by Ajay Agrawal
neurial rents. Policy can undo these redistributions by taxing windfall gains
while leaving the price system to work at the margin. The result also holds for
decreasing- returns- to-scale production functions if we interpret the profi ts
372 Anton Korinek and Joseph E. Stiglitz
Box 14.1
Machine Labor and Factor Earnings
Assume a constant-returns-to-scale production function that
produces output Y by combining capital K with labor, consisting
of the sum of human labor H and machine labor M:
Y = F( K, H + M).
In this formulation human labor and machine labor are perfect
substitutes, so machine technology is what we call worker- replacing.
In the competitive equilibrium, the wage is determined by the
marginal product of labor,
w = F .
L
Proposition 1: Machine Labor and Factor Earnings: adding a mar-
ginal unit of machine labor reduces human wages but increases
the returns to capital in a zero-sum manner, in addition to increas-
ing output by the marginal product of labor, which is equal to the
wage.
Proof: Using Euler’s Theorem, we rewrite the production func-
tion:
( H + M ) F ( ) + KF ( ) = F( K, H + M ).
L
K
We can now ascertain the eff ect of an additional unit of M:
F + ( H + M ) F + KF
= F ,
L
LL
KL
L
or, simplifi ed, ( H + M ) F +
KF
= 0.
LL
KL
decline in wage bill
increase in return to K
Source: Korinek and Stiglitz (2017).
earned by the owner of the technology as compensation for the implicit
factor “entrepreneurship,” which takes part in the zero- sum redistribution.
Let us also emphasize that taxes on previously accumulated factors that
suddenly earn an unexpected excess return are nondistortionary. This means
that at least in principle, there is a role for implementing costless redistri-
bution and generating a Pareto improvement. (In practice, there are some
natural caveats to this result. For example, it relies on the assumption that
we can distinguish between previously installed capital that earns windfall
gains and new capital that would be distorted if it were taxed.)
AI and Its Implications for Income Distribution and Unemployment 373
14.4.2 Dynamic Implications of Worker- Replacing Progress
In the longer run, worker- replacing technological change will lead to sig-
nifi cant economic change. It implies that the biggest constraint on output—
the scarcity of labor—is suddenly relaxed. As a result, greater amounts of
complementary factors, here capital, are accumulated.
Observation 7) Machine Labor and Abundance of Labor: If not only
capital, but also labor, is reproducible at suffi
ciently low cost, then the economy
will grow exponentially in AK fashion, driven purely by factor accumulation,
even in the absence of further technological change.
In Korinek and Stiglitz (2017), we describe the dynamics of this transi-
tion as machines made by machines get increasingly effi
cient or, equiva-
lently, as the cost of producing machines decreases. We identify a singularity
point at which it becomes cost eff ective for machines to start to fully replace
human labor.23 In the simplest case, when complementary factors such as
capital adjust without friction, the human wage may actually be unchanged
because capital K grows in proportion to eff ective labor ( H + M) so that the marginal productivity of labor and the wage remain unchanged. In other
words, investment is allocated between conventional machines and human-
replacing robots in such a way that the return is equal to the intertemporal
marginal rate of substitution. Under the assumption that workers only care
about their absolute income, not their relative income, this outcome would
not be too bad: in absolute terms, even though the human labor share would
go to zero as an increasing fraction of the labor in the economy is performed
by machines, workers are no worse off as a result of AI.
When factors are slow to adjust, the pattern of transition can be com-
plex, with demand for human labor typically going down temporarily.24 In
general, the pattern of adjustment depends on how fast the capital stock
versus the stock of labor adjust. (For example, if the capital stock rises in
anticipation of an increased supply of machine labor in the future that has
not yet materialized, then human wages may even go up at intermediate
stages.)25
However, the following observation describes that in the long run, workers
are actually worse off as a result of machine labor if there are nonreproduc-
ible complementary factors that are in scarce supply, such as land or other
natural resources.
23. This singularity captures the important economic aspects of what technologists such as Vernor Vinge (1993) or Ray Kurzweil (2005) call the technological singularity. A similar point is also made in Aghion, Jones, and Jones (2017).
24. Berg, Buffi
e, and Zanna (2018) shows that it may actually take decades for the economy’s
complementary capital stock to adjust after major revolutions in labor- saving technology.
25. This assumes that capital is “putty- putty,” that is, that capital investments made before AI arrives are equally productive after AI, as would be the case if humans and robots were in fact identical.
