vol. 20, no. 2 (Summer), 1988, pp. 207-234
Phillips Curves And Hayekian Triangles:
Two Perspectives on Monetary Dynamics
Don Bellante and Roger W. Garrison
During different phases of the Keynesian episode, the monetary theories
offered first by Keynes and later by the Keynesians have been judged by
changing standards. The standards changed along with the changing perceptions
of what constituted the most viable alternative to the Keynesian vision.
"[T]here was a time," wrote John Hicks (1967, p. 203), "when the new theories
of Hayek were the principal rival of the new theories of Keynes." But times
changed, and Milton Friedman (1969d), with his restatement of the quantity
theory of money, became the dominant alternative to Maynard Keynes. And
eventually, Friedman's Monetarism gave way to New Classicism and the idea
of "Rational Expectations."
The profession has been
treated to exhaustive comparisons of Keynes and then Keynesians with the
various opposing schools, but comparisons of the sequential alternatives
to Keynesianism have been all but lacking. The present paper begins to
fill this void by offering a critical comparison of the Monetarists and
the Austrians as represented by Friedman and Hayek.(1)
A statement by Robert Lucas (1981, p. 4) suggests how such a comparison
can be undertaken: "...I see no way to account for observed employment
patterns that does not rest on an understanding of the intertemporal substitutability
of labor." Lucas's concise way of identifying the understanding that explicitly
underlies his theories (and implicitly underlies Friedman's) hints at an
alternative way of accounting for observed unemployment patterns. While
Monetarism and New Classicism are based on the intertemporal substitutability
within the market for labor, Austrianism is based on the intertemporal
complementarity within the market for capital goods.
Differences between the
monetary theories of Friedman and those of Hayek, then, can be spelled
out in terms of the markets (for labor and for capital, respectively) that
serve as a focus for each.(2) With attention
to the major themes of each theorist, Sections II and III identify the
relevant aspects of monetary disturbances as seen by Friedman and as seen
by Hayek. Focusing on both substance and method, Section IV offers a critical
comparison of the two perspectives. Section V points out some implications
in terms of the conventionally defined categories of unemployment, monetary
lags, and the concept of "full" employment; and Section VI provides a summary
II. Friedman (and the Monetarists) on Monetary Dynamics
With conventional qualifications and allowances for real growth, increases
in the supply of money lead, in the long run, to proportionate increases
in the general price level. This proposition, which constitutes the kernel
of truth in the quantity theory of money, is not in dispute. But the nature
and significance of the monetary dynamics, the market process that translates
the initial cause into its ultimate effect, is and has long been a matter
of much controversy. This issue, in fact, is what separates the Monetarists
from the Austrians and gives scope for interpretation within both schools.
Friedman appears to be of
two minds on the issue of monetary dynamics—the transmission mechanism
linking money to prices. On occasions where the focus is on long-run comparative-statics
results, it is simply admitted that he (along with Anna Schwartz) has "little
confidence in [their] knowledge of the transmission mechanism, except in
such broad and vague terms as to constitute little more than an impressionistic
representation rather than an engineering blueprint" (Friedman, 1969b,
p. 222). But on occasions where the focus is on the transmission mechanism
itself, such as the market process that traces out a short-run Phillips
curve, there is an accounting of the the mechanism in terms of the market
for labor that would rival any blueprint(3)
(Friedman, 1976). These monetary dynamics, which are used by Friedman to
explain the short-run nature of the negatively sloped Phillips curve and
to suggest the existence of a vertical long-run Phillips curve, can be
used as a basis for evaluating the Monetarist view and for comparing it
to the alternative offered by the Austrians.
The monetary dynamics envisioned
by Friedman hinge on the sequential effects of perceived relative-price
changes in the market for labor. A number of heuristic assumptions about
the market for capital goods and about income effects in commodity markets
are invoked. These assumptions are required to narrow the focus so that
the Monetarists' story can be told. The lack of discussion in this context
of the allocation of capital goods or of the short-run effects of a monetary
injection within the market for capital goods is evidence enough that such
considerations are no part of the story.(4)
Implicitly, one of several alternative constructions is adopted: (1) Real
capital is taken to be completely homogeneous, or—to use the fiction invented
by Frank Knight—it is treated as a "Crusonia plant." (2) The structure
of real capital, which admittedly consists of heterogeneous elements, is
fixed. It cannot or, for some reason, is not altered—even in the short
run—as a result of a monetary injection. (3) Allocations within the market
for capital goods are (somehow) always governed, whether in the presence
or the absence of monetary disturbances, by "real factors only." This third
construct is in the spirit of the New Classicism.
One of these three or some
effectively similar construction or assumption must underlie any macroeconomic
theory that ignores the allocation of resources within the capital-goods
sector. Implicit assumptions of this sort about capital goods are not at
all at odds with the Chicago tradition more broadly conceived. The inattention
to capital theory stems from Frank Knight's grappling with the thorny issues
and conceptual difficulties that inhere in this subject matter (Knight,
1934). In general, Monetarists have taken comfort in the Knightian view
that the structure of capital, particularly the intertemporal structure,
can be safely ignored, and that theories in the Austrian tradition, which
make use of such concepts as "roundaboutness" and "stages of production,"
are especially misguided.(5)
The Knightian view of capital
permits the Monetarists to focus exclusively on the market for labor. But
at least two additional assumptions are needed to limit the focus to the
relative-price effects in the labor market. Distribution effects (who gets
the new money) and differential income effects (how it gets spent) must
be removed from consideration. Friedman's heuristic device for short-circuiting
the distribution effects is to assume that increases in the money supply
are brought about by a one-time dropping of money from a helicopter in
such a way that each individual picks up the new money in direct proportion
to the amount already in his possession.(6)
The differential income effects of "helicopter money"—as it has come to
be called—are assumed away. This mode of theorizing reflects the implicit
assumption that differential income effects are in fact negligible or the
heuristic assumption that indifference curves are both identical across
agents and homothetic.(7)
Within this theoretical
construction, the Quantity Theory holds in its strongest form, and any
divergence in the pattern of prices between the initial injection of money
and the eventual increase in the price level is purely stochastic. Thus,
disequilibrium relative-price movements within the markets for both capital
goods and final products are taken to be unsystematic. No generalizations
about such movements can be made. But movements in the price of labor relative
to the price of final output are systematic in the Monetarist view.
And the temporal pattern of these movements depend in a critical way on
differences in the ability of workers and of employers to perceive the
price changes most relevant to each.(8)
The spending of newly created helicopter money begins to bid up the prices
of final products in unsystematic ways. The individual worker, who clearly
perceives his yet-unchanged nominal wage, has no clear perception of the
change in his real wage—depending, as it does, on the change in some index
of final-product prices. The employer, whose perception of the general
price level is no better than the workers', is motivated by a different
concern. From the employer's perspective, the relevant real wage is the
Ricardian real wage, which depends upon a single price—the price of the
product that the employer produces. If the output price has increased,
the wage that he pays to the workers—relative to the output price—has clearly
decreased. Alternatively stated, workers and employers alike have no clear
perception of the real wage, where real is understood in the market-basket,
or Fisherian, sense; but employers have a clear perception of the real
wage, where real is understood in the Ricardian sense.)
Quantity adjustments in
the labor market are made in ways that correspond to the differing perceptions
in real-wage movements. The behavior of workers, who make their labor-leisure
decisions on the basis of yet unchanged perceptions, is depicted by an
unchanged labor supply curve; the behavior of employers who now enjoy higher
output prices is depicted by a rightward shift in the demand for labor.(9)
The nominal wage rate rises as does the level of employment. (Figure 1
shows the corresponding movement (from A to B) along the initial Short-run
Phillips Curve.) The higher nominal wages paid to a larger number of workers
exert upward pressure on the prices of final products; increases in output
exert downward pressure on final-product prices. And with the passage of
time, workers begin to get a clearer perception of their real wage rate.
A temporal pattern of the real wage rate is traced out by a series of iterative
steps in which the reinforcing and counteracting forces interact. In the
end, after a "long and variable"—and fundamentally indeterminate—time lag,
the worker-employer perception differential is eliminated; the short-run
Phillips curve shifts rightward. The level of employment, the level of
output, and the real wage (both Fisherian and Ricardian) return to the
levels that characterized the economy before the monetary injection. (Point
C in Figure 1 differs from point A only in terms of absolute prices and
III. Hayek (and the Austrians) on Monetary Dynamics
If Frank Knight accounts for the Monetarists' inattention to capital
theory, Eugen von Böhm-Bawerk (1959) accounts for the Austrians' preoccupation
with it. Where Friedman's treatment of monetary dynamics requires some
key assumptions about the workings of the market for capital goods, Hayek's
treatment(10) requires similar assumptions
about the workings of the labor market. Recognizing this symmetry allows
us to describe the alternative treatments in a way that facilitates a comparison.
