In early 2020 economists and statisticians faced a similar problem as 90 years ago in the very beginning of the Great Depression. After the economy collapsed in the fourth quarter of 1929 the US government needed to find a way to measure the magnitude of the recession. Fast forward to the late 1930s and the first official US national accounts were born from the efforts of one Simon Kuzents. This was the introduction of the modern concept of ‘gross domestic product’, a central concept of modern economic analysis that has changed relatively little in the intervening decades despite improvements in data collection and processing, expanded coverage across economic activities and more.1
While GDP can confirm the speed and magnitude of changes in economic activity, by virtue of being estimated on a quarterly basis, significant changes in economic activity (upward or downward) typically only appear three to four months post-fact. GDP certainly cannot measure up to a COVID-19 economic experience.
Even leading indicators—introduced first in 1938 by Wesley C. Mitchell and Arthur F. Burns, founders of business cycle research—with the benefit of monthly and even daily frequencies, as well as positive track records for predicting recessions, are ill equipped for COVID-19. Consider, for example, the US economic index published by the Conference Board; on March 19, it predicted economic expansion even though alternative data indicators already pointed to recession.
In this article, we discuss how leading indicators became lagging and show how alternative data can supplement traditional measures to track economic conditions.
Leading Indicators Rely Heavily on Lagging Data
Experts from the Conference Board notified readers in a March 19 press release that despite encouraging predictions represented by the index, “The economy may already be entering into a period of contraction”.2 This warning suggests that traditional leading indicators have been rendered null and void in the age of coronavirus.
Six out of 10 components of the Conference Board’s leading index—manufacturers’ new orders (consumer goods, capital goods, and the ISM new orders index), building permits, consumer expectations and average weekly hours in manufacturing—suggested stable economic expansion in January/February. Only two components clearly signaled recession – weekly unemployment claims and stock prices.
The Conference Board is not alone. Composite leading indices have failed across the board to reflect the extent of the economic downturn caused by the coronavirus pandemic. Even the OECD canceled the March 2020 release of its leading indices for OECD and other major developing countries. Instead, in early March the OECD released a note stating, “Composite Leading Index (CLI) sub-components for many countries are not yet able to capture the effects of the more widespread COVID-19 outbreak. Timely as they are, sub-components are not able to capture significant unpredictable changes in the macro-economy nor in expectations that may have occurred in the days and weeks following their collection”.3
There’s a fundamental incompatibility between reliance on monthly data that runs on a minimum one month lag when COVID-19 is transforming economies and lives on a daily basis. In February 2020, the economy was still growing, hence leading indices predicted growth for March even though the world economy was almost certainly already headed toward recession … if only we could measure it.
Making Leading Indices Leading Again
The week and a half from April 7 (when this article was written) to April 17 will provide additional color to the ‘leading’ nature of traditional economic measures. Several research institutions are scheduled to release their leading indices for the United States and other major economies. But let’s not lose sight of the fact that while these leading indices will begin to show signs of stress under the COVID-19 pandemic, this information is already a month late relative to reality on the ground.
Returning as always to the data, but this time looking beyond traditional economic data, new leading indicators are available. In the beginning of March 2020, when COVID-19 took hold in the United States, Knoema started publishing reports capturing the downturn in economic activity through the lens of alternative data. Here are just a few examples:
- In February, foot traffic in major US and international airports showed signs of contraction.
- Data on seated diners at restaurants revealed a significant decrease in visitors by early March.
- By March 15, data on mobile app downloads had already demonstrated changing behavior of US mobile app users.
- Data on web traffic in the first 14 days of March also pointed to contraction in tourism, retail, car rentals, and hotels.
Alternative data can offer insights into business activity throughout an economy, agnostic to location, sector, and means of delivery. Credit card transactions, web traffic, geolocation data from cell phones, satellite images and weather forecasts can all help track changes in our economic environment. Institutional investors have been using alternative data for years to predict corporate performance indicators (KPIs), especially sales and revenue.
Although statistical agencies are by and large only beginning to make use of alternative data, we believe that the COVID-19 crisis has the power to alter the world of traditional statistical measurements as did the Great Depression 90 years ago.
- The Art and Science of Measuring National Economic Performance. Knoema Insights. January, 2020. Link
- The Conference Board Leading Economic Index (LEI) for the U.S. Increased Slightly in February. The Conference Board. March19, 2020. Link
- Release of OECD Composite Leading Indicators Cancelled for March 2020. 09 March, 2020. OECD. Link