374 Anton Korinek and Joseph E. Stiglitz
Observation 8) Machine Labor and Return of Scarcity: if there are non-
reproducible complementary factors, they eventually limit growth; human real
wages fall, and the owners of nonreproducible factors absorb all the rents.
Intuitively, as the supply of eff ective labor proliferates due to the intro-
duction of machine labor, agents in the economy will compete for scarce
nonreproducible resources like land, driving up their price.
A similar argument holds for nonreproducible consumption goods: even
if all factors in the production process are reproducible so that produc-
tive output in the economy exhibits AK- style growth and workers’ product
wages remain unchanged, competition for fi xed resources that are part of
their consumption basket, such as land used for housing, may lead workers
to eventually be worse off . This may be particularly important in urban set-
tings where, say, economic activity occurs at the center. Rich rentiers may
occupy the more desirable locations near the center, with workers having to
obtain less expensive housing at the periphery, spending more time commut-
ing. The advent of AI will thus lower their utility.
However, just as in the earlier case, at the margin, the redistribution from
workers to nonreproducible factor owners is zero sum. Since taxes on non-
reproducible factors are by defi nition nondistortionary, there is scope for
nondistortionary redistribution.
Observation 9) Nonreproducible Factors and Pareto Improvements: so
long as nondistortionary taxes on factor rents are feasible, labor- replacing
innovation can be a Pareto i
mprovement.
14.4.3 Redistributing the Innovators’ Surplus
via Changes in Institutions
If outright redistribution is not feasible or limited, there may be other
institutional changes that result in market distributions that are more favor-
able to workers. For example, intervention to steer technological progress
may act as a second- best device.
In this section, we provide an example in which a change in intellectual
property rights—a shortening of the term of patent protection—eff ectively
redistributes some of the innovators’ surplus to workers (consumers) to
mitigate the pecuniary externalities on wages that they experience, with the
ultimate goal that the benefi ts of the innovation are more widely shared.
If an innovation results in a lower cost of production, then the innovator
enjoys the benefi ts of the innovation in the form of higher profi ts during the
life of the patent; but after the expiration of the patent, society enjoys the
benefi ts in terms of lower prices. The trade- off is that shortening the life of
the patent may reduce the pace of innovation. But in the spirit of the theory
of the second- best, there is generally an “optimal” patent life, in which there
is still some innovation, but in which the well- being of workers is protected.
Box 14.2
Intellectual Property Regime and Redistribution
Consider an economy with a unit mass of workers H = 1, in
which the capital stock supplied each period K(τ) is a function
solely of a distortionary capital tax τ, the proceeds of which are
distributed to workers, and the eff ective stock of machine labor
M( z) is an increasing function of patent life z.
A worker’s total income I consists of her wage plus the revenue
of the capital tax,
I = w + τ K(τ).
For any level of M( z), we defi ne τ( M) as the value of the capital tax that keeps workers just as well off as they were before the
introduction of machine labor.
Proposition 1. As long as elasticity of capital supply is not too
large, we can always increase z from z = 0 and compensate work-
ers by raising the capital tax τ.
Steady-State Dynamics
Consider an intertemporal setting in which the growth rate g =
g ( z, τ) is a function of the length of the patent z and the tax rate τ, by which we now denote the tax rate on innovators. Assume
that the share of output that is invested is a function of the
growth rate ( i ( g)) and that the fraction of output not spent on
investment that is appropriated by the innovator is b( z, τ). In
steady state, the present discounted value of the income of work-
ers can be approximated as
PDV = (1 – i ( g)) [1 – (1 – τ) b( z, τ)]/ ( r – g), where r is the discount rate. If we choose { z, τ} to maximize the
PDV, in general, the optimum will not be a corner solution in
which any innovation hurts workers.
Proposition 2. In general, the optimal { z*, τ*} entails g > 0.
It is easy to write down suffi
cient conditions under which Propo-
sition 2 holds: setting τ* equal to zero, all that we require is that
| g | is not too large relative to | b |.
z
z
376 Anton Korinek and Joseph E. Stiglitz
With network externalities the innovator may be able to maintain a domi-
nant position even after the end of the patent, and may continue to earn the
surplus from her innovation. With taxes on monopoly profi ts, it should be
possible to ensure that the innovations are Pareto improving and that even
human worker- replacing technological change can improve the well- being
of workers.