More often than not, Hayek's
assumptions about labor, like Friedman's assumptions about capital, are
implicit. Labor skills are assumed to be non-specific. Individual occupations
are defined in terms of the particular capital goods that are complementary
to labor. Wage rates are flexible, but not perfectly flexible. Workers
can be bid away from some occupations and into others, but not instantaneously.
Adjustments in the labor market that involve a reduction in labor demand
in some occupations will be characterized by temporary increases in the
level of unemployment—the greater the adjustment, the greater and longer-lasting
the temporary increase. With allowance for frictions of this sort, workers
are able correctly to perceive and respond to changes in the pattern of
real wage rates. Expectations about wage rates and prices can come into
play—but not expectations whose formulation requires a theoretical understanding
of economic relationships, such as (correct or "rational") expectations
about the upper turning point of a business cycle.
It might be noted at this
point that Lucas (1981, pp. 215-17) sees a certain kinship between his
own ideas and those of Hayek. But Lucas parts company with the Austrians
when he treats knowledge of the economy's structure in the same
way as knowledge within the structure. Hayek (1948b, pp. 79-81)
makes a first-order distinction between theoretical knowledge and knowledge
of the marketplace. Market participants can be expected to make use of
information conveyed by prices along with other particular knowledge that
they might have, but they cannot be expected to know—even in a probabalistic
sense—the parameters of the economy's structure. That is, they cannot be
expected to know, or to behave as if they know, the answers to questions
that economists have been debating amongst themselves for more than two
The assumptions spelled
out above about the workings of the market for labor allow the Austrians
to deal with monetary dynamics exclusively in terms of the market for capital
goods. The dynamics within the capital-goods market, coupled with these
assumptions, will have clear implications about the corresponding pattern
and time path of the employment of labor. The effects of a monetary disturbance
within the market for capital goods reflect several considerations. Capital
goods in the Austrian view are heterogeneous in the extreme, and the structure
of capital involves multidimensional complexity. Individual capital goods
are characterized by different degrees of specificity and are related to
one another, both intertemporally and atemporally, with various degrees
of substitutability and complementarity.(12)
Thus, a set of heuristic assumptions about the capital structure is required
to allow the identification of its most essential features and to render
the treatment of monetary dynamics tractable.
In the Austrian view, the
central problem in macroeconomics is the problem of intertemporal discoordination.
(O'Driscoll, 1977, pp. 70-79; Garrison, 1984, 1985) Whether the focus is
on the coordination of investment decisions with consumption decisions
or on the time-pattern of macroeconomic magnitudes over the course of a
business cycle, the temporal element is essential to the macroeconomic
perspective. Hayek incorporated this temporal element into his monetary
dynamics by focusing on the intertemporal aspects of the economy's capital
structure. To avoid undue complexity, Hayek envisioned a heavily stylized
His vision gives recognition,
sometimes explicitly and sometimes implicitly, to a set of corresponding
heuristic assumptions. Each production process is characterized by a modifiable
sequence of inputs and a point output; some production processes are more
time-consuming than others. The sequence of inputs is conceived as consisting
of "stages" of production. The once-
triangles(13) represent a stylization of
the entire economy's production process (See Figure 2): The horizontal
leg represents the time element in the process—the depth of the capital
structure. In the simplest case this leg represents the time that sepatates
the earliest stage of production from the final output; the vertical leg
represents the nominal value of the final output. The height of the hypotenuse
at successive points in time represents the value of semi-finished goods
as they move through time from the earliest to the latest stage of production.(14)
Capital goods can be shifted—within
limits—from one production process to another in response to relative-price
changes. More importantly, capital goods can be shifted from one stage
of production to another in ways that modify the intertemporal pattern
of output. Some types of capital goods that are employed in the earlier
stages of production (or that, with some modification, can be so employed)
may be needed in the later stages of production as well. That is, while
competing for the use of individual capital goods, the stages themselves
exhibit a certain degree of intertemporal complementarity. There is no
one-to-one relationship between stages of production and business firms.
Some firms may operate within one stage, some in several sequential stages,
and some in different stages of different production processes.
Spelling out the characteristics
of the Hayekian structure of production is, by itself, almost enough to
specify the nature of the corresponding monetary dynamics. The general
pattern of events set into motion by a monetary injection can be identified
as soon as the method of injecting the new money is specified. Because
of the historical relevance, Hayek assumed that the new money is injected
through credit markets—that the central bank, in effect, pads the supply
of loanable funds with newly created money.(15)
Clearing the market for
loanable funds in the face of such a monetary injection requires that the
rate of interest fall until the quantity of funds demanded matches the
increased supply.(16) In turn, this lowered
rate of interest has implications for the intertemporal structure of capital.
To the extent that the temporal relationship between the various types
of capital goods and the ultimate output of the production processes is
perceived by entrepreneurs, the prices of the capital goods will be affected
in a systematic way. In the earlier phases of the market's reaction to
the credit expansion, the greater the time between the use of the capital
good and the emergence of the ultimate output, the greater the relative
increase in the price of the capital good. This pattern of relative-price
changes follows from the application of standard discounting techniques.
There will be a corresponding pattern of quantity adjustments. Capital
will be bid away from relatively less time-consuming production process
into relatively more time-consuming processes and away from relatively
late stages of production into relatively early stages of production. Marginal
adjustments of these sorts will be made throughout the capital structure
as firms seek to take advantage of the favorable credit conditions. In
Figure 2 the shift towards more time-consuming processes is represented
by a movement from A to B. (The change in the slope of the hypotenuse reflects
smaller profit differentials between the stages of production, which in
turn reflect cheaper credit.)
The later phases of this
dynamic process are marked by a reversal of the price and quantity movements
that characterized the earlier phases (Hayek, 1967, p. 58). The passage
of time takes production projects that were begun as a result of credit
expansion into their later stages. Some of these later stages require the
use of capital that was used up in—or irrevocably committed to—earlier
stages. That is, the money-induced expansion caused more capital to be
committed to the earlier stages of production without providing the additional
resources—as would have been provided had the expansion been savings-induced—necessary
for the completion of all the production processes. In Figure 2 the value
of the production projects in their final stages is represented by a broken
line, indicating that a credit-induced boom cannot be sustained.
Capital goods complementary
to the yet-uncompleted production processes are now in short supply. Their
prices are bid up, and as a result of these higher prices, the demand for
credit increases. The rate of interest, which had been artificially low
during the monetary expansion, is now bid up. Two significant differences
between the Austrian and the Monetarist views can be noted here: First,
because of intertemporal complementarity, the initial investment raises
the demand for capital(17); second the
resulting rise in the interest rate is separate from any Fisher effect,
which depends upon a rising price level.(18)
Both directly through the
market for capital goods and indirectly through credit markets, the prices
of capital goods committed to the early stages of production are bid down.
Uncommitted capital goods are bid away from the earlier stages of production
and into the later stages.(19) But in this
phase of the dynamic process, marginal adjustments may not be possible.
Some capital goods attracted to the earlier stages of production by the
monetary expansion may not be retrievable. As a consequence, many production
projects may have to be abandoned; many others can be completed but only
with a great delay and/or at a much higher cost than could have been anticipated.
(The claim, made in the spirit of the New Classicism, that market participants
anticipate the money-induced capital shortage rings hollow. Such an anticipation
would require that they know—or behave as if they know—the "real scarcities"
independent of the price system which supposedly communicates that information
to them. In the Austrian view, if a monetary injection distorts the price
signals, market participants will be economizing on the basis of erroneous
This later phase of the
adjustment process takes on the characteristics of a crisis, a sharp down-turn.
But with time, some types of capital can be liquidated to accommodate the
excess demands for other types. Eventually, the economy's capital can be
restructured in a way that reflects real resource availabilities, and the
credit market can be adjusted to reflect the supply and demand for loanable
funds. Abstracting from the capital that is lost forever as a result of
the credit expansion and from possible long-run effects on the distribution
of income, the rate of interest and the corresponding structure of production
will return to the level and configuration that characterized the economy
before the credit expansion.
Figures 1 and 2 can serve
to depict the correspondence between the Monetarist and the Austrian views.
