During the inflationary surge of 2021-2022, PPI demonstrated that the inflation rate for digital goods and services was lower than the inflation rate for “physical-economy” goods and services such as food, energy, and housing. In the digital sector, price increases were moderated by faster productivity growth and higher investment rates. In particular, key digital sectors such as broadband and ecommerce did not experience the sort of capacity squeeze which drove up prices in other parts of the economy.
The Federal Trade Commission, however, takes the direct opposite position. In its antitrust complaint against Amazon, the FTC argues that Amazon is behaving in a way that drives up online prices — not just for Amazon, but for other online sellers. The FTC writes:
Amazon’s conduct causes online shoppers to face artificially higher prices even when shopping somewhere other than Amazon.
Amazon deploys a series of anticompetitive practices that suppress price competition and push prices higher across much of the internet by creating an artificial price floor and penalizing sellers that offer lower prices off Amazon.
In order to justify its claim of Amazon monopoly power, the FTC paints an odd picture of high and rising online prices relative to brick-and-mortar prices. In particular, the FTC’s complaint would imply that online inflation is higher than brick-and-mortar inflation.
What does the data show? To answer this question, we analyze private sector and government data from Adobe, the Bureau of Labor Statistics, and from the Census Bureau. Each of these have shortcomings, but together they tell a consistent story of online prices rising slower than offline prices.
We start with the Adobe Digital Price Index (ADPI), which tracks online prices for 18 different categories of goods, including books, groceries, electronics, pet products, and apparel. This index goes back to 2014, but we focused on the period since 2019, when the FTC’s argument would suggest that any potential Amazon effect on prices would be larger.
We matched inflation in 16 of the 18 ADPI categories with comparable categories in the BLS Consumer Price Index, which is mostly weighted towards brick-and-mortar sales. We found that the median online price increase was 3.1% for the four years ending December 2023. Over the same period, the median price increase across the comparable 16 BLS categories, including mainly brick-and-mortar sales, was 10.4%.
For example, in the category of appliances, the ADPI showed a price increase of 1.6% from December 2019 to December 2023, while the CPI showed a price increase of 9.3%. In the category of personal care products, the ADPI showed a price increase of 7.2%, compared to a 10.8% price increase for the CPI. And in the category of sporting goods, the ADPI showed a price increase of 4.4%, compared to a 9.9% increase for the CPI.
True, there are some categories where online prices have risen faster than the comparable BLS CPI index. For example, online apparel prices rose by 8.7% according to the ADPI, compared to 5.6% in the CPI. But overall, online prices rose slower in 10 of the 16 categories.
We now look at a different data set from the BLS, the producer price indexes for retail trade. These price indexes measure trade margins—that is, the difference between the acquisition price of a good and the sale price to consumers. If margins are expanding faster in a particular retail industry, that is a sign that prices to consumers are increasing faster than the acquisition price of good.
Through December 2022, the BLS published a margin price index for “electronic and mail order shopping.” That margin only rose by 3.2% from December 2019 to December 2022. Over the same period, the margin price index for general merchandise stories — including department stores and big box retailers such as Walmart and Target — rose by 24.2%. This sort of disparity is not consistent with the FTC’s story of online prices increasing faster.
Finally, we look at the Census Bureau’s data on e-commerce spending as a share of total retail sales. Before the pandemic, the ecommerce share was rising at just under 1% per year. It obviously jumped during the pandemic, but then levelled off to 15.4% in 2023 (figure).
Based on pre-pandemic trends, we would have projected the ecommerce share to be 14.2% in 2023 and 15.0% in 2024 (the dashed line in the figure). To put it a slightly different way, the Census Bureau data suggests that the pandemic ended up moving the shift to ecommerce by only 1 year, or 1 percentage point.
This behavior is inconsistent with the FTC’s argument that online price inflation is higher than offline inflation. Higher online inflation, as the FTC claims, would have driven up the ecommerce share higher rather than lower, because consumers would have been spending more for the benefits of buying online. (To be a bit technical here, this conclusion also requires a low cross-price elasticity of demand between online and offline markets, which the FTC has already implicitly assumed in its complaint).
Indeed, the stagnation of the ecommerce share since the pandemic is more consistent with online inflation being slower than brick-and-mortar inflation, thus holding down ecommerce spending.
Let’s be clear: The data analyzed here are not perfect. Our analysis does not rule out the possibility that online prices basically track offline prices. Consumers are not dumb, and there’s nothing holding them back from buying at their local Target or Walmart rather than Amazon if particular online prices veer higher, or shift back to online purchases if offline prices go up. But the weight of the evidence suggests that online inflation has been lower than offline inflation across this period.
