Leijonhufvud, A. Information and Coordination Cap. 9 y 10

INFORMATION
AND COORDINATION
Essays in Macroeconomic Theory
AXEL LEIJONHUFVUD
New York Oxford

OXFORD UNIVERSITY PRESS
1981
CHAPTER NINE
Costs and Consequences
of Inftation *
l. INTRODUCTION
One approach to the problem of the microeconomic founda-
tions of macro-economics takes the frame and the components
of standard "neoclassical" theory as the given starting point.
One asks what can be used and what needs modification for
purposes of representing the movement of a macro-system
through time and into a future that is in sorne respects un-
knowable. The aim is to define and, if possible, solve the analyt-
ical problems that emerge at the levels of individual conceptual
experiments, market experiments, and general equilibrium ex-
periments. I have pursued this approach in other recent
papers 1 but am running into diminishing returns.
An alternative approach is to start from the other end with
sorne "applied" problem, preferably one of such importance
that no macroeconomist can really afford to dodge it, and con-
sider the difficulties that arise in trying to handle it in a "rea-
sonable" way with standard micro-theoretical tools. From this
viewpoint we get a different critica] angle on the problems
requiring solution if micro- and macro-theory are to be made to
mesh. This is the approach taken in this paper.
T he "practica!" macro-question to be considered here is that
of the social costs and consequences of inflation. A new view of
the welfare costs of inflation has emerged in the last ten or fif-
* I am thankful to Armen Alchian, Roben Clower, Ben Klein, John McCall and Sidney
Afriat for comments and obliged to d eclare them free from responsibility. Financia!
support of the Liberty Fund is gratefully acknowledged.
l. Leijonhufvud (1974, 1974-5).
227
INFORMATION AND COORDINATION
teen years. It trivialises the cost of inflation. This new view is

undergirded by essentially "neoclassical" theoretical construc-
tions and may, indeed, be regarded as a by-product of work
primarily oriented toward seeking neoclassical foundations for
macro-theory. In the analytical exercise that is central to this
view, inflation is treated as a foreseen tax on money balances
and its costs are seen to líe in the productive and transactional
inefficiencies induced by su ch a tax. Even a quite high rate of
inflation will not imply a very sizable tax as taxes go in modero
mixed economies; the inefficiencies that it may induce will be
correspondingly trivial.
Sorne economists will feel that this work has helped u s put
the undesirability of inflation into proper perspective by dis-
pelling old and murky myths about its dangers. To those, my
tapie will not seem a promising avenue towards a fuller under-
standing of the trouble we are having with microeconomic
foundations.
It should thus be obvious and shall in any case be openly ad-
mitted that my choice of tapie is predicated on the prior convic-
tion that in advocating or letting go unopposed this new view of
inflation we have been guilty of profound and appalling
na"iveté. I fear that the spreading influence of the new view is
dangerous in so far as it directly or indirectly influences policy.
The new view on inflation is not altogether unassailable on its
own terms. But the questions about it that may be raised strictly
within the neoclassical framework are probably not the impor-
tant ones. Neoclassical theory-or, more precisely, its scope-is
itself at issue. The social consequences of inflation most ger-
mane to "wise" conduct of economic policy may fall largely out-
side its purview. For this once, I do not think inside ("im-
manent") criticism is the tack to take. This p aper wilfu lly
refuses obedience to the neoclassical rules of the game. We
begin by taking an "institutionalist" view of monetary exchange.
T he institutional approach has, of course, its own limitations.
One cannot be perfectly "general" (i.e., refer to all times and a li
places) and still retain content. The time-space reference co-
ordinates that I have had in mind in writing this paper are (i)
the last ten years or so, and (ii) the United States. Similarly,
the term "inflation" in the title is not to be read as denoting a
228
COSTS ANO CONSEQ UENCES OF INFLATION
theoretically defined "pure" concept but as referring to infla-

tionary processes "like" the one of recent years.
11. AN INSTITUTIONALIST SKETCH

OF MONETARY EXCHANGE
Whether th e tru e idea of money, as such, is not altogether that of a tickel
or counter?
-Bishop Berkeley, The Querist
Sorne of the questions on the present theoretical agenda are