Arguing respectively in terms of the wage-rate effect on the employment
of labor and the interest-rate effect on capital utilization, Friedman
and Hayek have traced out the consequences of a monetary injection from
A to B to C, where, in both diagrams, points A and C represent identical
sets of real parameters. Point B represents—in Figure 1—a rate of unemployment
temporarily and unsustainably below the (Friedmanian) natural rate and—in
Figure 2—a depth of capital temporarily and unsustainably maintained by
a loan rate of interest kept below the (Wicksellian) natural rate.
To this point the Austrian
story has been told exclusively in terms of the market for capital goods.
Yet a major purpose of the story is to account for the cyclical unemployment
of labor. But filling in the blanks about how the labor market is affected
by the dynamics in the capital market is not difficult. In the trivial
case of perfect wage flexibility and instantaneous adjustments, there need
be no unemployment at all. Labor would simply be shifted around during
the unfolding of the dynamic process so as to be employed in ways consistent
with the changing pattern of relative prices.(20)
This is clearly not what Hayek had in mind.
The recognition that labor
is complementary to (some) capital and that frictions inherent in the market
process prevent adjustments from taking place instantaneously is all that
is required to translate the story about capital into a story about labor.
During the early phases of the dynamic process, there is a net increase
in the demand for labor. The new money injected through credit markets
is used not only to bid workers away from some jobs and into others but
to attract new workers as well. To use Friedman's terminology, unemployment
falls below its natural rate.
But during the late phases
of the process, there are actual decreases in the demand for labor in the
early stages of production. And the increases in the demand for labor in
the later stages is only partially offsetting—due to the shortage of complementary
capital goods. The frictions involved in the economy-wide movements of
labor out of the early stages and into the late stages coupled with the
net reduction in the demand for labor account for a considerable amount
of supernatural unemployment during this phase of the dynamic process.
Once the misallocated capital
has been liquidated and the wage rates have adjusted to the underlying
market conditions, the cyclically unemployed workers can be reabsorbed.
In Hayek's story as in Friedman's, unemployment eventually returns to its
IV. A Critical Comparison
While the two views of monetary dynamics spelled out in the previous
two sections differ in important respects, they are in several fundamental
respects quite similar. Comparing the differences, then, must be prefaced
by a clear recognition of the underlying similarities. Five points of commonality
are noteworthy: (1) Both theories can be fully squared with the kernel
of truth in the quantity theory of money. (2) Both theories deal with disequilibrium
phenomena, but neither denies that equilibrating forces dominate in the
end. (3) Both hinge in a critical way on the distinction between short-run
effects and long-run effects. (4) Both involve a market process that is
necessarily, or endogenously, self-reversing. Monetary disturbances cause
certain kinds of distortions in market signals. These distortions give
rise in the short run to movements in certain prices and quantities, movements
which in the long run create market conditions for counter-movements in
those same prices and quantities. (5) With appropriate qualifications (about
what constitutes the long-run) both theories are characterized by monetary
disturbances whose short-run effects are non-neutral but whose long-run
effects are neutral.
With allowance for these
points of commonality, Friedman and Hayek are offering in some respects
complementary views, in other respects competing views. Our comparison
will attempt to separate the two kinds of differences. Comparing aspects
of the two views that are at odds with one another must await a closer
look at each view separately. But ways in which the two views are different
but complementary can be readily identified.
At the root of these kinds
of differences is the fact that Friedman focuses his analysis on the market
for labor while Hayek focuses his on the market for capital goods.(21)
The trade-off that gets distorted by monetary disturbances is labor vs.
leisure for Friedman and goods now vs. goods later for Hayek. And due in
large part to the differing natures of labor and capital (the more radical
heterogeneity of capital as compared to labor(22)),
Friedman argues in terms of substitutability and Hayek in terms of complementarity.(23)
In an important sense there
is simply no scope for conflict here. Labor-leisure preferences and intertemporal
preferences interact with the perceived constraints that are relevant to
each. The dimensions of the two market processes can be seen as orthogonal
to one another. Trading off labor against leisure in a sequence of periods
can be understood as substituting labor in some periods for labor in others.
Thus, with both views laid out intertemporally, we see that Friedman is
dealing with the intertemporal substitutability of labor while Hayek is
dealing with the intertemporal complementarity of capital. Spelled out
in these general terms, then, we have a harmony, rather than a conflict,
1. A Critical Retelling of Friedman's Story
But when we turn again to the specifics of the market processes envisioned
by Friedman and Hayek, we see first-order differences. Some of these differences
put the two views at odds with one another—or at least allow for a preference
of one over the other on the basis of plausibility or fruitfulness. Identifying
these differences and their significance can begin with a critical assessment
of Friedman's view, which we preface with a recognition of existing criticism
in the literature.
Sketchy expositions of the
Monetarist view have given much scope for misinterpretation. Gardner Ackley
(1983, p. 10), for instance, sees the Monetarist dynamics purely in terms
of "tricks" played on employers and workers. It is a "trick for an inflation
to fool both employers and workers—in opposite directions—about
the movements of the real wage paid by one and received by the other."
But to call this a trick is to miss Friedman's point. The "real wage" means
one thing to workers and something else to employers. Neither workers nor
employers have a clear perception about what is happening to the general
price level, but both are responding in conventional ways to the incentives
that they face. Recognizing these incentives gives Friedman's view a microeconomic
footing and insulates it against criticism of the type offered by Ackley.
Another criticism of Friedman's
view (Birch, Rabin, and Yeager, 1982) is based on a perceived contradiction
between the market process envisioned and the equation of exchange.(25)
The familiar identity MV = PQ implies that when M, or more properly MV,
increases, the corresponding increase in PQ is split in some way
between an increase in P and an increase in Q. That is, Q can increase
only to the extent that P does not increase. By contrast and according
to the Monetarist account of the dynamic process, it is increases in P
that cause increases in employment and hence increases in Q. That is, Q
increases to the extent that P does increase—hence the perceived contradiction.
But what appears at first blush to be a contradiction is in fact a manifestation
of a self-reversing process. There is nothing logically contradictory
about a process in which P begins rising before Q but in which Q falls
as P becomes fully adjusted to the increase in M. While the Ps and the
Qs can be squared with the equation of exchange at each point in the process,
Q's initial upturn and inevitable downturn, which reflect similar movements
in the employment of labor, constitute the essential self reversal that
lies at the heart of the Monetarist view.
Our own criticism is fundamentally
different from the ones identified above. While the sequence of movements
in P and Q are not evidence of a contradiction, they are evidence of a
certain anomaly in the Phillips Curve story. An initial rise in prices
is a prerequisite to any movement along a short-run Phillips curve, but
the story accounts inadequately for the nature of this initial rise. To
make the story stick we must recognize that there is some other, logically
prior, market process that is set into motion by a monetary injection.
The process can be easily identified by drawing more broadly from the Monetarist
literature. Helicopter money adds to each individual's cash balances, thereby
inducing greater spending and the bidding up of prices (Friedman, 1969c,
p. 5). But if buyers of labor services receive their pro-rata share
of the helicopter money, they would be bidding up the price of labor at
the same time. The simultaneous rise in the price of output and the price
of labor would preclude the fall (as perceived by the employer) in the
real wage, which itself constitutes a critical aspect of the self reversing
dynamic process. The real-cash-balance effect, then, becomes the whole
story rather than a prelude to the Phillips-curve story.
The Phillips Curve story
can be saved by assuming a monetary injection in which the new money falls
first into the hands of consumers and only later into the hands of producers.
While this kind of assumption, which highlights a particular distribution
effect, violates the spirit of Monetarism, it allows in a straightforward
way for a variation on the Austrian theme. We turn now to consider the
2. Heuristic Assumptions and Domain Assumptions
The assumptions made by Friedman and Hayek about the nature of the
hypothesized monetary injection are not just two alternative assumptions
that serve the same purpose. Friedman's assumption (that money is dropped
from a helicopter) is a heuristic assumption. It is a deliberate
fiction whose purpose is to bypass any questions that relate directly to
the actual injection of money while still allowing for the derivation of
implications that can be tested empirically. (Friedman's uninhibited use
of such fictions identifies his method as instrumentalism.(26))
Hayek's assumption that
money is injected through credit markets is a domain assumption.(27)
In many instances money actually is injected through credit markets.
Thus Hayek's story applies directly and without modification to those instances.
A lower rate of interest resulting from the credit expansion increases
the quantity of credit demanded. Given the relative volumes of commercial
lending and consumer lending, we can say that the new money falls first
into the hands of producers and only later into the hands of consumers.
This disproportionate distribution of the new money is consistent with
Hayek's story about the effect of a monetary injection on the structure
of capital: Production for future consumption is temporarily favored
over production for present consumption.
