The Economics of Artificial Intelligence
Page 46
cult or too costly to hire a high- skill worker;
furthermore, and perhaps more important, the low- occupation employee
is expected to stay longer in the fi rm than higher- skill employees, which in
turn encourages the fi rm to invest more in trust- building and fi rm- specifi c
human capital and knowledge. Overall, such low- occupation employees can
make a big diff erence to the fi rm’s performance.
This alternative view of AI and fi rms is consistent with the work of theo-
rists of the fi rm such as Luis Garicano. Thus in Garicano (2000) down-
stream, low- occupation employees are consistently facing new problems;
among these new problems they sort out are those they can solve them-
selves (the easier problems) and the more diffi
cult questions they pass on
to upstream—higher- skill—employees in the fi rm’s hierarchy. Presum-
ably, the more innovative or more AI- intensive the fi rm is, the harder it
is to solve the more diffi
cult questions, and therefore the more valuable
the time of upstream high- occupation employees becomes; this in turn
makes it all the more important to employ downstream, low- occupation
employees with higher ability to make sure that less problems will be passed
on to the upstream, high- occupation employees within the fi rm so that
these high- occupation employees will have more free time to concentrate
on solving the most diffi
cult tasks. Another interpretation of the higher
complementarity between low- occupation and high- occupation employees
in more innovative (or more AI- intensive) fi rms, is that the potential loss
from unreliable low- occupation employees is bigger in such fi rms: hence
the need to select out those low- occupation employees that are not reliable.
268 Philippe Aghion, Benjamin F. Jones, and Charles I. Jones This higher complementarity between low- occupation workers and other
production factors in more innovative (or more AI- intensive) fi rms in turn
increases the bargaining power of low- occupation workers within the fi rm
(it increases their Shapley Value if we follow Stole and Zwiebel [1996]).
This in turn explains the higher payoff for low- occupation workers. It also
predicts that job turnover should be lower (tenure should be higher) among
low- occupation workers who work for more innovative (more AI- intensive)
fi rms than for low- occupation workers who work for less innovative fi rms,
whereas the turnover diff erence should be less between high- occupation
workers employed by these two types of fi rms. This additional prediction is
also confronted to the data in Aghion et al. (2017).
Note that so far R&D investment has been used as the measure of the
fi rm’s innovativeness or frontierness. We would like to test the same predic-
tions, but using explicit measures of AI intensity as the RHS variable in the
regressions (investment in robots, reliance on digital platforms). Artifi cial
intelligence and fi rm organizational form: recent empirical studies (e.g., see
Bloom et al. 2014) have shown that the IT revolution has led fi rms to elimi-
nate middle- range jobs and move toward fl atter organizational structure.
The development of AI should reinforce that trend, while perhaps also
reducing the ratio to low- occupation to high- occupation jobs within fi rms
as we argued above.
A potentially helpful framework to think about fi rms’ organizational
forms is Aghion and Tirole (1997). There, a principal can decide whether or
not to delegate authority to a downstream agent. She can delegate author-
ity in two ways: (a) by formally allocating control rights to the agent (in
that case we say that the principal delegates formal authority to the agent);
or (b) informally through the design of the organization, for example, by
increasing the span of control or by engaging in multiple activities: these
devices enable the principal to commit to leave initiative to the agent (in
that case we say that the principal delegates real authority to the agent).
And agents’ initiative particularly matters if the fi rm needs to be innova-
tive, which is particularly the case for more frontier fi rms in their sectors.
Whether she decides to delegate formal or only real authority to her agent,
the principal faces the following trade- off : more delegation of authority to
the agent induces the agent to take more initiative; on the other hand, this
implies that the principal will lose some control over the fi rm, and there-
fore face the possibility that suboptimal decisions (from her viewpoint) be
taken more often. Which of these two counteracting eff ects of delegation
dominates, will in turn depend upon the degree of congruence between the
principal’s and the agent’s preference, but also about the principal’s ability
to reverse suboptimal decisions.
How should the introduction of AI aff ect this trade- off between loss of
control and initiative? To the extent that AI makes it easier for the princi-
pal to monitor the agent, more delegation of authority will be required in
Artifi cial Intelligence and Economic Growth 269
order to still elicit initiative from the agent. The incentive to delegate more
authority to downstream agents, will also be enhanced by the fact that with
AI, suboptimal decision- making by downstream agents can be more easily
corrected and reversed: in other words, AI should reduce the loss of control
involved in delegating authority downstream. A third reason for why AI may
encourage decentralization in decision- making has to do with coordination
costs: namely, it may be costly for the principal to delegate decision- making
to downstream units if this prevents these units from coordinating within the
fi rm (see Hart and Holmstrom 2010). But here again, AI may help overcome
this problem by reducing the monitoring costs between the principal and
its multiple downstream units, and thereby induce more decentralization
of authority.
