Economics & Growth | Fiscal Policy
- A new OECD working paper asks whether job losses during the pandemic were more likely in less productive firms, and whether more productive ones absorbed them.
- They were: the authors find a 6.8pp differential in employment growth between high and low revenue-per-worker firms from February 2020 to February 2021.
- Tech-savvy firms were more often classified as highly productive and more likely to increase employment during the pandemic.
During recessions, labour turnover increases as struggling firms attempt to reduce costs. In theory, these workers get reallocated to more productive firms that survive because they have more adaptable business models. This process of creative destruction, where less productive firms contract/fail and more productive ones expand, leads to a more efficient labour allocation and increased productivity.
During the pandemic, however, governments prioritised preservation over reallocation, introducing job-retention schemes that saved jobs and firms indiscriminately. And so, a new OECD working paper examines whether the drop in creative destruction has reduced worker reallocation, thereby restraining aggregate productivity growth. They find:
- Job reallocation and creative destruction fell once the pandemic began.
- Job losses were still significantly more likely in the least productive firms, while highly productive firms tended to expand.
- App-centric businesses were much more resilient to the pandemic’s adverse effects, reaffirming their importance for future productivity growth.
Productivity During Recessions
According to economist Joseph Schumpeter (1939), recessions accelerate productivity-enhancing reallocation. That is, the tendency for more productive firms to expand and less productive ones to contract (or exit). During this process, jobs are simultaneously created and destroyed, and in theory, workers move from the contracting to expanding firms.
This process of job creation and destruction is well documented. Its speed depends on a country’s market structure (e.g., younger and smaller firms typically account for a larger share of job creation and destruction), on government policies (e.g., business grants for new firms unable to access credit), and on firm-level productivity (e.g., more productive firms are more likely to expand).
Overall, in countries where reallocation occurs more often, aggregate productivity is generally higher. And if markets increasingly select the most productive firms during economic downturns, recessions can raise aggregate productivity. Indeed, recessionary episodes in the US from the 1940s to the early 2000s generally displayed such cleansing dynamics.
The Importance of Reallocation and Covid-19
Productivity growth in many OECD countries has slowed over the last two decades. Part of this is likely related to the fall in labour reallocation resulting from lower firm entry and tighter credit conditions following the financial crisis. In essence, jobs are being destroyed, but few high-quality alternatives are being created because of, for example, more difficult lending conditions.
The pandemic’s impact on productivity-enhancing reallocation is hard to tell. Consequently, so is whether aggregate productivity will rise post-Covid. Retention schemes and crisis policies reduced the reallocation rate, and those who lost their jobs were unlikely to find similar-quality ones at more productive firms within the same sector.
But an alternative view posits that the reallocation-productivity link remained, even if overall reallocation fell. The Covid shock required nimble and innovative business models to thrive, and – due to their superior management practices – these were likely to be highly productive firms. If these firms created jobs, labour would have reallocated to them.
Methodology and Data
A lack of suitable microdata, particularly measures of firm-level productivity, has limited timely analysis of the pandemic’s impact on productivity-enhancing reallocation. However, the authors use new microdata from Xero – a cloud accounting platform with over 2.7 million global subscribers – on business revenue, employment, hours worked, wages and industry.
They obtain data on Xero’s subscribers in the UK, Australia and New Zealand aggregated to a monthly frequency from 2018 to February 2021. Overall, the sample of businesses from Xero closely matches the characteristics of businesses from national statistics offices, proving a good representative dataset for their analysis (Chart 1).
Results – Job Reallocation
To start, the authors look at wages and employment changes within SMEs from March 2020. Across countries, average small business employment and hours worked significantly declined when the pandemic began. Furthermore, between February and May 2020, the share of expanding firms (i.e., those adding workers) declined relative to the same period in 2019, and the share of contracting firms increased (Chart 2). In other words, job reallocation (proxied by the sum of contracting and expanding shares, and inversely related to the static share) fell when the pandemic began in New Zealand and the UK.
Results – Productivity
Next, the authors estimate the change in firm-level employment for high-productivity (revenue per worker one standard deviation above industry mean) and low-productivity firms. They find that, over the year to February 2021, the difference in employment growth between the two was 6.8pp. In other words, reallocation remained productivity enhancing, despite the onset of the pandemic and crisis-phase policies that prioritised preservation over reallocation.
Reallocation was particularly productive over the initial months of the pandemic. On average across countries, productivity-enhancing reallocation accelerated over the first half of 2020, relative to the same period in 2019 (Chart 3). This is especially so in the UK, where the difference in employment growth between high- and low-productivity firms was noticeably higher from February to September 2020. Put differently, the most productive UK firms were significantly more resilient to the Covid-19 shock.
The same patterns are visible in hours worked: highly productive firms were less likely to reduce hours worked when the pandemic began. However, even the most productive UK firms still reduced hours by more than 20% between March and June 2020. Together, though, the results show a productivity-enhancing labour reallocation during the pandemic.
Results – Apps
The authors find a positive correlation between labour productivity and app usage. That is, the most productive firms were also more likely to use apps. Given the last results, it is unsurprising they also find that more tech-savvy firms saw 2.1pp higher employment growth from February 2020 to February 2021 than other firms. Also, firms that use apps more intensively were more likely to have increased hours during the pandemic.
Business and employment support programmes saved many firms during the pandemic. There were concerns this blanket approach would artificially preserve doomed businesses. The research shows that labour reallocation from low- to high-productivity firms continued, despite policy intervention. Tech-savvy firms drove this reallocation, supporting aggregate productivity growth. The paper also confirms an age-old belief remains true: markets increasingly select the most productive firms during economic downturns.
Andrews, Charlton and Moore (2021), COVID-19, productivity and reallocation: Timely evidence from three OECD countries, OECD Working Paper No 1676
Sam van de Schootbrugge is a Macro Research Analyst at Macro Hive, currently completing his PhD in international finance. He has a master’s degree in economic research from the University of Cambridge and has worked in research roles for over 3 years in both the public and private sector.