Central bankers are obsessed with price stability. Today, something like two dozen countries target inflation explicitly. The ECB says it doesn’t, but the Maastricht Treaty’s price stability mandate certainly makes the Eurosystem more of an inflation targeter than a number of countries that say they are. And top Federal Reserve officials have argued that the Fed’s three statutory objectives – maximum employment, stable prices, and moderate long-term interest rates – are complementary so that they are all met by keeping inflation low and stable. It really is the case that central bankers around the world agree that they should use their control of money and credit to keep prices from rising or falling on average.
This fixation on price stability has yielded results. Twenty-five years ago when inflation was high, it didn’t matter much how you measured it. At the height of the Great Inflation of the 1970s and early 1980s, prices were rising at more than a 10% rate in more than half the world’s industrialised countries. Regardless of how you used the price data, the result was the same: Inflation was high. Today, inflation is below 4% in every one of the world’s 30 developed economies. And when inflation is low, it matters how you measure it. This is where core inflation comes in.
But we are getting ahead of ourselves. Before discussing why we compute core inflation, let’s take a step back. The basics of inflation are straightforward: it is the extent to which prices are rising on average. If all prices were to go up in equal proportion, measuring inflation would be easy. But since some prices go up more than others, getting an accurate reading requires averaging across all of them. Doing the job well turns out to be surprisingly difficult, so there is a variety of ways to do it. The most common is a consumer or retail price index. The European Harmonized Index of Consumer Prices (HICP) is an example of such an index. Such measures are designed to answer the question “How much more would it cost for people to purchase today the same basket of goods and services that they actually bought at some fixed time in the past?" To construct a Consumer Price Index (CPI), government statisticians survey people every few years to measure what they purchased and combine this with monthly information on individual prices. Inflation measured using this index tells us how much more money we need to give someone to restore the purchasing power they had in the earlier period when the survey was done.
Consumer price indexes have two well known problems, one long-term and one short-term. Because of the way in which they are constructed, these indexes systematically and permanently overestimate inflation. To put it bluntly: The numbers are wrong. This problem is complex and difficult to solve, so instead, policymakers work around it by setting their inflation objectives slightly higher than they would if they had a perfect index.
The short-term problem with traditional price indexes is that over short intervals of a few months or so they contain temporary noise. For example, floods or droughts might drive prices of raw food temporarily higher, only to have those increases quickly reversed when the weather improves. Since this volatility is transitory, it is important that policymakers not respond to it. If interest rates were adjusted in response to the short-term inflation moves, it would amplify the noise, destabilizing the economy – the opposite of what we want them to do. This brings us back to core inflation: Core inflation measures are designed to solve the problem of transitory noise.
Before continuing, I should note that every country’s national statistician, as well as Eurostat, produces a variety of inflation measures that differ both conceptually and in important details. First, while CPI’s are based on out-of-pocket expenditure, there are also measures of inflation based on total household consumption, including items that people do not pay for directly, like health care. Second, different countries treat owner-occupied housing very differently. For example, in the US CPI, owners are assumed to rent their homes from themselves, so there is a very high weight, with this implicit rent accounting for nearly one-quarter of the index! By contrast, the current HICP has no role for owner-occupied housing. Third, there are differences in the frequency at which the weights are changed. This can be every several years, as is typical for CPI-type measures, or continuously, as some total consumption measures do. Experts believe that the more frequent adjustment of the weights minimises the bias.
Returning to core inflation, the indexes that exclude food and energy were born in the 1970s during the heyday of the Organization of Petroleum Exporting Countries as oil prices rose from $3.50 to $10 a barrel in 1974 and then to $40 in 1980. I saw the impact on inflation measurement close up. As a young economist on the staff of the White House Council of Economic Advisers in 1980, I saw US consumer price inflation hit its modern peak of nearly 20 per cent. In an attempt to improve appearances, we started computing various alternative measures of inflation. Half jokingly, one of my colleagues noted that our job was to remove all the components of the price index that rose by more than the average. What was left was the core.
More seriously, the purpose of a core inflation index is to get an accurate measure of the current inflation trend measured at a frequency of a year at most. As the interval over which inflation is calculated increases beyond twelve months, a properly constructed core measure should yield the same quantitative estimate of inflation as the headline index does. Let me say that again: Over a period of several years, headline and core inflation should be the same.
This has several important implications. First, it means that core inflation is not a forecast of future headline inflation. Anyone who wants an inflation forecast can do much better than simply look at core inflation. Second, the fact that core inflation is a measure of the current inflation trend, purged of transitory noise, means that there is no justification for using it as an explicit objective unless the central bank is targeting inflation over a horizon of a year or less – something that seems imprudent under nearly all circumstances.
Recent history suggests a danger inherent in targeting core inflation rather than the headline number. Since 1995 energy prices have been increasing at rates that are consistently faster than those of other prices. As a result, the traditional measure of core inflation that simply excludes food and energy prices has been a biased measure of headline inflation, running at roughly half a percentage point below the headline measure.
Not only do we all consume food and energy – that’s what they are doing in the price index in the first place – but it turns out that excluding them has imparted a bias to medium-term measures of inflation. Since the goal of policymakers is stable prices overall, including those of food and energy, they should turn their attention to forecasts of headline inflation and stop focusing on core measures.