much older than the current movement to provide microeco-
nomic foundations for macro-theory: Why do people hold
money? Why is the set of goods serving as means of payment so
small? Why are "indexed" contracts so uncommon? Etc.
One approach to these questions starts by interpreting the
mathematical structure of a standard general equilibrium
model as representing a multilateral "barter" system. One then
seeks precise formulation of realistic assumptions about infor-
mation imperfections and transactions costs that can be shown
to lend a "monetary" transactions structure to the G.E. model.
It is not part of my aim to criticise this research, much of which
I find interesting and promising.
The point to be made here is simply that these conceptual ex-
periments should not be given historical interpretations. The
proposition that "barter is costly and inefficient" will no doubt
be part of any explanation of the "use of money." T hat "the
inefficiency of barter leads to the use of money," would, how-
ever, be false as an historical generalisation. Monetary ex-
change systems have not evolved out of non-monetary ex-
change ("barter") systems but out of non-exchange systems.
Both intertemporal and cross-cultural comparisons show us
that in the spheres of economic activity where monetary ex-
change does not prevail, neither do we find predominantly
"private" property rights, commercial contracting, and orga-
nised markets. (These are however institutional features pre-
sumed by the "non-monetary" G.E. model.) We will still expect
to find a fairly extensive division of labour but the institutional
arrangements- the systems of rights and obligations governing
229
INFORMATION ANO COORDINATION
the activities of individuals-devised to ensure that the commu-

nity can depend on the benefits from the division of labour will
be different in kind. "Custom and Command," in the terms of
Classical Economics, or "Reciprocity and Redistribution," in
those of Anthropology-not barter exchange-are the alterna-
tives to monetary exchange.2 The development of monetary
exchange is, consequently, part of a complex evolution of insti-
tutions. Perhaps the best short statement is Wesley Mitchell's
famous passage: 3
W h en money is introduced into the dea lings of men, it enlarges their
freedom . . . . By virtue of its generalised purchasing power, money
emancipates its users from numberless restrictions upon what they do
a nd what they get. As a society learns to use money confidently, it gradu-
a lly abandons restriction s upon th e places peop le shall live, the occupa-
tions they shall follow, t h e circles they shall serve, t h e prices they shall
ch arge, a nd the goods they ca n buy.
In largely non-monetary economies, important economic rights

and obligations will be inseparable from particularised rela-
tionships of social status and political allegiance and will be in
the same measure permanent, inalienable, and irrevocable. 4 As-
surance of stability of the economic arder is sought in tying eco-
nomic functions to social roles that carry particular rights and
duties vis-a-vis particular individuals or groups. In monetary
exchange systems, in contrast, "the value to the owner of [his
human capital or) a physical asset derives from rights, privi-
leges, powers, and immunities against society generally rather
than from the obligation of sorne par ticular person." 5 And,
paraphrasing J. S. Mill, "competition is the governing principie
of such contracts" as leave particular agents \tÚth a debt-claim
relationship.
2. Ali I can do al this point is to provide a personally favoured select list of "further
reading": J. S. Mili ( 1909) Book 11 , Chapter IV; F. H . Knight (1965) Chapter I; Dalton
(ed ) (1968) esp. Dalton's lntroduction); Hicks (1967, Chapter 9; 1969); Michael Polanyi
( 1969).
3. W. C. Mitchell (1944). For a summar·y of Mitchell's views and further references to
his writings, cf. Friedman (1952).
4. Feudal land-rents, for example, cannot be "decomposed" into a renta! price on land
"plus" a tax on the cul tivator of it; nor can the overlapping rights and interlocking
obligations of a peasant, of other village members, of the manorial lord , and of the sov-
ereign with regard to a particular piece of land be disentangled in terms of modern no-
tions of "ownership. "
5. Cf. Burstein ( 1963), p. l05.
230
COSTS AND CONSEQUENCES OF INFLATION
Neoclassical theories rest on a set of abstractions that separate

"economic" transactions from the totality of social and political
interactions in the system.6 For a very large set of important
problems, this separation "works"-since we are usually dealing
with monetary exchange systems. But it assumes that the events
that we make the subject of conceptual experiments with the
neoclassical model of the "economic system" do not affect the
"socio-political system" so as to engender repercussions on the
economy of such significance as to invalida te the institutional ce-
teris paribus clauses of that model.
It is not "in the nature of things" that this assumption neces-
sarily holds. There can be no epistemological guarantee 't hat in-
teractions between the "economic," the "political," and the "so-
cial spheres" of the system we study will be negligible.
Double-digit inflation may label a class of events for which the
assumption is a bad one. The neoclassical conceptual experi-
ment of a steady-state inflation, which in time becomes accura-
tely foreseen, and to which "everything adjusts"--except prop-
erty rights, contract forms, and the organisation of
markets 7-is at the very least a most instructive exercise. But
that <loes not suffice to make it a good theory. It is a long-run
theory. But its institutional ceteris paribus assumptions may not
hold approximately true for that long.
We should at least keep an open mind to this disturbing pos-
sibility. We do not now have the empirical knowledge to rule it
out. It may be the case that in the world we inhabit, before the
"near-neutral" adjustments can all be smoothly achieved, "soci-
ety unlearns to use money confide ntly" and reacts by restric-
tions on "the circles people shall serve, the prices they shall
charge, and the goods they can buy." 8 If such reactions are in
6. Exactly what ali these abstractions are and what conditions will allow thern validity,
we are not very clear about. We are conlent to live with the correspondingly hazy defi-
nition of the boundaries between economics and other social sciences for, I think, the
simple reason that most of the time our work is shaped by the "economic rnethod" we
use---and letting "the way econornists think" establish the limits to our "territorial irn-
perative" will alrnost always be good enough.
7. E.g,, the line between legal and "black" rnarkets,
8. Historical processes are not reversible, The paraphrase of Mitchell here is not in-
tended to convey sorne silly suggestion of a return to feudalism or even mercantilisrn. It
is in tended to convey the j udgrnent that an analysis of inflation that <loes not atternpt to
take política! feedback on the econornic process systematically into account is, in con-
temporary jargon, "irrelevant."
231
fact endogenous to the social system, we mis-identify the conse-

