Tag: Median CPI

Is Sugar All You Need?

Dylan’s Candy Bar in New York City (Photo from the New York Times)

Can prices of one type of good track inflation accurately? As we’ve discussed in a number of blog posts (for instance, here, here, and here), there is a debate among economists about which of the data series on the price level does the best job of tracking the underlying rate of inflation.

The most familiar data series on the price level is the consumer price index (CPI). Core CPI excludes the—typically volatile—food and energy prices. In gauging whether it is achieving its goal of 2 percent annual inflation, the Federal Reserve uses the personal consumption expenditures (PCE) price index. The PCE price index includes the prices of all the goods and services included in the consumption category of GDP, which makes it a broader measure of inflation than the CPI. To understand the underlying rate of inflation, the Fed often focuses on movements in core PCE.

With the increase in inflation that started in the spring of 2021, some economists noted that the prices of particular goods and services—such as new and used cars and housing—were increasing much more rapidly than other prices. So some economists concentrated on calculating inflation rates that excluded these or other prices from either the CPI or the PCE.

For example, the following figure shows the inflation rate measured by the percentage change from the same month in the previous year using the median CPI and using the trimmed mean PCE. If we list the inflation rate in each individual good or service in the CPI, median inflation is the inflation rate of the good or service that is in the middle of the list—that is, the inflation rate in the price of the good or service that has an equal number of higher and lower inflation rates. The trimmed mean measure of PCE inflation is compiled by economists at the Federal Reserve Bank of Dallas by dropping from the PCE the goods and services that have the highest and lowest rates of inflation. During the period when the inflation rate was increasing rapidly during 2021 and 2022, CPI inflation increased more and was more volatile than PCE inflation. That difference between movements in the two price level series is heightened when comparing median inflation in the CPI with trimmed mean inflation in the PCE. In particular, using trimmed mean PCE, the inflation of late 2021 and 2022 seems significantly milder than it does using median CPI.

The United States last experienced high inflation rates in the 1970s, when few people used personal computers and easily downloading macroeconomic data from the internet wasn’t yet possible. Today, it’s comparatively easy to download data on the CPI and PCE and manipulate them to investigate how the inflation rate would be affected by dropping the prices of various goods and services. It’s not clear, though, that this approach is always helpful in determining the underlying inflation rate. In a market system, the prices of many goods and services will be affected in a given month by shifts in demand and supply that aren’t related to general macroeconomic conditions.

In a recent blog post, economists B. Ravikumar and Amy Smaldone of the Federal Reserve Bank of St. Louis note that there is a strong correlation between movements in the prices of the “Sugar and Sweets” component of the CPI and movements in the overall CPI. Their post includes the following two figures. The first shows the price level since 1947 calculated using the prices of all the goods and services in the CPI (blue line) and the price level calculated just using the prices of goods included in the “Sugar and Sweets” category (red line). The data are adjusted to an index where the value for each series in January 1990 equals 100. The second figure shows the percentage change from the previous month for both series for the months since January 2000.

The two figures show an interesting—and perhaps surprising—correlation between sugar and sweets prices and all prices included in the CPI. The St. Louis Fed economists note that although the CPI is only published once per month, prices on sugar and sweets are available weekly. Does that mean that we could use prices on sugar and sweets to predict the CPI? That seems unlikely. First, consider that the sugar and sweets category of the CPI consists of three sub-categories:

  1. White, brown, and raw sugar and natural and artificial sweetners
  2. Chocolate and other types of candy, fruit flavored rolls, chewing gum and breath mints
  3. Other sweets, including jelly and jams, honey, pancake syrup, marshmallows, and chocolate syrup

Taken together these products are less than 3 percent of the products included in the CPI. In addition, the prices of the goods in this category can be heavily dependent on movements in sugar and cocoa prices, which are determined in world wide markets. For instance, the following figure shows the world price of raw cocoa, which soared in 2024 due to bad weather in West Africa, where most cocoa is grown. There’s no particular reason to think that factors affecting the markets for sugar and cocoa will also affect the markets in the United States for automobiles, gasoline, furniture, or most other products.

