MNI INTERVIEW:Covid Distorts Productivity Data-OECD's Schreyer

Distorted statistics due to Covid are making it more difficult to assess productivity growth, but apparent sharp divergences between countries' performance during the crisis were exaggerated by methodological differences and should narrow in annual data, the chief statistician at the Organisation for Economic Cooperation and Development told MNI.

National statistics authorities that aimed to better capture the actual provision of services in areas like health and education, rather than inputs such as wages, are likely to have seen more volatile higher frequency data during lockdowns, when classrooms emptied and many workers were idle but paid under furlough schemes, Paul Schreyer said in an interview.

In the UK, the effects of the first furlough scheme meant that output per hour rose 1.0% in Q1 2020, while output per worker plunged 4.6%.

Source: Office for National Statistics, MNI Market News

Such effects complicate the task of measuring productivity, Schreyer noted, pointing to a key metric used by policymakers to judge whether economies are running too hot or too cold.

"You have to be really careful to see whether you are measuring what you want to measure … Hours paid could be very different under Covid from hours worked so what do you use to construct your labour productivity figures?" he said. "I would really approach the quarterly productivity figures with a lot of caution. Again, I think we have to wait for some of the annual data to come in to get a better picture of that."

In the case of education, the UK measures outputs based on teacher-pupil hours, while many other countries collated inputs based on teaching costs.

"Conceptually speaking what the UK has been doing is something that I would welcome," Schreyer said, but he added: "The price to pay in the short-run is that you face an acute volatility, especially if you look at quarterly figures."

As official data do not capture domestic activity, home schooling meant much education "moved out of the production numbers," he said.

JOBS DATA

Another divergence came with the treatment of furloughed employees, with statisticians in Europe treating them as employed and in the U.S. as unemployed. Prior to the unusual circumstances of Covid, this methodological difference was of little import, Schreyer noted.

Source: Reuters Datastream, MNI Market News

"Normally … whether you consider those people unemployed or employed is almost a matter of taste … It never really played out in international comparisons," he said, adding that statisticians should learn the lessons of the crisis to improve their procedures

"I would expect some more convergence as more solid data comes in," Schreyer said. While he cautioned that he had little hard evidence for the statement, he added that it was an empirically-based "gut instinct".