What does Mona Lisa do? At first glance, the subject of the most famous painting in the world seems to be smiling. Look again and her smile fades. When it reappears next, it’s a different kind of smile. Leonardo da Vinci achieved this ambiguous effect with the use of sfumato, where he blurred the lines around the Mona Lisa’s face. No matter how many times you look, you’re not sure what’s going on.
The post-pandemic economy is like the Mona Lisa. Every time you look you see something different. After the chaos in the banking sector, many analysts are now convinced that the global economy is headed for a “brutal” recession. Few seem to be expecting a “no-landing” scenario, in which the economy is undisturbed by rising interest rates – a view fashionable just weeks ago, and which itself supplanted the common view late last year that a mild recession was certain.
In short: forecasting has rarely been more difficult. Over the past year, the range of analysts’ expectations for quarterly U.S. GDP growth has been twice as wide as in 2019. The word “uncertainty” appears more than 60 times in the IMF’s latest global economic outlook, about twice as many as in April and October 2022. At the last monetary policy meeting of the European Central Bank last month, Christine Lagarde, its president, was candid about the role of her institution. “It is not possible to determine at this stage what the way forward will be,” she said.
Official statisticians struggle to understand the table. Of course, they update their estimates of everything from GDP to employment as more data comes in. But something profound has changed. Eurozone GDP revisions are four times larger than normal. In March, the UK Statistics Office released significant revisions. The statement showed that real business investment was in line with its pre-pandemic level, not 8% below as once believed. Australian statisticians last month more than halved their estimate of productivity growth in the third quarter of 2022. That year, the US Bureau of Labor Statistics (bls) released revisions to its nonfarm payrolls estimate (unadjusted for seasonality) of 59,000 per month between the first and third estimates, down from 40,000 in 2019.
What is going on? Maybe the world is just more unstable. Last year, Europe experienced its biggest war in seven decades, supply chain upheavals, an energy crisis and a bank run. The rest of the rich world has been only a little more stable.
Yet deeper changes are also at play. The first concerns the disruptions linked to covid-19. The world went from collapsing to skyrocketing as lockdowns came and went. This upset the “seasonal adjustments” common to most economic numbers. In February, the bls changed the factors it applies to inflation, which makes the interpretation of monthly rates much more difficult. Annualized core inflation in the last quarter of 2022 “increased” from 3.1% to 4.3%. It is also more difficult than usual to understand eurozone inflation. Kamil Kovar of Moody’s Analytics, a consultancy, notes that depending on how the seasonal adjustment is done, monthly core inflation in March was as low as 0.2% or as high as 0.4%.
The second change concerns sample sizes. The pandemic has accelerated a trend in which a growing share of people are not responding to official surveys. In America, the response rate to the survey used to estimate vacancies fell from nearly 60% just before the pandemic to around 30%. When covid hit, the UK Labor Force Survey response rate dropped by around half. During the lockdowns, some businesses closed. People have lost the habit of filling out questionnaires. Distrust of government may also have increased, leaving people reluctant to help statisticians.
Bleached
Declining response rates likely increase data volatility. They can also lead to biases. People who have stopped taking surveys appear to be less successful than those who continue to do so, which misleadingly inflates earnings. Jonathan Rothbaum of the Census Bureau suggests that the actual median household income growth in America from 2019 to 2020 was 4.1%, not 6.8% as originally reported, after appropriate adjustments for nonresponse. Since 2020, non-response has continued to drive up income statistics by around 2%. A report by Omair Sharif of Inflation Insights, a consultancy, suggests that the correction of “nonresponse bias” may also have contributed to recent large revisions to US earnings data.
The third reason for confusion stems from the disparity between ‘hard’ and ‘soft’ data – objective measures such as the level of unemployment and subjective measures such as people’s future expectations. Normally, the two types move in sync. For now, they are distant from each other. “Soft” measures seem recessive. “Hard” measurements indicate decent expansion. The divergence may reflect people’s discontent with inflation. Prices in the rich world continue to rise 9% year-over-year.
Investors and statisticians will better understand the global economy during periods of volatility and inflation. As the effects of the pandemic fade, the distortions of seasonal adjustments will also fade. Economists have already made progress in incorporating alternative data into forecasts, helping to overcome the problem of declining responses. But that’s hardly reassuring for governments and businesses having to make decisions right now, or for people just trying to keep up with the news. Don’t be surprised if the global economy remains sfumata for some time to come.
For more expert analysis of the biggest stories in economics, finance and markets, sign up for Money Talks, our weekly subscriber-only newsletter.
© 2023, The Economist Newspaper Limited. All rights reserved. From The Economist, published under licence. Original content can be found at www.economist.com
Enjoy unlimited digital access with HT Premium
Subscribe now to continue reading