In fact, as the first figure below shows, if we look at the inflation rate calculated as the percentage change from the same month in the previous year, movements in sugar and sweets prices don’t track very closely movements in the overall CPI. Beginning in the summer of 2022—an important period when the inflation that began in the spring of 2021 peaked—inflation in sugar and sweets was much higher than overall CPI inflation. Anyone using prices of sugar and sweets to forecast what was happening to overal CPI inflation would have made very poor predictions. We get the same conclusion from comparing inflation calculated by compounding the current month’s rate over an entire year: Inflation in sugar and sweets prices is much more volatile than is overall CPI inflation. That conclusion is unsurprising given that food prices are generally more volatile than are the prices of most other goods.

It can be interesting to experiment with excluding various prices from the CPI or the PCE or with focusing on subcategories of these series. But it’s not clear at this point whether any of these adjustments to the CPI and the PCE, apart from excluding all food and energy prices, gives an improved estimate of the underlying rate of inflation.

How Will the Fed React to Another High Inflation Report?

In a recent podcast we discussed what actions the Fed may take if inflation continues to run well above the Fed’s 2 percent target. We are likely a step closer to finding out with the release this morning (April 10) by the Bureau of Labor Statistics (BLS) of data on the consumer price index (CPI) for March. The inflation rate measured by the percentage change in the CPI from the same month in the previous month—headline inflation—was 3.5 percent, slightly higher than expected (as indicated here and here). As the following figure shows, core inflation—which excludes the prices of food and energy—was 3.8 percent, the same as in January.

If we look at the 1-month inflation rate for headline and core inflation—that is the annual inflation rate calculated by compounding the current month’s rate over an entire year—the values seem to confirm that inflation, while still far below its peak in mid-2022, has been running somewhat higher than it did during the last months of 2023. Headline CPI inflation in March was 4.6 percent (down from 5.4 percent in February) and core CPI inflation was 4.4 percent (unchanged from February). It’s worth bearing in mind that the Fed’s inflation target is measured using the personal consumption expenditures (PCE) price index, not the CPI. But CPI inflation at these levels is not consistent with PCE inflation of only 2 percent.

As has been true in recent months, the path of inflation in the prices of services has been concerning. As we’ve noted in earlier posts, Federal Reserve Chair Jerome Powell has emphasized that as supply chain problems have gradually been resolved, inflation in the prices of goods has been rapidly declining. But inflaion in services hasn’t declined nearly as much. Last summer he stated the point this way:

“Part of the reason for the modest decline of nonhousing services inflation so far is that many of these services were less affected by global supply chain bottlenecks and are generally thought to be less interest sensitive than other sectors such as housing or durable goods. Production of these services is also relatively labor intensive, and the labor market remains tight. Given the size of this sector, some further progress here will be essential to restoring price stability.”

The following figure shows the 1-month inflation rate in services prices and in services prices not included including housing rent. Some economists believe that the rent component of the CPI isn’t well measured and can be volatile, so it’s worthwhile to look at inflation in service prices not including rent. The figure shows that inflation in all service prices has been above 4 percent in every month since July 2023. Inflation in service prices increased from 5.8 percent in February to 6.6 percent in March . Inflation in service prices not including housing rent was even higher, increasing from 7.5 percent in February to 8.9 percent in March. Such large increases in the prices of services, if they were to continue, wouldn’t be consistent with the Fed meeting its 2 percent inflation target.

Finally, some economists and policymakers look at median inflation to gain insight into the underlying trend in the inflation rate. If we listed the inflation rate in each individual good or service in the CPI, median inflation is the inflation rate of the good or service that is in the middle of the list—that is, the inflation rate in the price of the good or service that has an equal number of higher and lower inflation rates. As the following figure shows, although median inflation declined in March, it was still high at 4.3 percent. Median inflation is volatile, but the trend has been generally upward since July 2023.