More delegation of authority in turn can be achieved through various
means: in particular, by eliminating intermediate layers in the fi rm’s hier-
archy, by turning downstream units into profi t centers or fully independent
fi rms, or through horizontal integration that will commit the principal to
spending time on other activities. Overall, one can imagine that the develop-
ment of AI in more frontier sectors should lead to larger and more horizon-
tally integrated fi rms, to fl atter fi rms with more profi t centers, which out-
source an increasing number of tasks to independent self- employed agents.
The increased reliance on self- employed independent agents will in turn
be facilitated by the fact that, as well explained by Tirole (2017), AI helps
agents to quickly develop individual reputations. This brings us to the third
aspect of AI and organizations on self- employment. Artifi cial intelligence
and self- employment: as stressed above, AI favors the development of self-
employment for at least two reasons: fi rst, it may induce AI intensive fi rms
to outsource tasks, starting with low- occupation tasks; second, it makes it
easier for independent agents to dev
elop individual reputations. Does that
imply that AI should result in the end of large integrated fi rms with individu-
als only interacting with each other through platforms? And which agents
are more likely to become self- employed?
On the fi rst question: Tirole (2017) provides at least two reasons for why
fi rms should survive the introduction of AI. First, some activities involve
large sunk costs and/or large fi xed costs that cannot be borne by a single indi-
vidual. Second, some activities involve a level of risk- taking that also may
not be borne by one single agent. To this we should add the transaction cost
argument that vertical integration facilitates relation- specifi c investments in
situations of contractual incompleteness: Can we truly imagine that AI will
by itself fully overcome contractual incompleteness?
On the second question: our above discussion suggests that low- skill ac-
tivities involving limited risk and for which AI helps develop individual rep-
utations (hotel or transport services, health assistance to the elder and/or
handicapped, catering services, house cleaning) are primary candidates for
increasingly becoming self- employment jobs as AI diff uses in the economy.
270 Philippe Aghion, Benjamin F. Jones, and Charles I. Jones And indeed recent studies by Saez (2010), Chetty et al. (2011), and Kleven
and Waseem (2013) point to low- income individuals being more respon-
sive to tax or regulatory changes aimed at facilitating self- employment.
Natural extensions of these studies would be to explore the extent to which
such regulatory changes have had more impact in sectors with higher AI
penetration.
The interplay between AI and self- employment also involves potentially
interesting dynamic aspects. Thus it might be worth looking at whether
self- employment helps individuals accumulate human capital (or at least
protects them against the risk of human capital depreciation following the
loss of a formal job), and the more so in sectors with higher AI penetra-
tion. Also interesting would be to look at how the interplay between self-
employment and AI is itself aff ected by government policies and institu-
tions, and here we have primarily in mind education policy and social or
income insurance for the self- employed. How do these policies aff ect the
future performance of currently self- employed individuals, and are they
at all complemented by the introduction of AI? In particular, do currently
self- employed individuals move back to working for larger fi rms, and how
does the probability of moving back to a regular employment vary with
AI, government policy, and the interplay between the two? Presumably, a
more performing basic education system and a more comprehensive social
insurance system should both encourage self- employed individuals to bet-
ter take advantage of AI opportunities and support to accumulate skills
and reputation and thereby improve their future career prospects. On the
other hand, some may argue that AI will have a discouraging eff ect on
self- employed individuals, if it lowers their prospects of ever reintegrating
a regular fi rm in the future, as more AI- intensive fi rms reduce their demand
for low- occupation workers.
9.6 Evidence on Capital Shares and Automation to Date
Models that conceptualize AI as a force of increasing automation suggest
that an upswing in automation may be seen in the factor payments going to
capital—the capital share. In recent years, the rise in the capital share in the
United States and around the world has been a central topic of research.
For example, see Karabarbounis and Neiman (2013), Elsby, Hobijn, and
S¸ahin (2013), and Kehrig and Vincent (2017). In this section, we explore this
evidence, fi rst for industries within the United States, second for the motor
vehicles industry in the United States and Europe, and fi nally by looking
at how changes in capital shares over time correlate with the adoption of
robots.