For instances in which money
is injected by some other means, Hayek's story applies only after suitable
modifications are made. Suppose, for example, that a monetary expansion
takes the form of increased transfer payments to consumers. This type of
monetary injection would temporarily favor present consumption over
consumption. Capital goods would be reallocated accordingly. In many respects,
the self-reversing market process triggered by such an injection would
be a mirror image of the process triggered by a credit expansion. Capital
goods in the later stages of production would be first in short supply
and subsequently in surplus. The demand for labor would rise and then fall
as workers were bid into the later stages of production only to become
unemployed when the demand for present consumption fell to its "natural"
level. The downturn associated with such a transfer expansion may
be less severe than one associated with a credit expansion if only
because short-term capital can be liquidated more quickly than long-term
Two observations are warranted
here. First, Friedman needs to incorporate a transfer expansion, or something
like it, into his theoretical construction if his Phillips curve story
is to become coherent. Second, inflation-rate and unemployment data that
suggest a negatively sloped short-run Phillips curve and a vertical long-run
Phillips curve are consistent with Hayek's story whether the new money
is lent into existence or transferred into existence.(28)
Before we suggest what empirical, or historical, considerations might constitute
a basis for choosing between the alternative stories offered by Friedman
and Hayek, we turn to one further issue—the issue of generalizability.
Any theoretical construction that suggests how a particular market
works may be more or less generalizable—adaptable to different circumstances
or extendable to other markets—depending in large part upon the nature
of the assumptions used in the construction. The discussion above suggests
that theories based on domain assumptions are more generalizable than theories
based on heuristic assumptions. Hayek's story can be adapted to apply to
circumstances in which the new money favors consumption activity over investment
activity even though it was first told the other way around. The story
can be generalized to recognize that monetary expansion can cause intertemporal
discoordination—in one direction or the other—depending upon the particular
device used to inject the new money. In recent years Hayek has generalized
his own story even further by recognizing that monetary expansion causes
a self-reversing discoordination in many markets, whether or not it causes
any intertemporal discoordination (1975a, pp. 23-24). By generalizing in
this way, Hayek has not at all changed his mind about the effects of monetary
expansion, he has simply recognized that the domain (credit expansion)
that once dominated his subject matter is no longer so dominant. But money-induced
discoordination—of one sort or another—still dominates the story.(29)
Friedman's story makes use
of a monetary helicopter whose precise function is to assume away the discoordination
that Hayek's story deals with. The monetary helicopter does not constitute
one domain from which the theory can be extended; it constitutes the intentional
neglect of all such domains. The Friedmanian helicopter, like the Walrasian
auctioneer in a different story, should be seen as a "red flag," a marker
where something important has been left out of account so that some other
part of the story could be told. From this perspective we can admire the
outrageousness of the particular fiction employed: the helicopter is a
red flag that virtually no one could fail to see. But a red flag is no
basis for generalization. If we go back and take into account the effects
of actual monetary injections, we do not get a generalization of Friedman's
story; we get Hayek's story.
Hayek's story is generalizable
in another important respect. The task of identifying the effects of credit
expansion was made tractable by employing a heuristic assumption about
the structure of capital-using production processes—the assumption of multi-period
inputs and a point output. This particular assumption allowed for the abstraction
from many complexities while it retained the essential element—the time
element—in the analysis. Once the story is told in its most tractable form,
then, it can be generalized and extended to take into account some of the
complexities that were initially assumed away.(30)
Production processes with multi-period outputs can be taken into account
as well as processes that make use of capital goods of a greater or lesser
degree of durability or whose final output is a durable consumer good.
And by recognizing the ways in which the market for "human capital" is
like the market for capital goods, Hayek's analysis can be extended in
a direct way to labor markets as well (Bellante, 1983).
Unfortunately, Hayek's attempt
to tell his story in the contexts of several different circumstances was
seen as evidence of confusion on Hayek's part. In response to criticism
that he assumed an initial state of full employment and placed too much
emphasis on changes in the terms of credit, Hayek (1975c, pp. 3-37) offered
an alternative account in which widespread unemployment was assumed and
the loan rate of interest was held constant. The resulting story was then
criticized (Kaldor, 1942) on the grounds that it contradicted Hayek's own
earlier rendition. By uncritically adopting Kaldor's assessment of what
he dubbed the "concertina effect," modern critics have failed to appreciate
the adaptability, the generalizability, that characterizes Hayek's formulation.(31)
A similar sort of generalization
and extension is not possible for Friedman's story. The essential feature
on which his story depends is unique to the particular context in which
it is told. For Friedman, the differing perceptions of workers and employers
is what gives rise to the story; they are what "drive the system." The
market process that Friedman identifies has no direct counterpart in the
market for capital goods. The owner of a diverse capital stock, for instance,
is unlikely to have perceptions that differ in a systematic way from those
who may buy the services of that capital stock. Such perception differences
are even less likely in the more prominent cases in which capital goods
are owned indirectly through financial markets. Thus, independent of any
empirical considerations, we have some basis for preferring Hayek's story
over Friedman's. Hayek's domain assumptions and even his heuristic assumptions
allow for generalization and extension in ways that Friedman's do not.
Hence, Hayek's theoretical construction is the richer, the more fruitful,
of the two.
4. Historical Applications and Policy Implications
Extensive empirical testing has favored Friedman's view of the nature
of the short-run Phillips curve over beliefs or hopes that there is a more
permanent trade-off between inflation and unemployment. But this same empirical
testing provides no clue at all about the nature of the market process
that moves the economy along a short-run Phillips curve and then shifts
that curve so as to conform with the long-run vertical Phillips curve.
That is, the testing allows to us choose between a Keynesian view and a
Friedman-Hayek view, but not between the Friedman view and the Hayek view.
Direct empirical evidence
about the nature of the monetary dynamics involved in adjusting the economy
to a monetary injection may be all but impossible. Directly substantiating
Friedman's story would require data on perceived real wage rates; directly
substantiating Hayek's story would require the quantification of actual
changes in the complex structure of capital.(32)
We must look for implications of the two views that allow for empirical
differentiation. As it turns out, our comparison of the two views provides
the needed clue about differing implications.
In Friedman's story a general
rise in prices of final output is a necessary link in the self-reversing
market process triggered by monetary expansion. Without this price inflation,
which has a different significance for workers and for employers, there
is no story to tell. While Hayek's story allows for price inflation, his
story does not depend upon it (1967, pp. 22-30; 1975b, pp. 104, 121, 196,
and passim). An initially depressed interest rate followed by subsequent
resource constraints within the market for capital goods can serve alone
as the basis for the self-reversing market process envisioned by Hayek.(33)
This realization that price inflation plays a leading role in one story
and, at best, a supporting role in the other allows us to identify important
differences in the way the two stories are used not only in the interpretation
of history but also in the prescription of policy.
Historical episodes in which
monetary expansion is accompanied by an unchanging price level are interpreted
differently by Friedman and Hayek. The decade of the 1920s provides the
most dramatic illustration of this difference. Real economic growth during
this decade, which in the absence of monetary expansion would have produced
a decline in the price level, served instead to offset the inflationary
effects of the monetary expansion. The self-reversing process that, in
Friedman's view, characterizes other episodes of monetary expansion does
not get triggered in this one. There is no Phillips-curve story to tell.
Economic problems that surfaced at the end of the '20s are not, according
to the Monetarist view, a result of a market process that began much earlier
in the decade. The economic downturn is blamed instead on exogenous factors—the
incompetence and irresponsible behavior of the central bank (Friedman,
1963, pp. 299-419).
From an Austrian perspective,
the same historical episode is seen quite differently (Hayek, 1975b, pp.
18-20; Robbins, 1934; Rothbard, 1963). The fact that there was virtually
no price inflation is irrelevant. There is no scope for the idea that the
monetary expansion simply accommodated real growth: real growth, in the
Austrian view, must be accommodated by real saving. The credit expansion
during the 1920s caused the rate of interest to be lower than it otherwise
would have been and thereby triggered a self-reversing process within the
market for capital goods. The actual self reversal came in 1929 causing
the economic downturn. This historical episode, then, is an illustration
of Hayek's story and not an exception to it.
Interpretations of history
have their counterpart in policy recommendations. Again, the differing
significance of price inflation is the basis for differences in preferred
policies. In the Monetarist view, so long as the price level is stable,
monetary expansion is not disruptive. Monetary expansion may even be necessary
to keep prices from falling during periods of real economic growth. In
the Austrian view, monetary expansion is a disruptive force, whether or
not the price level is changing as a result of the expansion. The particular
nature of the disruption will depend upon the particular form of the expansion.