Financial investors, who had been expecting that this CPI report would show inflation slowing, reacted strongly to the news that, in fact, inflation had ticked up. As of late morning, the Dow Jones Industrial Average had decline by nearly 500 points and the S&P 5o0 had declined by 59 points. (We discuss the stock market indexes in Macroeconomics, Chapter 6, Section 6.2 and in Microeconomics and Economics, Chapter 8, Section 8.2.) The following figure from the Wall Street Journal shows the sharp reaction in the bond market as the interest rate on the 10-year Treasury note rose sharply following the release of the CPI report.

Lower stock prices and higher long-term interest rates reflect the fact that investors have changed their views concerning when the Fed’s Federal Open Market Committee (FOMC) will cut its target for the federal funds and how many rate cuts there may be this year. At the start of 2024, the consensus among investors was for six or seven rate cuts, starting as early as the FOMC’s meeting on March 19-20. But with inflation remaining persistently high, investors had recently been expecting only two or three rate cuts, with the first cut occurring at the FOMC’s meeting on June 11-12. Two days ago, Neel Kashkari, president of the Federal Reserve Bank of Minneapolis raised the possibility that the FOMC might not cut its target for the federal funds rate during 2024. Some economists have even begun to speculate that the FOMC might feel obliged to increase its target in the coming months.

After the FOMC’s next meeting on April 30-May 1 first, Chair Powell may provide some additional information on the committee’s current thinking.

Soft Landing, Hard Landing … or No Landing?

During the recovery from the Covid–19 pandemic, inflation as measured by the personal consumption expenditures (PCE) price index, first rose above the Federal Reserve’s target annual inflation rate of 2 percent in March 2021. Many economists inside and outside of the Fed believed the increase in inflation would be transitory because it was thought to be mainly the result of supply chain problems and an initial burst of spending as business lockdowns were ended or mitigated in most areas.

Accordingly, the Federal Open Market Committee (FOMC) kept its target for the federal funds rate at effectively zero (a range of 0 to 0.25 percent) until March 2022 and continued its quantitative easing (QE) program of buying long-term Treasury bonds and mortgage-backed securities (MBS) until that same month.

As the following figure shows, by March 2022 inflation had been well above the FOMC’s target for a year. The Fed responded by raising its target for the federal funds rate and switched from QE to quantitative tightening (QT). Although some supply chain problems were still contributing to the high inflation rate during the spring of 2022, the main driver appeared to be very expansionary monetary and fiscal policies. (This blog post from May 2021 has links to contributions to the debate over macro policy at the time. Glenn’s interview that month with the Financial Times can be found here. In November 2022, Glenn argued that overly expansionary fiscal policy was the main driver of inflation in this op-ed in the Financial Times (subscription or registration may be required).We discuss inconsistencies in the Fed’s forecasts of unemployment and inflation here. And in this post we discuss the question of whether the Fed made a mistake in not attempting to preempt inflation before it accelerated.)

Since March 2022, the FOMC has raised its target for the federal funds rate multiple times. In February 2023, the target was a range of 4.50 to 4.75 percent. Longer-term interest rates have also increased. In particular, the average interest rate on residential mortgage loans increased from 3 percent in March 2022 to 7 percent in November 2022, before falling back to around 6 percent in February 2023.  In the fall of 2022, there was optimism among some economists that the Fed had succeeded in slowing the economy enough to put inflation on a path back to its 2 percent target. Although many economists had expected that inflation would only return to the target if the U.S. economy experienced a recession—labeled a hard landing—the probability that inflation could be reduced without a recession—labeled a soft landing—appeared to be increasing. 

