarkably heterogeneous, with an interquartile range of 6.2 to 9.0 percentage points on an annual basis". – J Li Jun 3 '16 at 4:29
• Interesting paper! Perhaps I misunderstand what you are saying, but it seems to align well with what I said: inflation exposure is heterogeneous, but this does not defeat the purpose of hedging inflation risk. In their words: "The entire distribution of household inflation rates shifts in parallel with aggregate inflation. Deviations from aggregate inflation exhibit only slightly negative serial correlation within each household over time, implying that the difference between a household’s price level and the aggregate price level is persistent." – user7935 Jun 3 '16 at 12:33
• Although I note that I have not expressed something very well: it is not the dispersion of inflation rates among investors that matters, but the variation of one particular investor with respect to the mean inflation rate at a particular point in time. If investors will one year experience a much lower inflation rate, than a much higher one a year later, hedging with the mean rate adds risk. But if an investor's inflation rate is persistently higher than the mean inflation rate, hedging reduces risk. – user7935 Jun 3 '16 at 12:41

Your intuition is entirely correct. Inflation is calculated based on a "standard basket", whereas real baskets vary among different consumers. As such, the incidence of changes in prices affect consumers differently.

A good treatment of the issue is this piece by the IFS, UK. Basically, distributional effects of inflation would not exists if: