Automation, Worker Training, and Federal Tax Policy
May 27, 2022
Automation, Worker Training, and Federal Tax Policy
Technologies that partly or fully automate a variety of tasks
likely fill jobs created by automation requiring new skills
are being used with increasing frequency in a range of
and knowledge. The study also estimated that low-wage
industries and occupations. These technologies include
workers were four times more likely to be displaced by
robotics, machine learning, and other forms of artificial
automation in that decade than were highly paid workers.
intelligence (AI). This increasing use of automation has
fueled the concern that the substitution of machines for
Investment in Worker Training and
humans in a growing number of workplaces will result in
Education
massive job losses, especially for unskilled or low-skilled
Investment in worker training and education is a key source
workers. Some predict that if such a scenario were to arise,
of economic and productivity growth owing to its impact on
many displaced workers would face a bleak future marked
human capital. In theory, both employees and employers
by fewer job opportunities at lower wages, long-term
reap benefits from this investment. Employees gain new
earnings losses, and poor health.
skills and knowledge that they can use to obtain well-
paying and more interesting jobs. Employers obtain a more
The worker-displacing potential of automation has given
productive, skilled, loyal, and motivated workforce.
rise to a debate over what steps firms, governments,
postsecondary schools, and other entities should take, if
Yet there is a possible employer downside to such
any, to help displaced workers find well-paying jobs that
investment in the form of a market failure. In general, such
may or may not be linked to automation.
a failure refers to a condition that prevents or hinders
economically efficient outcomes. In the case of employer
This In Focus looks at how federal tax policy might be used
investments in worker education and training, a market
for that purpose. Specifically, it addresses the pros and cons
failure could arise when an employee transfers skills and
of possible new business tax incentives to encourage
knowledge he or she received from training from a former
employers to invest more in training their employees and to
employer to another employer. Such a transfer allows the
dissuade them from increasing automation investment. This
second employer to benefit from the enhanced human
overview does not address possible new tax incentives for
capital without bearing any of its cost. Economists regard
individuals to acquire on their own the skills and
these spillovers as a market failure because the inability of
knowledge they would need to find well-paying jobs.
employers to capture all returns to their training and
education investments is thought to lead them to invest too
Automation and Worker Displacement
little in worker training and education relative to its overall
The spread of workplace automation has raised many policy
economic benefits. Government subsidies might boost this
concerns in recent years. They include the jobs that are
investment to socially optimal levels.
likely to disappear, the jobs that will be created, the wage
effects of this churning automation’s impact on how and
Worker Training Investment and the
where jobs are performed, and actions that employers and
Federal Income Tax
federal, state, and local governments might take to facilitate
The federal income tax offers no targeted incentive for
the transition of displaced workers to new jobs.
employers to invest in worker training. Internal Revenue
Code (IRC) Section 162(a) permits businesses to deduct in
A number of studies have addressed these and other
full their “ordinary and necessary” expenses in the year
concerns. Of particular note is a 2020 paper by Daron
they are paid or incurred in calculating their taxable
Acemoglu and Pascual Restrepo on the impact of robots
income. Employee training expenses are among these
(narrowly defined) on U.S. employment and wages. The
expenses.
authors estimated that the addition of one robot for every
1,000 U.S. employees resulted in a slight decline in overall
Some limits apply to employees who pay for classes on
wages (0.42%) and the employment-to-population ratio (0.2
their own to improve their work skills and knowledge. In
percentage points, which meant the loss of about 400,000
general, they may deduct the cost of the classes only if they
jobs) from 1993 to 2007. They also found that the declines
are directly related to their current jobs.
varied by industry.
Arguably, the federal tax code contains an implicit tax
A broader perspective on the labor-saving potential of
incentive for firms to invest in worker training. This
automation came from a 2019 report by the McKinley
incentive is known as expensing, which allows an employer
Global Institute. The study predicted that 49.1 million U.S.
to deduct the full amount of qualified training costs in the
jobs might be lost because of automation from 2020 to
year they are incurred or paid. Expensing benefits
2030. But only one-third of those workers (14.9 million)
employers by lowering the cost of capital for qualifying
would have to find jobs completely unrelated to their
investments and boosting short-term cash flow—although
previous jobs; the other two-thirds (34.2 million) would
https://crsreports.congress.gov
Automation, Worker Training, and Federal Tax Policy
at the expense of reduced cash flow in the next several
to increase their worker training budgets, while managing
years.
the risk that firms would benefit from the credit for worker
training investments they would undertake in any event.