Economic data for January 2023 made a soft landing seem less likely. Consumer spending remained above its trend from before the pandemic, employment increases were unexpectedly high, and inflation reversed its downward trend. A continuation of low rates of unemployment and high rates of inflation wasn’t consistent with either a hard landing or a soft landing. Some observers, particularly in Wall Street financial firms, began describing the situation as no landing. But given the Fed’s strong commitment to returning to its 2 percent target, the no landing scenario couldn’t persist indefinitely.

Many investors had anticipated that the FOMC would end its increases in the federal funds target by mid-2023 and would have made one or more cuts to the target by the end of the year, but that outcome now seems unlikely. The FOMC had increased the federal funds target by only 0.25 percent at its February meeting but many economists now expected that it would announce a 0.50 percent increase at its next meeting on March 21 and 22. Unfortunately, the odds of a hard landing seem to be increasing.

A couple of notes: Although there are multiple ways of measuring inflation, the percentage increase in the PCE is the formal way in which the FOMC determines whether it is hitting its inflation target. To judge what the underlying inflation is—in other words, the inflation rate likely to persist in at least the near future—many economists look at core inflation. In the earlier figure we show movements in core inflation as measured by the PCE excluding prices of food and energy. Note that over the period shown PCE and core PCE follow the same pattern, although core PCE inflation begins to moderate earlier than does core PCE.

Some economists use other adjustments to PCE or to the consumer price index (CPI) in an attempt to better measure underlying inflation. For instance, housing rents and new and used car prices have been particularly volatile since early 2020, so some economists calculate PCE or CPI excluding those prices, as well as food and energy prices. As we discuss in this blog post from last September some economists prefer median CPI as the best measure of underlying inflation. (We discuss some of the alternative ways of measuring inflation in Macroeconomics, Chapter 15, Section 15.5 and Economics, Chapter 25, Section 25.5.) Nearly all these alternative measures of inflation indicated that the moderation in inflation that began in the summer of 2022 had ended in January 2023. So, choosing among measures of underlying inflation wasn’t critical to understanding the current path of inflation. 

Finally, the inflation, employment, and output measures that in January seemed to show that the U.S. economy was still in a strong expansion and that the inflation rate may have ticked up are all seasonally adjusted. Seasonal adjustment factors are applied to the raw (unadjusted) data to account for regular seasonal fluctuations in the series. For instance, unadjusted employment declined in January as measured by both the household and establishment series. Applying the seasonal adjustment factors to the data resulted in the actual decline in employment from December to January turning into an adjusted increase. In other words, employment declined by less than it typically does, so on a seasonally adjusted basis, the Bureau of Labor Statistics reported that it had increased. Seasonal adjustments for the holiday season may be distorted, however, because the 2020–2021 and 2021–2022 holiday seasons occurred during upsurges in Covid. Whether the reported data for January 2023 will be subject to significant revisions when the seasonal adjustments factors are subsequently revised remains to be seen.  The latest BLS employment report, showing seasonally adjusted and not seasonally adjusted data, can be found here.

Should the Fed Be Looking at the Median CPI?

For years, all the products for sale in Dollar Tree stores had a price of $1.00 or less. But as inflation increased, the company had to raise its maxium prices to $1.25. (Thanks to Lena Buonanno for sending us the photo.)

There are multiple ways to measure inflation. Economists and policymakers use different measures of inflation depending on the use they intend to put the measure of inflation to. For example, as we discuss in Macroeconomics, Chapter 9, Section 9.4 (Economics, Chapter 19, Section 19.4), the Bureau of Labor Statistics (BLS) constructs the consumer price index (CPI) as measure of the cost of living of a typical urban household. So the BLS intends the percentage change in the CPI to measure inflation in the cost of living as experienced by the roughly 93 percent of the population that lives in an urban household. (We are referring here to what the BLS labels CPI–U. As we discuss in this blog post, the BLS also compiles a CPI for urban wage earners and clerical workers (or CPI–W).)