Figure 9.6 reports capital shares by industry from the US KLEMS data of
Jorgenson, Ho, and Samuels (forthcoming); shares are smoothed using an
HP fi lter with smoothing parameter 400 to focus on the medium- to long-
Artifi cial Intelligence and Economic Growth 271
run trends. It is well- known that the aggregate capital share has increased
since at least the year 2000 in the US economy. Figure 9.6 shows that this
aggregate trend holds up across a large number of sectors, including agricul-
ture, construction, chemicals, computer equipment manufacturing, motor
vehicles, publishing, telecommunications, and wholesale and retail trade.
The main place where one does not see this trend is in services, including
education, government, and health. In those sectors, the capital share is
relatively stable or perhaps increasing slightly since 1990. But the big trend
one sees in these data from services is a large downward trend between 1950
and 1980. It would be interesting to know more about what accounts for
this trend.
While the facts are broadly consistent with automation (or an increase
in automation), it is also clear that capital and labor shares involve many
other economic forces as well. For example, Autor et al. (2017) suggest that a
composition eff ect involving a shift toward superstar fi rms with high capital
shares underlies the industry trends. That paper and Barkai (2017) propose
that a rise in industry concentration and markups may underlie some of
the increases in the capital share. Changes in unionization over time may
be another contributing factor to the dynamics of factor shares. This is all
to say that a much more careful analysis of factor shares and automation is
required before any conclusions can be drawn.
Keeping that important caveat in mind, fi gure 9.7 shows evidence on the
capital share in the manufacturing of transportation equipment for the
United States and several European countries. As Acemoglu and Restrepo
(2017) note (more on this below), the motor vehicles industry is by far the
industry that has invested most heavily in industrial robots during the past
two decades, so this industry is particularly interesting from the standpoint
of automation.
The capital share in transportation equipment (including motor vehicles,
but also aircraft and shipbuilding) shows a large increase in the United
States, France, Germany, and Spain in recent decades. Interestingly, Italy
and the United Kingdom exhibit declines in this capital share since 1995.
The absolute level diff erences in the capital share for transportation equip-
ment in 2014 are also interesting, ranging from a high of more than 50 per-
cent in the United States to a low of around 20 percent in recent years in
the United Kingdom. Clearly it would be valuable to better understand
these large diff erences in levels and trends. Automation is likely only a part
of the story.
Acemoglu and Restrepo (2017) use data from the International Federation
of Robots to study the impact of the adoption of industrial robots on the
US labor market. At the industry level, this data is available for the decade
2004 to 2014. Figure 9.8 shows data on the change in capital share by indus-
try versus the change in the use of industrial robots.
Two main facts stand out from the fi gure. First, as noted earlier, the motor
Fig. 9.6 US capital shares by industry
Source: The graph reports capital shares by industry from the U.S. KLEMS data of Jorgenson, Ho, and Samuels (2017).
Note: Shares are smoothed using an HP fi lter with smoothing parameter 400.
Fig. 9.7 The capital share for transportation equipment
Sources: Data for the European countries are from the EU- KLEMS project (http:// www
.euklems .net/ ) for the “transportation equipment” sector, which includes motor vehicles, but also aerospace and shipbuilding; see Jägger (2016). US data are from Jorgenson, Ho, and Samuels (2017) for motor vehicles.
Note: Shares are smoothed using an HP fi lter with smoothing parameter 400.
Fig. 9.8 Capital shares and robots, 2004– 2014
Sources: The graph plots the change in the capital share from Jorgenson, Ho, and Samuels (2017) against the change in the stock of robots relative to value added using the robots data from Acemoglu and Restrepo (2017).
274 Philippe Aghion, Benjamin F. Jones, and Charles I. Jones vehicles industry is by far the largest adopter of industrial robots. For example, more than 56 percent of new industrial robots purchased in 2014
were installed in the motor vehicles industry, the next highest share was
under 12 percent in computers and electronic products.
Second, there is little correlation between automation as measured by
robots and the change in the capital share between 2004 and 2014. The
overall level of industrial robot penetration is relatively small, and as we
discussed earlier, other forces including changes in market power, unioniza-
tion, and composition eff ects are moving capital shares around in a way that
makes it hard for a simple data plot to disentangle.
Graetz and Michaels (2017) conduct a more formal econometric study
using the EU- KLEMS data and the International Federation of Robotics
data from 1993 until 2007, studying the eff ect of robot adoption on wages
and productivity growth. Similar to what we show in fi gure 9.8, they fi nd
no systematic relationship between robot adoption and factor shares. They
do suggest that adoption is associated with boosts to labor productivity.