The differing recommendations can be stated concisely in terms of the equation
of exchange and an assumed constant velocity of money. The Monetarist recommendation:
Increase the supply of money to match long-term, secular increases in real
output; the Austrian recommendation: Abstain from monetary expansion even
in periods of economic growth; increased credit should come only from increased
saving; increasing output should be accommodated by a declining price level.(34)
These policy differences suggest that the critical comparison of alternative
monetary dynamics is more than an idle exercise.
V. Further Implications
Our critical comparison has important implications about the way we
think about macroeconomic problems. The monetary dynamics envisioned by
Hayek provide a richer understanding of the market processes that might
be triggered by a monetary expansion, but the Austrian alternative may
at the same time render less serviceable—or even misleading—some of the
standard macroeconomic concepts. At issue, in particular, are the conventionally
defined categories of unemployment, the notion of "long and variable" monetary
lags, and the concept of "full" employment.
1. Cyclical and Structural Unemployment
Beginning with Keynes, it has been standard practice to identify a
number of categories of unemployment in order to isolate conceptually one
particular component of special interest. Cyclical unemployment is the
major focus of macroeconomic and monetary theory. Traditionally, theorizing
about what causes unemployment of this category or about how it might be
reduced or eliminated have been facilitated by impounding other categories
of unemployment in a ceteris paribus assumption. Frictional unemployment,
which is inherent in any market economy, is wholly attributable to the
existence of search costs. Unemployed workers of this category always find
employment, but not instantaneously. Institutional unemployment, such as
might be caused by minimum-wage legislation, may be lamentable, but it
is to be fully accounted for in terms of the legal constraints. These two
categories of unemployment can be conceptually separated from cyclical
unemployment, whether we accept Friedman's treatment of monetary dynamics
involves a geographical or occupational mismatch of workers and employment
opportunities. Unemployment of this sort could result from autonomous
shifts in consumer demand or from technological innovations. In Keynesian
and Monetarist formulations, structural unemployment is combined with frictional
and institutional unemployment, and all three are collectively impounded
in a ceteris paribus assumption. Keynes's views on unemployment
of these sorts are virtually identical to the "Classical" views of, say,
Cecil Pigou; they were spelled out by Keynes (1936, p. 6) only to make
clear what he was not talking about. Friedman (1976, p. 217) makes
use of these categories to locate the vertical long-run Phillips curve.
The natural rate of unemployment,
which is simply the market rate given frictions, mismatches, and institutional
constraints, serves as the base point from which to analyze cyclical unemployment.
Operationally, this last category is defined as a residual. Cyclical unemployment
is calculated by subtracting an estimation of the natural rate from the
measured unemployment rate. Friedman's story begins with an economy in
which all unemployed workers are "naturally" unemployed. Initially, then,
a monetary expansion gives rise to negative cyclical unemployment,
which persists until employer/worker wage-perception differentials are
eliminated. In symmetrical fashion, a monetary contraction—or disinflation—causes
the actual unemployment rate temporarily to exceed the natural rate. Eventually,
the natural rate is reestablished, but—more significant for the present
discussion—the rate of structural unemployment remains constant all the
while. The market process by which the economy deviates from and then returns
to the long-run Phillips curve creates no mismatches between workers and
employment opportunities. Money is neutral with respect to the structure
of the economy.
Hayek's story can begin
at the same point as Friedman's, but once the economy begins to react to
the monetary expansion, the strict dichotomy between structural unemployment
and cyclical unemployment can no longer be maintained. The self-reversing
process plays itself out in terms of changes in the structure of capital
and corresponding changes in the structure of employment. Friedman's initial
increase in employment becomes, in Hayek's story, an increase of some particular
kinds of employment at the expense of other particular kinds. If the new
money enters the economy through credit markets, the "forced saving," as
Hayek uses that term, is financing production for the more remote future
at the expense of production for the more immediate future. The structure
of employment opportunities is modified accordingly. The unemployment that
characterizes the later phase of the dynamic process, then, is structural
unemployment. It differs from the structural unemployment that existed
prior to the monetary expansion only in terms of the causal factors. But
operationally, structural unemployment caused by autonomous changes in
tastes and technology and structural unemployment caused by monetary disturbances
may not be separable categories of unemployment. (Does the waning of smokestack
industries reflect a technological shift towards an information-based economy,
or does it represent a structural hang-over from an earlier money-induced
structural unemployment is likely to be long-lasting. The long run in Hayek's
formulation must be long enough so that all misallocated capital can be
liquidated and all capital/labor mismatches can be rectified. The amount
of time required may be so great as to make any propositions about long-run
neutrality highly misleading. (It would do violence to standard macroeconomic
terminology to categorize the Great Depression as an instance of "short-run
Hayek (1977, p. 282; 1975a,
p. 44) does allow for the possibility that some cyclical unemployment may
not be structural unemployment. The reduction in incomes during the downturn
can—through the reduction in effective demand—have an aggravating effect
on the problem of unemployment. This is what Hayek refers to as the "secondary
deflation," or the "cumulative process of contraction." The unemployment
associated with the economy's spiraling downwards is the type of cyclical
unemployment dealt with by Keynes. Hayek's recognizing the possibilitly
of a secondary deflation does not, as some have suggested, constitute a
capitulation to Keynes. In fact, this problem was put into perspective
(1975b, p. 19) well before the appearance of The General Theory.
Rather, the fact that Hayek sees the Keynesian component of cyclical unemployment
as a secondary effect draws attention to the primary effect which Keynes
2. The Long and Variable Lag
We turn now from the nature of the unemployment that is caused by a
monetary disturbance to consider the length of time between the initial
injection of new money and the economy's eventual adjustment to it. As
an empirical summary, the Monetarist phrase "long and variable lag," (Friedman,
1969a, p. 238) is appropriate for both Friedman's and Hayek's account of
the monetary injection and its ultimate effect. But further reflection
on the two views of monetary dynamics reveals important differences in
this respect. First, the "ultimate effect," which marks the end of the
lag, refers to different things in the two views. As suggested above, the
overall level of prices may become fully adjusted to the increased quantity
of money long before the money-induced distortions in the structure of
capital (and labor) are fully eliminated. Thus, the short-run and long-run
Phillips curves may be separated by a much shorter span of time than the
short-run and the long-run Hayekian triangles.
Second, the basis for the
lag differs between the two views in an important way. Friedman's lag depends
upon how long it takes for workers to straighten out their misperceptions
of the real wage; Hayek's lag depends upon characteristics of the capital
structure—the degree of specificity and intertemporal complementarity of
the capital goods that make up the structure. Hayek's story suggests that
credit expansion is less disruptive (and involves a shorter adjustment
lag) in underdeveloped countries than in industrialized countries. This
accords with casual empiricism. Does Friedman's story suggest that workers
in underdeveloped countries have better and/or more quickly adjusting perceptions
of their real wage rates in comparison to workers in industrialized countries?
Third, there is an important
difference in the basis for the variability of the lag from one expansionary
episode to another within a given economy. For Friedman, the length of
the lag, being based on workers' misperceptions, is fundamentally indeterminate.
But it is not at all clear why it should take workers so long to straighten
out their misperceptions about the real wage and why it should take much
longer in some episodes than in others.(35)
For Hayek, the length of the lag will depend upon the particular way in
which capital and labor markets are distorted, which in turn depends upon
how the new money is injected. Monetary injections that discoordinate intertemporally
are more likely—precisely because of the temporal nature of the discoordination—to
involve longer lags than injections that discoordinate in some atemporal
way. This relationship between the method of injection and the likely length
of the lag helps to explain why the Austrian theorists have always seen
credit expansion, which was characteristic of the 1920s' boom, as particularly
troublesome, and why, in their study of this and other expansionary episodes,
they are as much or more interested in how—rather than how much—money was
3. "Full" Employment
Finally, we turn to the concept of "full" employment as used by Keynes
and Friedman in the light of the monetary dynamics as envisioned by Friedman
and Hayek. For Keynes, full employment is achieved so long as investors
are sufficiently optimistic or sufficiently moved by the "animal spirits,"
or so long as public works takes up the slack created by any insufficiency
of private spending. But the very nature of a self-reversing dynamic process—as
incorporated into the visions of both Friedman and Hayek—suggests that
the critical issue is whether or not a particular pattern of employment
is sustainable. In either vision any level of employment above the
full-employment level is clearly not sustainable. But in Friedman's
formulation, full-employment is—by construction—a sustainable level of
For Hayek, the level of
employment that corresponds to Friedman's full-employment may be sustainable
or it may contain the seeds of its own undoing (1975c, pp. 60-62). That
is, even though the real wage rate consistent with the underlying real
factors is correctly perceived by both employers and workers, the existing
capital structure may be inconsistent with the intertemporal pattern of
consumer demand and resource availabilities. The market process through
which these inconsistencies are discovered and remedied may involve a considerable
period of cyclical (structural) unemployment. Again, the circumstances
existing in the late 1920s can illustrate the distinction. In the Austrian
view, the economy was experiencing, in those years, unsustainable full
Friedman does recognize
that irresponsible monetary policies may eventually increase the natural
rate of employment. In two different pieces of analysis (1976, pp. 232-33
and 1977, pp. 459-60), he suggests the possible existence of a "positively
sloped Phillips curve." But in each case, new considerations not integral
to his story of the adjustment process are introduced to account for such
a possibility.(36) In neither case are
structural considerations—as conceived by Hayek—integrated into the dynamic
process envisioned by Friedman. Friedman's strict dichotomization of structural
and cyclical unemployment stands in the way of his recognizing the possibility
of full (in the operational sense) but unsustainable (in the Hayekian sense)
employment. Hayek's formulation, which involves an interplay between structural
and cyclical unemployment, allows for the recognition of such possibilities
and for the prescription of policy most conducive to sustainable full employment.
