At least some of us were productivity optimists in the earlier days of pandemic recovery. There was some evidence of a productivity uptick in the period from the beginning of the pandemic, continuing to the fourth quarter of 2021. While productivity growth averaged just 1.0 percent from the fourth quarter of 2009 to the fourth quarter of 2019, growth averaged 2.3 percent in the two years from the fourth quarter of 2019 to the fourth quarter of 2021.
Anyone who follows productivity data closely knows that the data are erratic, and even two good years doesn’t amount to much. No one should assume we’re on a faster growth path based on an uptick over a relatively short time period.
However, there was also some anecdotal evidence that we might be a faster growth trajectory. The most obvious story had to with the increased use of Zoom and other technologies allowing for teleconferencing. Insofar as business can be conducted without people traveling across town or across the country, there is a gain in productivity. There were many other stories of businesses being forced to adopt new and often more efficient ways of operating to get around the disruptions of the pandemic.
That was an optimistic story, but the picture looks much worse after the last two quarters. Productivity fell at a 7.4 percent annual rate in the first quarter of 2022 and a 4.6 percent rate in the second quarter. These drops totally offset the positive picture of 2020 and 2021. With the more recent data, productivity growth since the fourth quarter of 2019 now averages just 0.6 percent.
While this might mean that the productivity optimists should throw in the towel, there still is some reason for hoping that the data from the last two quarters are an anomaly and we could be on a faster growth path. There are three directions in which the productivity optimists’ argument would go:
+ Data errors gave us the bad numbers for the first two quarters;
+ Temporary disruptions from the pandemic slowed productivity growth in the first half of 2022;
+ There are important gains in productivity not reflected in the official data.
I’ll lay out each of these arguments briefly below.
Gross Domestic Product and Gross Domestic Income
The extraordinary job growth in the first half of 2022 is hard to reconcile with the negative GDP growth reported for this period. Many of us have noted that Gross Domestic Income (GDI) was actually reported as growing at a healthy 1.8 percent annual rate in the first quarter. (We don’t have GDI data for the second quarter yet.)
In principle, GDP and GDI should be the same. GDI is simply measuring GDP based on the income received. While the two numbers are never exactly the same, the gap in the first quarter was especially large. In fact, GDI growth has been outpacing GDP growth ever since the start of the pandemic. While real GDP has grown a total of 2.7 percent between the fourth quarter of 2019 and the first quarter of 2022, real GDI has grown by 6.1 percent. That translates into a 2.7 percent annual rate of growth since the start of the pandemic, compared to a 1.2 percent rate for GDP.
If the GDI measure is in fact closer to the actual rate of the economy’s growth over this period, then productivity growth would be approximately 1.4 percentage points faster than productivity calculated based on the GDP number. Instead of the 1.1 percent rate now reported for the period between the fourth quarter of 2019 and the first quarter of 2022, the growth rate would be close to 2.5 percent. That would still be consistent with the productivity speed-up story.
Rather than just taking GDI growth or GDP growth, it is common for economists to look at the average for the two measures. The growth rate for the average of GDI and GDP between the fourth quarter of 2019 and the first quarter of 2021 was 1.9 percent. Using this as the basis for measuring productivity growth would give us an average growth rate of 1.7 percent for the nine quarters since the start of the pandemic. That is not clear evidence of a speedup in productivity, but it is notably better than the growth rate for the prior decade.
As we get further revisions to growth data over the next year, the picture on productivity growth since the start of the pandemic will become clearer. But it is worth noting that the data at this point are not conclusive. If the GDI growth data prove to be more accurate then we will still have evidence of a productivity uptick since the start of the pandemic.
Temporary Pandemic Related Disruptions
As we know, the pandemic gave us supply chain disruptions in many major sectors of the economy. At the level of consumers, this meant shortages and higher prices for a wide variety of products.
However, these supply chain disruptions also hit businesses. They had to deal with unanticipated delays in the shipping of many products, and in some cases were forced to go without some items altogether. It would be amazing if these disruptions did not lead to slower productivity growth.
To see this story, we can look at an ad hoc measure of productivity growth in the construction industry. The combined categories in the National Income and Product Accounts of non-residential construction and residential construction (Table 1.1.6, Lines 10 and 13) fell 6.8 percent between the fourth quarter of 2019 and the second quarter of 2022. By contrast, the Bureau of Labor Statistics (BLS) index of aggregate hours for the construction sector rose 1.0 percent over this period. This would imply roughly a 7.8 percent decline in productivity over this ten-quarter period.
It doesn’t seem plausible that either construction technology or the quality of labor in the industry could have deteriorated so much in such a short period of time. The more obvious explanation for a decline in productivity in construction is that many workers were effectively wasting their time waiting for parts or materials that were needed for them to do their jobs. It would be reasonable to expect that when we have gotten through the supply chain disruptions associated with the pandemic, productivity in construction will return at least to its pre-pandemic level.
Insofar as this sort of story describes the work situation in other industries, it would mean that supply chain disruptions may have been a drag on productivity growth throughout the economy. It is also worth noting that many businesses had to take pandemic related measures, such as screening people who entered restaurants and stores, or frequently cleaning and sanitizing facilities, that would have also been a drag on productivity growth. As businesses stop doing these pandemic related safety measures, the drag on productivity growth will be reversed.
