As businesses begin to pass tariff-related costs to consumers, one factor might make that sting feel worse: retail prices were moderating from pandemic surges before the levies began.
Research by Harvard Business School Professor Alberto Cavallo illustrates the downward trend in the price levels for many retail goods, followed by an acceleration after tariff announcements. Prices on both imported and domestic goods have climbed modestly but steadily since March, even if the hike is still small relative to the size of the tariffs.
Cavallo, the Thomas S. Murphy Professor of Business Administration, and HBS research assistants Paola Llamas and Franco Vazquez, both from the Universidad de San Andrés in Buenos Aires, have been tracking how tariffs impact prices. They coauthored the working paper “Tracking the Short-Run Price Impact of US Tariffs,” documenting their initial findings in April.
The researchers created indexes with daily prices collected by PriceStats, a private firm cofounded by Cavallo that provides online data for over 350,000 products sold by five major US retailers. The indexes allow them to track price changes in specific categories and from countries of origin. Overall, the prices of imported products have increased faster than those made in the US.
An extended analysis, going back to January 2024, explores price changes of goods relative to their pre-tariff trend.
Following rising prices
Cavallo’s team publishes findings weekly through the HBS Pricing Lab Tariff Tracker. Their latest data highlights steady increases in post-tariff prices, compared with charts we published in June.
Cavallo’s research found that Chinese products have been the most affected of the three main US trade partners, with prices rising faster than those of US domestic goods.
“These diverging country-specific price trajectories likely reflect the price-setter’s expectations about tariff exposure and bilateral trade negotiations. China faced the highest effective tariffs and the fewest exemptions,” the authors wrote.
The researchers identified the country of origin for 308,000 products from three US retailers and then used a generative AI model to find the source of about 15% of goods without a matching country. They validated the AI model’s results with 10,000 products with known origins, and found that it was 88 percent accurate in predicting whether a product was “domestic” or “imported” and 85 percent accurate in identifying the correct country.
About one third of the total products analyzed come from China and another 37% are domestic US goods.