VI. A Summary View
Purely as an instrumentalist's device, it could be argued, Friedman's
story has served the Monetarists well. It put them onto the idea that the
Phillips curve trade-off is a short-run trade-off only. After this idea
was confirmed empirically (or properly, after it failed to be falsified
over many trials) the story itself became superfluous. The realization
that the empirically defined long-run Phillips curve is vertical is enough
to carry the day. It provides a strong basis for arguing that in the long
run, "nominal magnitudes do not influence real magnitudes" and for warning
against all attempts to exploit the trade-off offered by the short-run
Friedman's story and Hayek's
story differ sharply in methodological terms. Because of its instrumentalist
qualities, Friedman's story fares poorly as a source of insights about
the market process that translates a monetary injection into its ultimate
consequences; it fails to increase our understanding of any actual market
process. Hayek's story identifies the essential workings of the self-reversing
market process triggered by a monetary expansion in a way that sheds light
on the structure and timing of the pattern of unemployment caused by such
monetary disturbances. The Austrian insights contained in the Hayekian
triangles square with but go beyond the empirical regularities of Monetarism.
From a broader perspective,
Friedman's story represents a recognition of the intertemporal substitutability
of labor and of the possibility that monetary disturbances can interfere
with the intertemporal allocation of labor. Hayek's story represents a
recognition of the intertemporal complementarity of capital and of the
possibility that monetary disturbances can interfere with the intertemporal
allocation of capital. Viewed as such, the two stories are themselves not
substitutes, but complements.
Ackley, Gardner 1978. Macroeconomics: theory and policy.
________ 1983. "Commodities and capital: prices and quantities."
Economic Review, 73.1 (March):1-30.
Bellante, Don 1983. "A subjectivist essay on modern labor
economics." Managerial and Decision Economics, 4.4:234-43.
Birch, Dan, Alan Rabin, and Leland Yeager 1982. "Inflation,
output, and employment: some clarifications." Economic Inquiry,
Blaug, Mark 1978. Economic Theory in Retrospect,
3rd ed. London.
Böhm-Bawerk, Eugen 1959. Capital and Interest,
vol. 2 (1889). South Holland, IL.
Boland, Lawrence A. 1979. "A critique of Friedman's critics."
of Economic Literature, 17.2 (June):503-22.
Brimelow, Peter 1982. "Talking money with Milton Friedman."
62.43 (Oct. 25):5-6.
Butos, William 1985. "Hayek and general equilibrium analysis."
Economic Journal, 52.2 (Oct.)332-43.
Clower, Robert W. 1969. "The Keynesian counter-revolution:
a theoretical appraisal" (1965). In Robert W. Clower, ed., Monetary
Darby, Michael 1976. Macroeconomics. New York.
Friedman, Milton 1953. "The methodology of positive economics."
In Milton Friedman, Essays in Positive Economics. Chicago.
________ 1969a. "The lag in effect of monetary policy."
(1956) In Milton Friedman, The Optimum Quantity of Money. Chicago.
________ 1969b. "Money and business cycles" (1963). In
Milton Friedman, The Optimum Quantity of Money. Chicago.
________ 1969c. "The optimum quantity of money." In Milton
Friedman, The Optimum Quantity of Money. Chicago.
________ 1969d. "The quantity theory of money: a restatement"
(1956). In Milton Friedman, The Optimum Quantity of Money. Chicago.
________ 1969e. "The role of monetary policy" (1968).
In Milton Friedman, The Optimum Quantity of Money. Chicago.
________ 1970. The Counter-Revolution in Monetary Theory.
________ 1976. "Wage determination and unemployment."
In Milton Friedman, Price Theory. Chicago.
________ 1977. "Nobel lecture: inflation and unemployment."
of Political Economy, 85.3 (June):451-72.
Friedman, Milton and David Meiselman 1963. "The relative
stability of monetary velocity and the investment multiplier in the United
States, 1897-1958." In Stabilization Policies. London.
Friedman, Milton and Anna Schwartz 1963. A Monetary
History of the United States, 1867-1960. Princeton.
________ 1982. Monetary Trends in the United States
and the United Kingdom. Chicago.
Garrison, Roger W. 1984. "Time and money: the universals
of macroeconomic theorizing." Journal of Macroeconomics, 6.2 (Spring):197-213.
________ 1985. "Intertemporal coordination and the invisible
hand: an Austrian perspective on the Keynesian vision." History of Political
Economy, 17.2 (summer):309-21.
________ 1986. "Hayekian trade cycle theory: a reappraisal."
Journal 6.2 (Fall): 437-53.
Hicks, John R. 1946. Capital and Value, 2nd ed.
________ 1967. "The Hayek story." In John R. Hicks, Critical
Essays in Monetary Theory. Oxford.
Hayek, Friedrich A. 1941. The Pure Theory of Capital.
________ 1948a. "The Ricardo effect" (1942). In Friedrich
A Hayek, Individualism and Economic Order. Chicago.
________ 1948b. "The use of knowledge in society" (1945).
In Friedrich A. Hayek, Individualism and Economic Order. Chicago.
________ 1967. Prices and Production, 2nd ed. (1935).
________ 1975a. Full Employment at Any Price, Occasional
Paper 45. London.
________ 1975b. Monetary Theory and the Trade Cycle
(1933). New York.
________ 1975c. Profits, Interest, and Investment
(1939). Clifton, NJ.
________ 1977. "Three elucidations of the Ricardo effect."
of Political Economy, 77 (Mar./Apr.):274-85.
________ 1984. "The future monetary unit of value." In
Barry N. Siegel, ed. Money in Crisis. Cambridge, MA.
Hoover, Kevin D. 1984. "Two types of monetarism." Journal
of Economic Literature, 22.1 (Mar.):58-76.
Jevons, William S. 1970. The Theory of Political Economy
Kaldor, Nicholas 1942. "Professor Hayek and the concertina
effect." Economica, ns 9(Nov.):359-82.
Keynes, John M. 1936. The General Theory of Employment,
Interest, and Money. New York.
Knight, Frank H. 1934. "Capital, time, and the interest
rate." Economica, ns 2 (Aug.):257-86.
Lachmann, Ludwig M. 1978. Capital and Its Structure
(1956). Kansas City.
Leijonhufvud, Axel 1968. On Keynesian Economics and
the Economics of Keynes. New York.
Lucas, Robert E. Jr. 1981. Studies in Business Cycle
Theory. Cambridge, MA.
Mises, Ludwig von 1953. The Theory of Money and Credit.
New Haven, CT. Originally published in German in 1912.
Musgrave, Alan 1981. "'Unreal assumptions' in economic
theory: the F-twist untwisted." Kyklos, 34.3:377-87.
O'Driscoll, Gerald P. Jr. 1977. Economics as a Coordination
Problem: The Contribution of Friedrich A. Hayek. Kansas City.
O'Driscoll, Gerald P., Jr. and Mario J. Rizzo 1985. The
Economics of Time and Ignorance. Oxford.
Phelps, Edmund S. 1970. "Money wage dynamics and money
market equilibrium." In Edmund Phelps et al., ed., Microeconomic Foundations
of Employment and Inflation Theory. New York.
Reder, Melvin W. 1982. "Chicago economics: permanence
and change." Journal of Economic Literature, 20.1 (Mar.)1-38.
