A handful of items are driving inflation in America

CONSUMER-PRICE inflation has risen to 5.4% in America, the highest in 30 years. On November 3rd the Federal Reserve said it would taper bond purchases, a step towards higher interest rates. Most economists say that this bout of inflation is a result of temporary disruptions caused by covid-19, and that it will pass. But some think it presages a longer-term trend.

A leading argument by inflation doves has been that just a few items have caused a large share of total price increases. In the quarter to August used cars, hotel rooms and airfares made up less than 5% of America’s consumer-price index, but together accounted for the majority of overall inflation. “This is really extreme,” Paul Krugman, an economist, wrote at the time, “and does suggest transitory bottlenecks rather than broad-based inflation pressure.”

This case rests on two claims, which both merit scrutiny. One clearly stands up: when compared with past periods with similar inflation, current price rises are indeed unusually concentrated. The other—that inflation is likely to slow down as a result—is also broadly true. However, this effect is too small for the Fed to breathe easy.

To test these hypotheses, we built a dataset of price levels since 1959 for every item—from housing to lottery tickets—in the personal consumption expenditures (PCE) index, one of the Fed’s preferred metrics. For each rolling 12-month period, we calculated a measure of how much price changes vary between items: their standard deviation. When a few components account for a large share of inflation, this number is high. When most items’ prices change by similar amounts, it is low.

In general, standard deviations are correlated with inflation: the higher the average increase in prices, the more specific items’ price changes differ from each other. However, some eras were unusual, with inflation that was either low but concentrated, or high but broad. To identify such outliers, we measured the “excess” concentration of inflation in each time period: the gap between the actual standard deviation of price changes and what you would expect based on overall inflation.

This measure is now abnormally high. During the year to May inflation was more excessively concentrated than in 97% of rolling 12-month periods since 1961. It has dipped slightly as used-car prices have levelled off, but still sits in the 89th percentile.

What does this mean for future inflation? Historically, when excess concentration has been high, the present has been a poor guide to the future. When inflation is above its ten-year average, as it is now, high excess concentration makes it more likely to fall. This pattern should lead forecasters to reduce their predictions for inflation.

The notion that a few big price changes can lead forecasters astray is hardly new. In the 1970s economists devised “core” inflation, which excludes food and energy. More recently, “trimmed-mean” measures, which drop the items whose prices have swung the most, have come into vogue. The Dallas Fed has published papers showing that its version, which excludes the bottom 24% and top 31% of the PCE index, predicts inflation better than core does.

However, both of these methods have flaws. Changes in food and energy prices are not necessarily unusually large or short-lived. And trimmed means’ weighting schemes are plagued by abrupt cliffs. In the Cleveland Fed’s version, which lops off the top and bottom 8% of the index, an item in the 93rd percentile when sorted by price changes is removed entirely, whereas one in the 92nd gets its full weight.

With this in mind, we have devised an alternative inflation index. Like trimmed means, it adjusts items’ weights based on their recent price changes. But its weights are shaped like a smooth hill rather than a box. Components with inflation near the median get the most emphasis, and those with the biggest price changes get the least.

Our hill looks a bit like Uluru in Australia: a broad central plateau, flanked by a steep slope on the left side and a gentler one on the right. (The Dallas Fed’s trimmed mean is also asymmetric, counteracting bias caused by price-change distributions’ lopsidedness.) Most items with negative or low inflation get a hefty weight; those whose prices are rising fastest count for 25% as much as those in the middle do.

When using the past year of data to predict PCE inflation during the following year, this method is more accurate than either using core inflation or expecting inflation to remain constant. Since 1959, its one-year forecasts have also outperformed those of the Dallas Fed’s trimmed mean.

Some of this apparent advantage stems from the design of our study: the Dallas Fed sought to maximise accuracy for different time periods and forecast horizons than ours. However, its trimmed mean’s errors in the 1970s illustrate the risks of such a deep trim. Amid two oil shocks, some of the fastest-rising prices—those of energy-related items—just kept rising. The less weight an index placed on such goods, the worse it predicted inflation one year out.

The current episode of concentrated inflation differs from the 1970s in many ways. Surging prices for goods like home appliances are unlikely to feed through to other costs, as oil prices do. And in general, treating outliers the same as more representative items has been a mistake. But in cases where such price changes foreshadow broader supply constraints, ignoring them entirely can be an even bigger error.

As a sense-check of our “Uluru” method, we also built a model that forecasts PCE inflation using only excess concentration and the one- and ten-year trailing inflation rates. Both this approach and the Uluru index yield a 4.1% prediction for inflation during the next 12 months. That is below the current level of 4.4%, but above the Fed’s target of 2% and the 2.3% value of the Dallas Fed’s trimmed mean for the past 12 months. If this forecast comes true, interest-rate rises will almost surely follow.■

Sources: Bureau of Economic Analysis; The Economist