Do prices rise at the same pace on the internet as they do on Main Street? Inflation is actually lower online than in brick-and-mortar stores, according to Chicago Booth’s Austan D. Goolsbee and Stanford’s Peter J. Klenow.
The accepted way of measuring inflation is with the Consumer Price Index, which is released monthly by the US Bureau of Labor Statistics. The CPI reflects the prices urban consumers pay for a basket of goods and services—including household goods, recreational goods, food and beverages, information and communications technology, and medicines and medical supplies—mostly in traditional (as opposed to e-commerce) transactions.
But the share of spending going to e-commerce has almost tripled in the past decade, to 10 percent overall, report the researchers, citing census data. They set to find out: How does inflation in online sales compare to inflation in overall sales, and how do new products change the results?
The CPI samples only a fraction of each retailer’s products. However, Adobe Analytics, which is used by 80 percent of Fortune 500 retailers, gathers monthly data about millions of online transactions, and the data represent about 15 percent of all retail e-commerce tabulated by the US Census Bureau.
The researchers used the Adobe data to create what they call the Digital Price Index, matching as many CPI categories as possible. They find lower inflation on the DPI than the CPI in all categories except medicine and medical supplies. Between 2014 and 2017, inflation for online goods was at least 1 full percentage point lower every year than inflation for equivalent goods in the CPI, Goolsbee and Klenow write.
Moreover, they find that the turnover in products offered online lowered inflation further.
Because the Adobe data includes the quantities of goods sold and average prices paid, Goolsbee and Klenow looked at spending on goods that were either new to the market or were offered one year but not the next. There were a lot of these: even after the researchers controlled for apparel products, which are sold seasonally, 44 percent of online sales in the data set came from new products.
The entering and exiting products had a strong impact on prices. When the researchers factored these in, inflation dropped by another 1.5 to 2.5 percentage points per year. Thus, adjusting the DPI inflation rate for new products suggests to Goolsbee and Klenow that online inflation has been more than 3 percentage points per year lower than the CPI inflation rate.
“If online pricing is fundamentally different than traditional retail, its spread could have a rising impact on the overall Consumer Price Index and, potentially, bias it,” the researchers write.