Robbins, Lionel 1934. The Great Depression. London.
Robertson, Dennis H. 1949. Banking Policy and the Price
Level, rev. ed. New York.
Robinson, Joan 1972. "The second crisis in economic theory."
Economic Review. 62.2(May):1-10.
Rothbard, Murray N. 1975. America's Great Depression (1963).
Wainhouse, Charles E. 1984. "Empirical evidence for Hayek's
theory of economic fluctuations." In Barry N. Siegel, ed., Money in
Crisis. Cambridge, MA.
Wagner, Richard E. 1979. "Comment: politics, monetary
control, and economic performance." In Mario J. Rizzo, ed., Time, Uncertainty,
and Disequilibrium. Lexington, MA.
Warburton, Clark 1966. Depression, Inflation, and Monetary
Policies: Selected Papers, 1945-1953. Baltimore.
Wicksell Knut 1936. Interest and Prices (1898).
1. Two related contributions are Hoover
(1984), who provides an insightful comparison between Old Monetarists in
the style of Friedman with the New Classicists in the style of Lucas, and
Butos (1985) who compares Hayek and Lucas in the context of a general-equilibrium
approach to business-cycle analysis. Although not originally conceived
for this purpose, the present paper can be seen as completing the trilogy
by comparing Hayek and Friedman on the issue of monetary dynamics.
2. We should note that the sharpness
of the labor-market/capital-market distinction that characterizes our paper
is not intended to suggest that Hayek has had nothing to say about labor
markets or that Friedman has had nothing to say about capital markets.
Hayek (1975a, pp. 15-29) specifically addresses the relationships among
"Inflation, the Misdirection of Labour, and Unemployment"; Friedman (1970,
p. 24-25) includes the existence of a cash-balance effect on asset prices
as one of the key propositions of Monetarism. Our paper focuses more narrowly
on a comparison of Phillips Curves and Hayekian Triangles as alternatives
bases for theorizing about monetary dynamics.
3. The article exhibiting Friedman's
agnostic stance was first published in 1963; the "engineering blueprint"
appeared in 1976 after having been presented at the Southern Economic Association
meetings two years earlier. Some might argue that a dozen or so years is
enough to transform vagueness into a blueprint. An alternative view is
that the agnosticism persists (see, for instance, Friedman, 1969c, p. 6);
the arguments in the later effort, which have their roots in his 1967 A.E.A
Presidential Address (1969e), are to be understood as Friedman's willingness
to take on his adversaries on their own turf. If this view is valid,
a distinction must be made between Friedman and his followers. Some Monetarists
(e.g. Darby, 1976, pp. 328-50) write as if Friedman's critical analysis
which demonstrates the absence of an exploitable long-run Phillips curve
is at the same time an exposition of the Monetarist view of the market
process which transforms monetary injections into increases in the price
level. Also, the formal analysis offered by Phelps (1970. pp. 124-66) is
similar to Friedman's critical analysis of the relationship between long-run
and short-run Phillips curves.
4. In other contexts, Friedman opts
for a more broadly conceived adjustment process in which asset prices change,
but he trivializes the corresponding quantity adjustments (e.g. that may
temporarily alter the structure of capital) as "first-round effects." See
Friedman and Schwartz (1982, ch. 2) and Friedman and Meiselman (1963, pp.
5. Reder (1982) could be interpreted
as denying the significance in this respect of Knightian capital theory
for the development of the Chicago tradition. "The contribution to economic
thought with which Knight is most readily identified (theory of the firm,
uncertainty and profit, capital theory, social cost, etc.) are only
tangentially related to the Chicago tradition" (Reder, 1982, p. 6, our
emphasis). An anonymous referee has suggested that Reder's assessment is
compatible with our own: "The modern Chicago school interpreted Knight
so as to allow them to ignore capital theory."
6. See Friedman (1969c, pp. 4-7). Friedman
clearly recognizes that the money drop has to be perceived as a one-time
event. "Let us suppose further that everyone is convinced that this is
a unique event which will never be repeated." This essential assumption
separates Friedman from the New Classicists in terms of the issues that
each is addressing. The money drop in Friedman's analysis does not constitute
a "policy" in the New Classicist view. Further, Friedman's treatment of
expectations creates a certain internal tension in his own view if his
analysis is to explain more than a single episode of monetary expansion:
Market participants have adaptive expectations (about the price
level and real wages) during the period that the market is adjusting to
a money drop, but static expectations about the likelihood of a future
7. As an alternative formulation, Friedman
does away with both the distribution effects and the differential income
effects by assuming "infinitely lived people" such that no transitory event
has any effect on permanent income (Friedman, 1969c, p. 6n). This alternative,
however, does not represent a net "relaxation" of simplifying assumptions;
it represents the replacement of one assumption with another, analytically
8. The exposition that follows draws
primarily from Friedman, 1976, pp. 213-37.
9. As Friedman makes clear, the shifting
of and movements along the curves are reversed if we take the employer's
point of view. The movement down the employer's demand for labor corresponds
to an apparent shift in the supply of labor (Friedman, 1976, p. 223).
10. The exposition of monetary dynamics
as seen by Hayek draws largely from his early writings (Hayek, 1967 and
1975b). Even the more advanced of these two books was intended only as
an outline (1967, p. vii). Hayek's monetary dynamics is based upon a theory
first set out by Mises (1953, pp. 339-66), which integrated the capital
theory of Böhm-Bawerk with a theory of interest-rate movements adapted
from Wicksell (1936).
11. The kernel of truth in the idea
of rational expectations is clearly recognized by Mises as early as 1953
as is demonstrated by the following passage on inflationary finance: "Here
the famous dictum of Lincoln holds true: You can't fool all of the people
all of the time. Eventually the masses come to understand the schemes of
their rulers. Then the cleverly concocted plans of inflation collapse....
[I]nflationism is not a monetary policy that can be considered as an alternative
to a sound money policy. It is at best a temporary makeshift. The main
problem of an inflationary policy is how to stop it before the masses have
seen through their rulers' artifices. It is a display of considerable naivete
to recommend openly a monetary system that can work only if its essential
features are ignored by the public" (Mises, 1953, p. 419). For a comparison
of Hayek's approach to the issue of expectations with that of New Classicism,
see Garrison (1986).
12. For a treatment of Austrian capital
theory that emphasizes these features, see Lachmann (1956).
13. Jevons (1970, pp. 229-36) had
earlier employed a triangular construction for similar purposes. Attempting
to characterize the economy's capital structure in terms of the dimensions
of a triangle is what qualifies Prices and Production as an outline.
In Hayek's more formal—and more formidable—Pure Theory of Capital
(1941), many of the heuristic assumptions of his earlier efforts were relaxed.
This volume was to serve as the basis for a more comprehensive treatment
of monetary theory (Hayek, 1941, p. v-vi), but no follow-on volume was
14. At the time of Hayek's London
lectures, which were to become Prices and Production, this formulation
was largely unintelligible to his British audience. Hicks (1967, p. 204)
was later to note that "Prices and Production was in English, but
it was not English economics." While the neo-Austrian Hicks pointed to
difficulties in Hayek's communication of the message, Joan Robinson, who
had difficulties with the message itself, dubbed the Hayekian episode a
"pitiful state of confusion" and categorized it as a crisis in economic
theory (Robinson, 1972, p. 2).
15. See Hayek, 1967, p. 54). In his
earlier writing (1975b, pp. 143-48) Hayek criticized L. A. Hahn and Ludwig
von Mises for beginning their stories with "arbitrary interferences on
the part of banks." Hayek regarded an exogenous credit expansion as a "specially
striking case," and as a sufficient condition but not a necessary one for
the occurrence of alternating boom and crisis. He did note (p. 150) that
"it is possible to assume, with Professor Mises, that the Central Banks,
under the pressure of an inflationist ideology, are always trying to expand
credit and thus provide the impetus for a new upward swing of the Trade
Cycle; and this assumption may be correct in many cases."
16. This aspect of the market process
reflects the insights of Wicksell (1936, pp. 102-21): a economic boom grows
out of the divergence between the loan rate of interest and the natural
rate. But as an anonymous referee has noted, Hayek differed from Wicksell
on the question of what causes the divergence of the two rates. Wicksell
believed that the natural rate increases first due to technological advance
and that the loan rate adjusts but with a lag.
17. This aspect of Hayek's theory
is the central focus of his 1937 article, "Investment that raises the demand
for capital." (See Hayek, 1975c, pp. 73-82.)