As we discuss in an Apply the Concept in Chapter 15, Section 15.5, because the Fed is charged by Congress with ensuring stability in the general price level, the Fed is interested in a broader measure of inflation than the CPI. So its preferred measure of inflation is the personal consumption expenditures (PCE) price index, which the Bureau of Economic Analysis (BEA) issues monthly. The PCE price index is a measure of the price level similar to the GDP deflator, except it includes only the prices of goods and services from the consumption category of GDP. Because the PCE price index includes more goods and services than the CPI, it is suits the Fed’s need for a broader measure of inflation. The Fed uses changes in the PCE to evaluate whether it’s meeting its target of a 2 percent annual inflation rate.

In using either the percentage change in the CPI or the percentage change in the PCE, we are looking at what inflation has been over the previous year. But economists and policymakers are also looking for indications of what inflation may be in the future. Prices of food and energy are particularly volatile, so the BLS issues data on the CPI excluding food and energy prices and the BEA does the same with respect to the PCE. These two measures help avoid the problem that, for example, a period of high gasoline prices might lead the inflation rate to temporarily increase. Note that inflation caclulated by excluding the prices of food and energy is called core inflation.

During the surge in inflation that began in the spring of 2021 and continued into the fall of 2022, some economists noted that supply chain problems and other effects of the pandemic on labor and product markets caused the prices of some goods and services to spike. For example, a shortage of computer chips led to a reduction in the supply of new cars and sharp increases in car prices. As with temporary spikes in prices of energy and food, spikes resulting from supply chain problems and other effects of the pandemic might lead the CPI and PCE—even excluding food and energy prices—to give a misleading measure of the underlying rate of inflation in the economy. 

To correct for this problem, some economists have been more attention to the measure of inflation calculated using the median CPI, which is compiled monthly by economists at the Federal Reserve Bank of Cleveland. The median CPI is calculated by ranking the price changes of every good or service in the index from the largest price change to the smallest price change, and then choosing the price change in the middle. The idea is to eliminate the effect on measured inflation of any short-lived events that cause the prices of some goods and services to be particularly high or particularly low. Economists at the Cleveland Fed have conducted research that shows that, in their words, “the median CPI provides a better signal of the underlying inflation trend than either the all-items CPI or the CPI excluding food and energy. The median CPI is even better at forecasting PCE inflation in the near and longer term than the core PCE price index.”

The following figure shows the three measures of inflation using the CPI for each month since January 2019. The red line shows the unadjusted CPI, the green line shows the CPI excluding food and energy prices, and the blue line shows median CPI. To focus on the inflation rate in a particular month, in this figure we calculate inflation as the percentage change in the index at an annual rate. That is, we calculate the annual inflation rate assuming that the inflation rate in that month continued for a year.

Note that for most of the period since early 2021, during which the inflation rate accelerated, median inflation was well below inflation measured by changes in the unadjusted CPI. That difference reflects some of the distortions in measuring inflation arising from the effects of the pandemic.

But the last two values—for July and August 2022—tell a different story. In those months, inflation measured by changes in the CPI excluding food and energy prices or by changes in median CPI were well above inflation measured by changes in the unadjusted CPI.  In August 2022, the unadjusted CPI shows a low rate of inflation—1.4 percent—whereas the CPI excluding food and energy prices shows an inflation rate of 7.0 percent and the median CPI shows an inflation rate of 9.2 percent. 

We should always be cautious when interpreting any economic data for a period as short as two months. But data for inflation measured by the change in median CPI may be sending a signal that the slowdown in inflation that many economists and policymakers had been predicting would occur in the summer of 2022 isn’t actually occurring. We’ll have to await the release of future data to draw a firmer conclusion.

Sources: Michael S. Derby, “Inflation Data Scrambles Fed Rate Outlook Again,” Wall Street Journal, September 14, 2022; Federal Reserve Bank of Cleveland, “Median CPI,” clevelandfed.org; and Federal Reserve Bank of St. Louis.