Inflation rose across much of the world as the COVID-19 pandemic disrupted supply chains and changed consumption patterns during a period of social sequestration. In the United States, inflation topped 9% at its peak, but has since come down sharply although it remains stubbornly above the Federal Reserve’s (Fed) 2%target. In the U.S., the rate of inflation is primarily measured by two indicators: The Consumer Price Index (CPI), calculated by the Bureau of Labor Statistics (BLS), and Personal Consumption Expenditures (PCE), calculated by the Bureau of Economic Analysis (BEA). These indicators contain a variety of prices across several sectors such as food, energy, durable goods, services, transportation, and rent.
As commodity prices can change depending on the pull and push of demand and supply, among other factors, how much do commodity prices correlate with inflation indicators? Is there a lag in commodity prices correlating with these indicators?
Why are these questions of interest?
Many commodities are economic inputs and may play a role in the production of food and non-food products. The pricing of commodities may impact the production cost, the cost to consumers, shift the supply-demand dynamic, and affect consumer expectations of inflation. These changes in prices may be a reason for market participants with commodity exposure to hedge their risk.
During COVID-19, the global economy contracted, leading to a decline in the prices of many commodities. As the world gradually moved past COVID-19, many commodity prices recovered from their lows, and in some cases continued to move higher. One can argue that the higher prices coming out of COVID-19 were due to a combination of factors such as issues related to supply chains, production disruption, transportation, labor shortages, price recovery from COVID-era price declines, fiscal expansion in some countries, and a global economic recovery. Global growth in 2021 was an estimated 5.5%, the largest post-recession global growth in about 80 years.1
As the COVID-19 era moves into the rearview mirror, what can the data suggest about the relationship between commodities and inflation indicators?
How did commodities behave before and around the COVID-19 era?
Since 2000, commodities have experienced rallies and declines, as represented by the Bloomberg Commodity Index (BCOM) (Figure 1). The BCOM index is a broad-based commodity index representing 24 agricultural, energy and metals products.
Figure 1: BCOM Index Daily Prices
The BCOM rally in 2008 was primarily due to the energy markets, and was followed by a decline of several commodities at the end of 2008 triggered by the financial crisis slowing global growth. As the global economy revived after the crisis, the BCOM Index moved higher.
The BCOM Index then gradually declined prior to COVID-19 after peaking in May 2018 and bottomed out in March 2020 by about 32%. However, most of that decline occurred from January 2020 to March 2020 for a drawdown of about 27%. From BCOM’s bottom in March and April of 2020, prices recovered to their pre-COVID-19 level by January 2021. The commodity rally continued, with BCOM peaking in June 2022 before declining about 32% to the recent bottom set in September 2024, placing the BCOM prices back to their August 2021 level.
Inflation Behavior
Figure 2 briefly defines four well-known inflation indices. The CPI and PCE, a preferred measure of the Fed, are similar, but with some differences. The CPI index tends to report at higher levels relative to the PCE. The core indices exclude the more volatile food and energy prices.
Figure 2: Definitions of Inflation Indicators
Name | Produced by | Definition | |
---|---|---|---|
CPI | Consumer Price Index | Bureau of Labor Statistics | Includes prices for goods and services an urban consumer is likely to purchase. Does not include exports.2 Weights based on household surveys.3 |
Core CPI | BLS | Same as above but excludes food and energy. | |
PCE | Personal Consumer Expenditures | Bureau of Economic Analysis | Measures urban and rural prices paid by consumers or paid by others on behalf of a consumer.4 Produced in the U.S. & imported to the U.S. It also tracks changes in consumer behavior. Weights based on business surveys.5 |
Core PCE | BEA | Same as above, but excludes food and energy |
The year-over-year inflation indices tend to have a high positive correlation of 0.94 for core CPI to core PCE as noted in Figure 3. The CPI to PCE relationship is 0.99. The correlation of CPI and PCE to their core counterparts maintains a relatively strong positive relationship at 0.77, but not as strong as they are to each other, implying the food and energy components are a likely influencing factor when excluded from the core indices.