14.4.4 Factor- Biased Technological Change
So far, we have simply assumed that technological change—the intro-
duction of AI—is worker replacing. But advances in technology also make
some machines more productive and others obsolete, aff ecting the (mar-
ginal) return to traditional capital.26 It is thus useful to think of the world
as having three groups: capitalists, workers, and innovators. Intellectual
property rights (and antitrust laws) determine the returns to innovators,
but the nature of technological change in a competitive market determines
the division between workers and capitalists.
A long- standing literature, going back to Kennedy (1964), von Weizacker
(1966), and Samuelson (1965), describes the endogenous determination of
the factor bias of technological progress.27 The central result is that as the
share of labor becomes smaller, the bias shifts toward capital- augmenting
technological progress. If the world works as these models suggest, this
should limit the decline in the share of labor (at least in a stable equilib-
rium) and in inequality.28 As the share of labor decreases, the incentive to
produce worker- replacing innovation such as AI decreases. But the relevant
discounted future wage share near the point of singularity—the point where
it is cost eff ective for machines to fully replace human labor and produce
more machines all by themselves—may be suffi
ciently great that there is
nonetheless an incentive to pass the point of singularity.
Let us assume that land becomes the binding constraint once human labor
is fully replaceable by machine labor. In that case, provided the elasticity of
substitution between land and the other production factors—capital cum
labor—is less than unity, the share of land increases over time, generating
the result (analogous to that where labor is the binding constraint in the
standard literature) that in the long run, all technological progress is land
augmenting. If the production function is constant returns to scale in land,
26. As we noted earlier, IA (intelligence assisting) innovation may increase the productivity of humans, and thus increase the demand for humans if the elasticity of substitution is less than unity.
27. Important contributions were also made by Drandakis and Phelps (1965). More recently, there has been some revival of the literature, with work of Acemoglu (2002), Stiglitz (2006, 2014b) and Acemoglu and Restrepo (2018), among others.
28. One can describe dynamics with standard wage- setting mechanisms. The system is stable so long as the elasticity of substitution between factors is less than unity (Acemoglu 1998; Stiglitz 2006, 2014b).
AI and Its Implications for Income Distribution and Unemployment 377
labor (including machine labor) and traditional capital, then the long- run
rate of growth is determined by the pace of land- augmenting technological
change.
Role of the Service Sector
Currently, progress in AI focuses on certain sectors of the economy, like
manufacturing. Partly because of the resulting lower cost of manufacturing,
and partly because of the shape of preferences, the economy is evolving
toward a service- sector economy. (If there is diff erential productivity across
sectors, and the elasticity of demand for the innovation sector is not too
high, then production factors will move out of that sector into other sec-
tors. This is even more so if preferences are nonhomothetic, for example,
demand for food and many manufactured goods having an income elastic-
ity less than unity.) Among the key service sectors are education, health,
the military, and other public services. The value of those services is in
large part socially determined, that is, by public policies not just a market
process. If we value those services highly—pay good wages, provide good
working conditions, and create a suffi
cient number of jobs—this will limit
increases in the inequality of market income. Governments typically play
an important role in these sectors, and their employment policies will thus
play an important role in the AI transition. Many of these service- sector
jobs have limited skill requirements. However, higher public- sector wages
will—through standard equilibrium eff ects—also raise wages in the private
sector, will improve the bargaining position of workers, and will result in
such jobs having higher “respect.” All of this will, of course, require tax
revenues. If the elasticity of entrepreneurial services is low, for example, if
entrepreneurs are driven partly by nonpecuniary motives, we can impose
high taxes to fi nance these jobs.
14.5 Technological
Unemployment
Unemployment is one of the most problematic societal implications
of technological progress—new technology often implies that old jobs
are destroyed and workers need to fi nd new jobs. Economists, of course,
understand the “lump- of-labor fallacy”—the false notion that there is a
fi xed number of jobs, and that automating a given job means that there will
forever be fewer jobs left in the economy. In a well- functioning economy,
we generally expect that technological progress creates additional income,
which in turn can support more jobs.
However, there are two sound economic reasons for why technological
unemployment may arise: fi rst, because wages do not adjust for some struc-
tural reason, as described, for example, by effi
ciency wage theories, and
second as a transition phenomenon. The two phenomena may also interact
378 Anton Korinek and Joseph E. Stiglitz
in important ways, for example, when effi
ciency wage considerations slow
down the transition to a new equilibrium. We discuss the two categories in
turn in the following subsections.
The unemployment implications are especially problematic when techno-
logical progress is labor saving, which—by defi nition—requires that either