Pandemic hits front-line tangible and intangible employment

December is following a bad November for the labor market. According to this morning’s employment data from the BLS, employment dropped by 140,000 in December. The biggest hit was in Accommodation & Food Services – down by almost 400,000. The second largest drop was in Arts, Entertainment & Recreation where employment decline by 102,600. Both of these industries are directly hit by the pandemic. The bright spots were increases of 191,000 in Trade, Transportation & Utilities and 157,200 in Professional & Business Services.

The most important part of the story, however, may be the across-the-board slowdown in hiring. While the four sectors discussed above had the biggest changes, the rest of the economy was essentially flat. Employment in some tangible-producing industries, such as Construction & Mining, Manufacturing, and Tangible business services, increased slightly; employment in others, such as Repair & Maintenance, Personal & Laundry Services, Telecommunications, and Tangible educational & health services declined. Employment in the intangible-producing sectors showed the same mixed pattern. Employment in Membership Associations, Education & Health Services, and Government declined somewhat. Employment in Information and in Financial Activities increased.

For more on the categories, see my explanation of the methodology in an earlier posting.

New data on adoption advanced information technologies

There is an exciting new survey out on firms’ adoption of advanced information technology. The survey is part of the Census Bureau’s 2018 Annual Business Survey and is described in a new NBER paper  Advanced Technologies Adoption and Use by U.S. Firms: Evidence from The Annual Business Survey.

The survey finds “that digitization is quite widespread, as is some use of cloud computing. In contrast, advanced technology adoption is rare and generally skewed towards larger and older firms. Adoption patterns are consistent with a hierarchy of increasing technological sophistication, in which most firms that adopt AI or other advanced business technologies also use the other, more widely diffused technologies.”

The implication is that these firms using these technologies will continue to be the driver of productivity. As the report states, “while few firms are at the technology frontier, they tend to be large so technology exposure of the average worker is significantly higher.”

Use of business technologies by industry

For me, the most interesting finding are about the use of advanced business technologies. One of the more interesting finding concerns how different industries use different business technologies. The findings are both expected (in hindsight) and illuminating. (see table 16.)

As would be expected, warehousing and wholesalers are the top users of RFID technologies, along with motor vehicle parts manufacturers. Manufacturing industries are the biggest user of robotics, machine learning and machine vision. Surprisingly, software publishers are a top adopter of machine vision as well. Software publishers are also top users of augmented reality, and natural language processing. Motion picture and video industries, and specialized design services are the other heavy users of augmented reality.

The more enlightening finding of the survey concerned uses in industries not generally thought of as heavy into advanced technologies. Augmented guided vehicles are most used in the subindustry groups of Support Activities for Crop Production, Farm Product Raw Material Merchant Wholesalers, and Highway, Street and Bridge Construction. The usage rates in these industries are well about the mean usage rate for all industries. This indicates the existence of niche markets for autonomous/semi-autonomous vehicles.

Voice recognition technologies are most heavily used in the health-related industries of Medical and Diagnostic Laboratories, Offices of Physicians, and Outpatient Care Centers. Again, not where one might expect to find the greatest use of this technology.

Most intriguing is the use of touchscreens. The top user is Nursing Care Facilities (Skilled Nursing Facilities) followed by Beverage Manufacturing and Other Amusement and Recreation Industries. Given the heterogenous mix of industry users, it would be interesting to see how the technology is being used in the different industries. I can see touchscreens begun used in nursing. And other recreation industries might reflect the seemingly ubiquitous use of touchscreens in professional sports. But beverage manufacturing seems like an anomaly.

Looking at the data in a different way, the survey found that touchscreens technology is the most used business technology in every sector of the economy except manufacturing where it is ranked third. Machine learning and voice recognition are the technologies that show up as the second most used technology across all sectors. (see table 15)

I should note that none of these advanced business technologies should be considered widespread. As I mentioned earlier, while digitization is wide spread, the utilization of advanced business technology is still very limited (see fig 9). The most adopted technology is touchscreens with only a 7% overall use rate (in current use or being tested for future use). All the rest have overall use rates between 3.6% and 1% (see table 13).

The data provides interesting indications of where the technology is being used and which niche markets may provide a foundation for future adoption. More detailed research is needed, based on this data, as to what might be the opportunities and challenges to the spread of the technology across industry sectors.

Clustering of business technologies and technological hierarchies

One of the helpful insights in the paper is the discussion of technological hierarchies and complementarities. The data shows a clear hierarchy of technological sophistication. Logically, digitization is the first step. The Sankey diagram in fig. 9a shows that adoption of advanced business technologies is predicated on digitization and the usage of cloud computing. And, as the example machine learning shows, the adoption of the machine learning is helped by higher levels of digitization and usage of cloud computing (see the Sankey diagram in fig. 9b).

The data also shows that adoption of business technologies are tied together. For example, adoption of machine learning is correlated with machine vision (table 14). Interestingly, automated vehicles is correlated with augmented reality and with RFID. As the report notes, “These correlations suggest that certain technologies may need to be adopted in tandem to fully reap the benefits of the technology.”

Some remaining questions

As mentioned earlier, the paper reports at the data on size and age indicates greater adoption in general by larger, established firms. However, this overall finding leaves me with questions. First, it is unclear as to the differences in utilization among larger and older firms. Why does one incumbent adopt while another does not?

I would also like to see more analysis of the different industries use different advanced business technologies. Specifically, is the difference in utilization by size & age consistent across industries? Are there industries and business technologies where younger, smaller firms lead in adoption (rather than the overall finding for large, established companies)? For example, while I would expect the use of robotics in manufacturing to tend toward larger, established firms, is the same true for use of touchscreens in manufacturing? Or touchscreens in nursing care facilities?

The paper hints at this when it states “there are important differences in the diffusion and intensive use rates across sectors. The analysis of the connection between the prevalence of different technologies and firm life-cycle indicators (firm size and age) reveals that technology adoption and use is not always monotonically related to these indicators.”

Business technologies, productivity and public policy

The survey opens up a number of questions about the impact of business technologies. From a technology and economy policy perspective, the goal is to facilitate the adoption of technologies that increase productivity. Adoption patterns aside, what does the data tell us about use of business technologies and productivity? Are there business technologies that increase productivity across the board? If so, are the barriers to adoption of those technologies the same across industries? If so, the broad-based policies might be the most appropriate (such as an expansion of the MEP centers). If barriers are either technology or industry specific, then more targeted policies might be called for.

The discussion on technological hierarchies mentioned earlier also raises questions about public policy. The example of machine learning (fig. 9b) shows that the leap from digitization and use of the cloud to business technologies is still daunting. Thus, from a public policy perspective, this points to the need for both facilitating more intensive digitization and use of cloud computing and for strategies to help move beyond digitization and cloud computing.

Conclusion

The discussion above is but a small discussion of the findings of the report. And the paper itself only begins to scratch the surface of the richness of the data. As the report notes, “It is our hope that this paper serves as an impetus for further research using this new data set to help answer these important questions.”

I second that call. There is a lot more here waiting to be discovered.