While human capital’s rate of depreciation is hard to
measure, it seems reasonable to assume that the economic
An incremental credit’s incentive effect would largely
value from human capital created by worker education and
depend on both its statutory rate and how the base amount
training is not always used up in one year, as expensing
was determined. For example, companion bills in the 117th
presupposes. To the extent there is a gap between the useful
Congress (H.R. 2984 and S. 1422) would establish a 20%
life of human capital and one-year life under expensing,
incremental credit for worker training expenses above a
expensing would constitute a tax subsidy for employer
base amount that would equal an employer’s average
investments in worker education and training.
annual such expenses in the three previous years. This
formula ties the base amount to a firm’s recent worker
In addition, there is a tax incentive for employers to directly
training investments but dilutes the credit’s incentive effect
and indirectly provide educational assistance to employees.
in the three years after an increase in this investment.
IRC Section 127 permits an employee to exclude up to
$5,250 a year in employer educational assistance that is
By contrast, a flat credit would apply to an employer’s
provided under a written program. The cap has remained
entire qualified training expenses in a tax year. As a result,
the same since 1986, and the program can include college-
a flat credit would be more generous to recipients than an
level courses and does not need to be job-related.
incremental credit with the same statutory rate. A flat credit
Employers can deduct those costs only if they meet the
generally may be easier to administer and comply with than
criteria set forth in IRC Section 162(a).
an incremental credit. The main argument against a flat
credit is that it would subsidize worker training investments
Automation Investment and the Federal
an employer would make in any event.
Income Tax
The federal tax code also provides no targeted incentive for
Other notable design considerations are whether a tax credit
businesses to invest in automation. But as with worker
would apply to training expenses for all employees or only
training, it implicitly encourages employer investment in
for low- or middle-wage employees, and the kinds of
automation through two expensing allowances. IRC Section
training programs that would qualify for the credit (e.g.,
168(k) permits companies to deduct the full cost of certain
industry apprenticeships or partnerships involving
assets (mostly software and equipment) in the year when
employers and educators).
they are placed in service. (This expensing option phases
out between 2023 and 2026.) IRC Section 179 allows a
Tax Incentives for Deterring Automation
company to expense up to $1.08 million in qualified assets
Investment
placed in service in 2022. This allowance phases out dollar
A more controversial approach to helping workers gain the
for dollar when the firm’s total investment in those assets
knowledge and skills they would need for well-paying jobs
exceeds $2.7 million, a limitation that confines the
at a time of increasing automation is to tax the use of
allowance’s benefits to small- and medium-sized firms.
automated technologies like robots. Proponents of such a
Those amounts are indexed for inflation.
tax claim that it would (1) protect current jobs by
dissuading employers from replacing humans with robots
Policy Issues
for specific tasks; (2) curb the use of automation that does
The question of how the federal government should respond
little to improve labor productivity; and (3) raise revenue
to job market dislocations associated with automation has
that could pay for local, state, or federal training and
generated numerous policy proposals. Some would increase
reskilling programs for displaced workers.
federal funding for job training tied to industry-based
apprenticeships. Others would create new tax incentives to
Critics of such a tax say that it would do more economic
encourage greater employer investment in worker training
harm than good. They cite evidence that firms adopting
for in-demand skills, and to discourage further use of
automated technologies hire more workers than comparable
automation. This section looks at proposed tax incentives.
nonautomated firms. In their view, taxing robots and other
forms of automation would have the undesirable effect of
Tax Incentives for Investment in Training
slowing domestic job and productivity growth. Then there
One idea that has gained support among some policymakers
is the difficult question of defining the automated
is a tax credit for increased employer spending on worker
technologies that would be taxed. Some say that the answer
training. Bills to adopt such a proposal have been
would be controversial, as some firms investing in
introduced in the 115th through 117th Congresses.
automation might pay more in taxes than others making
similar investments.
The efficacy and cost of such a credit hinges on its design
and policy goals. Several design elements are key to a
No bills to deter automation through the federal tax code
worker training credit’s incentive effect.
have been introduced in the 117th Congress.
One element concerns whether the credit should be
incremental or flat. An incremental credit would apply to an
employer’s qualified training costs in a tax year above a
Gary Guenther, Analyst in Public Finance
base amount only. As such, it would incentivize employers
IF12124
https://crsreports.congress.gov
Automation, Worker Training, and Federal Tax Policy
Disclaimer
This document was prepared by the Congressional Research Service (CRS). CRS serves as nonpartisan shared staff to
congressional committees and Members of Congress. It operates solely at the behest of and under the direction of Congress.
Information in a CRS Report should not be relied upon for purposes other than public understanding of information that has
been provided by CRS to Members of Congress in connection with CRS’s institutional role. CRS Reports, as a work of the
United States Government, are not subject to copyright protection in the United States. Any CRS Report may be
reproduced and distributed in its entirety without permission from CRS. However, as a CRS Report may include
copyrighted images or material from a third party, you may need to obtain the permission of the copyright holder if you
wish to copy or otherwise use copyrighted material.
https://crsreports.congress.gov | IF12124 · VERSION 1 · NEW