18. Early on, Hayek was acutely aware
that his own monetary theory of the trade cycle was distinctive in that
it did not hinge on changes in the general level of prices: "We are in
no way concerned to explain the effect of the monetary factor on trade
fluctuations through changes in the value of money and variations in the
price level..." Instead, his explanations hinged upon "changes [monetary
in origin] which are bound to disturb the equilibrium inter-relationships
existing in the natural economy, whether the disturbance shows itself
in a change in the so-called 'general value of money' or not" (1975b.
19. Hayek uses the term "Ricardo Effect"
(1948a) to refer to the quantity adjustments in markets for capital goods
brought about by this upturn in interest rates. Also, see O'Driscoll (1977,
20. Richard Wagner (1979, pp. 182-85)
demonstrates that the applicability of Austrian business-cycle theory is
independent of any unemployed labor that may be associated with the downturn
and that to estimate the social costs of a monetary disturbance by the
unemployment caused by it is to underestimate those costs.
21. As Hayek (1967, p. 33) clearly
recognized, this same distinction separated his own theorizing (1967) from
the contemporaneous theorizing of Dennis Robertson (1949).
22. The claim that capital is more
radically heterogeneous than labor is not based simply on a difference
in degree. Different man-hours of labor are not perfectly substitutable
for one another. But our attempt to construct an analogous claim for capital
is revealing: Different ________ of capital are not perfectly substitutable
for one another. The difficulty of even filling in the blank is a clear
sign of dimensional, or radical, heterogeneity. We must resort
either to some universal non-unit, such as "hunks," or to some value measure,
such as "dollar's worth," which leads to an inescapable conflation of the
price of capital and its corresponding quantity. Further, differences between
different man-hours of labor are conventionally attributed to differences
in human capital. Theories about labor that hinge on such differences
are faced with the same thorny problems that are inherent in theories about
capital. It is largely because of these well-recognized problems that most
monetary theorists have preferred to think in terms of homogeneous, perfectly
23. Hayek recognized that some
capital goods are substitutes for one another. We emphasize the notion
of capital complementarity to contrast Hayek with most other theorists,
who, explicitly or implicitly, take all capital goods to be substitutes.
24. For an alternative treatment of
the relationship between labor-based and capital-based theories, see O'Driscoll
and Rizzo, 1985, pp. 160-87.
25. These critics would not take Hayek
to be the most viable alternative to Friedman. They would opt instead for
theories of monetary disequilibrium in the style of Warburton (1966) and
Clower (1969). Their preference is based upon a perceived asymmetry between
price adjustments associated with a falling price level and price adjustments
associated with a monetary injection. "Deflation has to work through a
sequence of millions of piecemeal price and wage decisions; the alternative
of nominal monetary expansion puts no such demands on the economy's coordinating
mechanism" (Birch, Rabin, and Yeager, p. 213). But the asymmetry is non-existent
unless the question-begging assumption of "helicopter money" is employed.
Actual deflations and actual monetary expansions "[have] to work through
a sequence of millions of piecemeal price and wage decisions."
26. For an instrumentalist defense
of Friedman's economics, see Boland (1979).
27. Musgrave (1981) uses the modifiers
"negligibility," "domain," and "heuristic" to define three categories of
assumptions that underlie various theoretical constructions. Generically,
a domain assumption identifies the domain, or scope, of direct applicability;
a heuristic assumption is a deliberate fiction used to facilitate the logical
development of a theory. Musgrave uses these distinctions to evaluate Friedman's
treatment of methodological issues (1953) in which assumptions are categorized
as "realistic" or "unrealistic."
28. For a treatment of Hayek's analysis
in terms of the Phillips curve, see O'Driscoll (1977, pp. 115-18). It should
be noted that O'Driscoll uses Phillips' original construction, which has
wage inflation on the vertical axis; the analysis is applicable whether
or not there is price inflation as well.
29. It may be misleading to claim
that Hayek generalized his theory only in recent years. The case involving
the expansion of credit to producers provided a focus for Prices and
Production where Hayek "concentrated on the successive changes in
the real structure of production, which constitute [cyclical]
fluctuations." In his earlier work where he was concerned with "the monetary
causes which start the cyclical fluctuations," his arguments
were much more general in nature. (See, for instance 1975b, pp. 91, 146-47,
and passim; the quoted passages are from the preface, p. 17.) Modern
Hayekians may even judge that Hayek generalized a little too much: "The
initial change [eliciting a cyclical fluctuation] need have no specific
character at all, it may be any one among a thousand different factors
which may at any time increase the profitability of any group of enterprises"
(1975b, pp. 182-83).
30. Drawing from Hicks (1946, p. 222),
Leijonhufvud (1968, p. 222) contends that the Austrian construction cannot
be generalized in this way: "the result reached for the point-input point-output
case ... 'does not generalize in the sort of way in which it might have
been expected to generalize.'" But the obstacles that make generalizing
difficult are not to be attributed to the Austrian formulation but to the
very nature of capital and hence of capital theory. (See footnote 20 above.)
If measured in terms of the "period of production," the time element in
the production process depends upon the "quantity" of capital employed
at each point in time; the quantity of capital must be measured in value
terms, which in turn depend upon the rate of interest. Thus, the "lengthening"
of the period of production brought about by a fall in the rate of interest
may involve a "lengthening" that follows from the very definition of the
period of production plus a "lengthening" that results from a market process
set into motion by the fall in the rate of interest. Hayek (1941, pp. 76-77,
140-45, 191-92 and passim) was aware of the hoary problems of this sort
well before the so-called "reswitching" debate of the 1960s. The fact that
the Austrian formulation is capable of making some important distinctions,
such as between the two senses of "lengthening," should be seen as a credit
to that theory and not as grounds for condemning it.
31. For modern treatments of Hayek
that rely heavily on Kaldor's critique, see Blaug (1978, pp. 571-74) and
Ackley (1978, p. 632). Ackley refers to the Austrian view as a fantasy
and is particularly vitriolic in his treatment of it. 'In what must be
one of the most devastating critiques in the history of economics, Nicholas
Kaldor showed the absurdity of what he called Hayek's "concertina effect"—the
cyclical "lengthening" and "shortening" of the production process. Remnants
of this idea may still be circulating in the byways of economics, but the
writer has not encountered them for several decades.'
32. Wainhouse (1984), whose comparison
of Friedman and Hayek is compatible with the one offered in the present
paper, employs "Granger causality tests" in an attempt to establish the
empirical relevance of Hayek's story.
33. In recent years Friedman has been
attributing high interest rates, not to an inflation premium, but to the
volatility of money growth (Brimelow, 1982, p. 5-6). And his argument sounds
more Hayekian than Friedmanian: In 1980 the Federal Reserve began a relatively
rapid monetary expansion. Businessmen assumed the cheap credit was going
to last for a considerable period, but when it didn't, "you had the business
community left with all sorts of commitments based on a wrong diagnosis,
a wrong prediction. And you had a great deal of distress borrowing that
was highly insensitive to interest rates."
34. This policy prescription observes
the stricture of separating policy implications from judgments of political
feasibility. There is no evidence that Hayek has changed his mind about
the implications of his theory, but in recent years he has tempered his
recommendations. Prefacing an argument for a stable price level (1975a,
pp. 26), Hayek "confess[ed] that 40 years ago I argued differently. I have
since altered my opinion—not about the theoretical explanation of the events
but about the practical possibility of removing the obstacles to the functioning
of the system by allowing deflation to proceed...." And looking to future
monetary systems, Hayek (1984, pp. 325-26) notes that "Sixty years ago
I began my work on monetary theory by questioning the belief, then universally
accepted, that the purchasing power of money should be kept stable. I then
suggested that it was more desirable for the purchasing power of money
over consumer goods to increase over time. But I have since become convinced
that a money of stable value is really the best we can hope for."
35. When Friedman (1969a, pp. 255-56)
explicitly considers the question of why the lag should be so long, he
gives a Hayekian-triangle answer rather than a Phillips-curve answer. The
process of monetary expansion effects the demand for "equities, houses,
durable producer goods, durable consumer goods, and so on.... The effects
can be described as operating on 'interest rates,' if a more cosmopolitan
interpretation of 'interest rates' is adopted than the usual one which
refers to a small range of marketable securities." This same process later
generates "reactions [which] undo the initial effects on interest rates."
36. The positive slope is attributed
in one piece of analysis (1976, p. 233) to "slow adjustments of anticipations
to experience, long-term contracts, government interventions into specific
markets, and other elements of 'friction' or 'rigidity'...."; and in the
other piece of analysis (1977, p. 460) to "the interdependence of experience
and political developments."