Figure 3: Correlations of YoY Inflation. Jan 2000 to Nov 2024
Core PCE | Core CPI | PCE | CPI | |
---|---|---|---|---|
Core PCE | 1 | |||
Core CPI | 0.94 | 1 | ||
PCE | 0.90 | 0.77 | 1 | |
CPI | 0.86 | 0.77 | 0.99 | 1 |
Source: Bloomberg Professional (PCE DEFY Index, CPI YOY Index, PCE CYOY Index, CPI XYOY), CME Economics Research Calculations
The inflation indices in Figure 4 demonstrate the year-over-year inflation between CPI and PCE tend to be similar as it includes food and energy. It also shows that when food and energy are included, the inflation data shows a wider dispersion of YoY inflation relative to the core indices.
Figure 4: Year-over-Year PCE and CPI
Inflation experienced moments exceeding 4% in the early 2000s. The U.S. also briefly experienced deflation (declining prices), during the financial crisis as the global economy quickly slowed during that period. However, the core inflation remained above zero during this time and remained relatively stable around the PCE’s two percent inflation target until about April 2021. During the summer and fall of 2022, inflation peaked at between 5.6% and 9.10%, depending on which inflation index is cited. As inflation began its decline in 2022/ 2023, in late 2024, core CPI found at least a temporary bottom and moved sideways around 3.3%. The other inflation indicators started moving higher.
What is the relationship between commodity prices and inflation indices?
The monthly year-over-year correlations in Figure 5 shows the core inflation metrics with a lower correlation to BCOM. This is likely due to the core indices exclude food and energy.
Figure 5: Correlations of YoY BCOM to inflation indices Jan 2000 to Nov 2024
The correlations slightly increased when examining the data from July 2010. Once again, the core inflation retained a lower positive correlation as noted in Figure 6.6
Figure 6: BCOM Correlations to Inflation Indices
Since the CPI and PCE have about the same correlation to the BCOM index and the PCE is considered the Fed’s preferred inflation index as it better accounts for consumers’ spending habits and the weights can be changed more frequently. In addition, the 2% inflation target is based on the PCE7 and not the CPI, therefore the PCE subsectors were examined for their correlation to BCOM (Figure 7).
Figure 7: YoY BCOM Correlation to PCE Subsectors
The nondurable goods subsector had the highest correlation at 0.76. This is the category with food and energy, therefore appropriate for a commodity index to have a higher positive correlation to the nondurable goods category.
Measuring the R-Squared of BCOM to the PCE index suggests BCOM impacts about 45% of PCE’s variance. The R-Squared of BCOM to nondurable goods suggests BCOM impacts about 58% of the nondurable’s variance.
It can take some time for commodity price changes to make their way through the supply chain and reach the inflation indices, therefore it makes sense to test for a lagged correlation effect of commodity price changes to PCE changes (Figure 8). The results show the positive relationship of BCOM to PCE increases from 0.67 to 0.78 about four to five months later. The BCOM to nondurable goods correlation increases from 0.76 to 0.85 with a two- to three-month delay.
Figure 8: Lagged Correlation YoY BCOM to the YoY PCE Index and the Nondurable Subsector
The R-Squared of BCOM to the five-month lagged PCE suggests BCOM impacts about 61% of PCE’s variance. The R-Squared of BCOM to the four-month lagged nondurable goods suggests BCOM impacts about 72% of the nondurable variance.
Bottom Line
The data suggests commodity prices (BCOM) have a relatively high positive correlation to the PCE index, and to the PCE’s nondurable goods subsector. The results also show that by lagging the data for three to five months may result in a higher correlation as it takes time for the commodity price changes to find their way through the supply chain. Going forward, these relationships could be interesting to keep abreast of.
References
- https://www.usitc.gov/research_and_analysis/tradeshifts/2021/special_topic
- https://fredblog.stlouisfed.org/2023/03/the-differences-among-price-indexes/
- https://www.bea.gov/research/papers/2007/reconciliation-between-consumer-price-index-and-personal-consumption
- Johnson, N. (2017). A comparison of PCE and CPI: Methodological Differences in U.S. Inflation Calculation and their Implications. Bureau of Labor Statistics, Washington, D.C.
- https://www.bea.gov/research/papers/2007/reconciliation-between-consumer-price-index-and-personal-consumption
- Examined data beginning in July 2010, due to the subsector data available as of July 2010.
- https://www.federalreserve.gov/economy-at-a-glance-inflation-pce.htm
Commodity Index Products
Gain exposure to commodities while potentially reducing volatility on your portfolio.
All examples in this report are hypothetical interpretations of situations and are used for explanation purposes only. The views in this report reflect solely those of the author and not necessarily those of CME Group or its affiliated institutions. This report and the information herein should not be considered investment advice or the results of actual market experience.