Toggle Mobile Menu
Public

Alberto Cavallo on Covid Inflation: Evidence from Real-Time Data

With introductions by Markus Brunnermeier
June 17, 2021
12:30 pm
Markus' Academy

More from this series
Online: Register to Receive Zoom Details

Video & Slides

On Thursday, June 17, Alberto Cavallo joined Markus’ Academy for a lecture. Alberto Cavallo is the Edgerley Family Associate Professor at Harvard Business School, a Faculty Research Fellow at the National Bureau of Economic Research, and a member of the Technical Advisory Committee of the US Bureau of Labor Statistics (BLS).

Watch the full presentation below and download the presentation slides. You can also watch all Markus’ Academy webinars on the Markus’ Academy YouTube channel.

Executive Summary

  • Online price indices have been produced in real-time since 2007, when Cavallo used a technique called web-scraping to show that the Argentinian government was lying about the inflation rate. In 2008, Cavallo and Rigobon created the Billion Prices Project to expand the data collection to more countries. Since 2010, a private firm called PriceStats has been collecting daily online prices from over 1000 large retailers and publishing real-time inflation data in 22 countries, PPP indicators in 8 countries, and sharing the result with Central Banks, policymakers, NSOs, and researchers. The micro data is used for research purposes.  
  • Online price data are different from the CPIs for many reasons, but are particularly useful to anticipate changes in inflation trends. Online indices are available with daily frequency in real time (just a 3-day lag). The data provide the full price history for each good, and for all goods within each retailer. Some disadvantages include a lack of hedonics, seasonal adjustments, and other special index methods, as well as a more limited coverage (must be online retailers, and it is still difficult to cover services and housing). Historically, online prices have anticipated changes in US CPI inflation trends by 2-3 months, mostly due to CPI publications lags and the speed at which retailers change online prices.
  • Online indices detected a significant increase in the US inflation trend back in November 2020. The trend remains stable at a 5% annualized inflation rate. Similar Covid “turning points” were seen in other countries, with some differences in timing and magnitude. In almost all economies tracked with these data, annual inflation is now at the highest level in over a decade.   
  • Measurement distortions with CPI weights are affecting the annual inflation rate, exacerbating base effects and over-weighting price increases in transportation. The measurement of inflation during Covid has been distorted by changes in spending patterns that are not accounted by the CPI sectoral weights, which were last updated in December with data collected back in 2018. Covid is making US consumers spend relatively more on food and less transportation, a trend that continues today in real-time data. Cavallo argues this means that the US CPI was under-estimating inflation in 2020 (when food prices were increasing and transportation prices falling), therefore creating stronger “base effects” in the annual inflation rates in 2021. Moreover, today the CPI is over-estimating inflation because it is still placing too much weight on transportation, a sector with strong price increase in recent months (gas and used cars and trucks). The distortions also affected core sectors, and caused inflation inequality because the spending patterns of low-income households differ from high-income households. Although temporary, these measurement distortions added about 0.9% to the US annual inflation rate in May 2021.
  • Supply disruptions caused by Covid led to a significant increase in stockouts and the annual inflation rate for affected goods. Disruptions were primarily caused by operational shut-downs, hoarding, changing distribution channels, higher operating costs, and global supply-chain bottlenecks. Temporary stockouts increased from about 12% pre-Covid and peaked around 23%, before returning to a lower level in late 2020. However, many goods disappeared from the stores, meaning product availability remains about 20% below Pre-Covid levels in 2021. These “permanent stockouts” are back to normal in sectors such as health and household goods, but remain persistently high in food and electronics. The impact on prices, which peaks at 6 weeks, can explain about half of the difference in annual inflation this past year relative to historical averages. 
  • Supply disruptions appear to be temporary, but there are also reasons to worry. Stockouts are starting to fall in many sectors, and the price impact is limited to sectors where the disruption is more persistent. However, the price effects will remain for a while, potentially contributing to higher inflation expectations. Additionally, Covid moved many transactions online, where prices tend to react faster to shocks. This will likely introduce additional volatility to the inflation rate.