In short, we should have expected that the pandemic would have created a major drag on productivity growth, both through its impact on supply chains, and by forcing businesses to engage in pandemic-related safety practices that would not ordinarily be required. This means that even if businesses had sustained trend rates of productivity improvement in the way they conducted business, reported productivity would show a falloff due to the effects of the pandemic. It also would mean that we should see above trend productivity growth as the impact of the pandemic fades.
Unmeasured Gains in Productivity and Living Standards
As I argued at the start of the pandemic, changes in practices due to the pandemic are likely to lead to improvements in living standards that are not picked up in our measures of GDP and productivity. Many of these are related to the increased ability of people to work from home.
In the most recent jobs report, more than 11 million people still reported that they were working from home due to the pandemic. This does not include all the people who are still working from home, but had not previously, as a result of changes in their work situation independent of the current state of the pandemic. Clearly, a large segment of the workforce is now working from home, full-time or part-time, who had not previously had this option.
There are two ways this leads to improvements in living standards. The first is simply the time saved on commuting. The average time spent each day commuting in 2019, for those who commuted to work, was 55 minutes. This means that a person who can now entirely work from home, as opposed to commuting five days a week, would save 225 minutes commuting a week. This comes to 3 hours and 45 minutes or more than 10 percent of the length of the average workweek.
To put this another way, the length of the average workweek is currently 34.6 hours. If we think of the worker’s pay as being for both their time at work, and their time getting to and from work, saving 3 hours and 45 minutes commuting each week, amounts to a 10 percent real increase in their hourly pay. This is a big deal.
Of course, many of the people who now work from home as a result of changes induced by the pandemic, only work two or three days a week at home. But this still implies substantial savings in time spent commuting, which are presumably are of considerable value to the workers who now have this option.
In addition to time, there are also expenses associated with commuting. If people drive, in addition to gas, they also will need more maintenance for their car and have higher insurance costs due to their commutes. In the case of public transportation, they will have daily costs associated with train and bus fares. People also need to buy and clean business clothes for working in an office that they would not need for working at home.
It is not easy to get good data on savings on driving as a result of less commuting, but real spending on public ground transportation, which would include commuter trains and buses, fell from $61.5 billion in the fourth quarter of 2019 to $48.7 billion in the second quarter of 2022 (NIPA Table 2.4.5U, Line 198). This is a bit less than 0.1 percent of consumption spending.
Real spending on personal care services, which includes items like hair salons and dry cleaning, fell from $167.6 billion in the fourth quarter of 2019 to $146.8 billion in the second quarter of 2022 (NIPA Table 2.4.5U, Line 306). The savings of $20.8 billion would be a bit more than 0.1 percent of total spending.
These savings on work-related expenses do not appear in the data as productivity gains. In fact, most immediately they appear as lower GDP, as we have less consumption spending. Although if people spend this money in other areas, GDP will not be affected.
The savings from less commuting are offset to some extent by costs associated with working from home, such as the need for additional space for an office and possibly spending more on Internet services. But the largest monetary savings from less commuting are likely those associated with less driving. Also, the increased ability to work from home has allowed millions of families to move to locations that they find more desirable, either due to physical amenities or to be closer to family or friends. This would of course not be picked up in GDP data.
Other changes from the pandemic are also likely to have enduring benefits. One obvious one is the increased use of telemedicine. I am not aware of good data on this, but during the pandemic, many doctors allowed for patients to have consultations through Zoom, rather than having to travel to their office. For patients suffering from health problems, avoiding a trip that could be time-consuming and painful, has to be of substantial value.
This is another gain that would not appear in GDP or be recorded as an increase in productivity. Insofar as they were increased opportunities for remote services in other areas, there could be substantial benefits that are not showing up in our data. It is likely that such gains will grow through time as service providers find new and better ways to take advantage of technology that allows services to provided over the Internet. This process was already going on before the pandemic, but it is likely that the pandemic has hastened the pace.
The Jury is Still Out on a Pandemic Productivity Speedup
The last two quarters of terrible productivity data have undermined the most obvious basis for believing that the pandemic was leading to faster productivity growth. However, there are reasons for questions these data, and believing that revised data may look substantially better. It is also clear that disruptions directly associated with the pandemic have lowered productivity over the last two and a half years. When we overcome these disruptions, productivity should rebound in the affected areas. And, there are gains associated with a change in patterns of work and consumption that are not picked up in productivity data.
For these reasons, we can still have some hope that we are on a faster productivity path than we were before the pandemic. However, we will need to see more evidence to believe this is actually the case.
 The gap between GDI growth and GDP growth cannot be exactly added to productivity growth, since our standard measure of productivity is for the non-farm business sector. That excludes some areas of GDP, most notably the government sector.
 This is a very crude productivity measure. On the output side, the residential construction data includes some services related to mortgage issuance, which would not be produced by construction workers. The BLS measure of hours only counts payroll employment, excluding self-employed and likely also many workers who might work off the books.
This first appeared on Dean Baker’s Beat the Press blog.