quences of inflation to the extent that we regard them as fortui-
tous "political" events exogenously impinging on "the econ-
omy."
Mitchell uses the term "money" in a sense so broad as to cover

not just ali of money and banking but also the "Legal Fouhda-
tions of Capitalism" (J. R. Commons) and even the psycho-
logical attitudes and calculating modes of decision-making that
go with life in a society where the range of alternatives subject
to the common measuring-rod of money is very wide. But
something need be said also about how "money," in the narrow
sense of "M," fits into such an institutionalist sch ema.
The stability of any social order requires (i) an exhaustive
and consistent a llocation of rights to economic resources, and
(ii) rules for the transfer of these rights and means for keeping
track of the legitimate succession to them. Disputes over the
possession of rights, where the legal entitlement of the parties
cannot be tracked clown or otherwise "fairly" determined, must
as far as possible be avoided-since the residual method of set-
tling conflicts will be the use of force.
In monetary exchange systems, the problem of keeping track
takes a particular form. The "typical" basic forms of wealth are
defined in terms of rights and immunities vis-a-vis "society in
general." Transactors have discretion in what they choose to
sell and buy and whom they choose to sell to or buy from. The
institutional problem is to ensure that no one takes more out of
the system than he puts in, so that everyone is assured of being
allowed to appropriate resources from the rest of society
"equal" to what he has contributed to others.
Hawtrey's insistence that every transaction generates a claim
and a matching d ebt is h elpful here in leaving ali questions of
settlements temporarily open. T he first problem is the mea-
surement of debts and claims. We may assume them to be re-
corded at the prices in terms of unit of account agreed u pon by
the p arties. In the simplest multilateral exercise, we wou ld have
only "real" transactions-involving the transfer of a physical
asset, real good or service (or the forward contract for such a
transfer)-to consider. Assuming "rules" allowing no financia!
232
COSTS ANO CONSEQUENCES OF INFLATION
transactions or the running of financia! surplus and deficits, a

purely "imaginary money ," 9 tied to no real numéraire good,
could perfectly well serve as unit of account. The conceivable
ways of "policing the rules" are legion.
As an illustration, suppose we find a short closed loop in this
system where repeatedly real resources are transferred from A
to B, from B to C, C to D, and from D to A, and all links in the
chain happen to be quoted by the two parties at the same value
in terms of "imaginary money":
l. We might decide to run a central social bookkeeping office
charged with keeping the respective balance-sheets of A, B,
C, and D continuously up-to-date by adding on the debts in-
curred and claims gained in each period. The object is sim-
ply to check that each balance sheet continues to balance. If
accurate addition is cheap enough to come by, we could as
well !et the balance sheets go on lengthening indefinitely.
Going through the motions of extinguishing debts would be
superfluous.
2. We might feed ali debits and credits arising from resource
transfers into a computer programmed to hunt for "closed
loops" and to wipe out all debts and claims (up to the largest
common numerator) in ali such loops found. Shrill bells
should sound and red lights flash whenever the computer
ends its daily exercise with a residue of net claims, etc. With
this system, debts are systematically "extinguished," putting
less of a burden on central archives, but they are not "paid."
None of the goods in the system is identifiable as the "means
of payment."
3. A social abacus might be cheaper than a central computer.
We might issue little pellets, "tickets or counters" (called
"Berkeleys"), pronounce them legal tender and instruct
every transactor to keep "paying" them out until his debts
are zero. We could leave A, B, C and D alone to agree on
how many Berkeleys extinguish a debt of one "imaginary"
unit or we could try to help them out. Record-keeping and
computational requirements will be drastically simplified by
the expedient of handling "counters" around; even people
9. Luigi Einaudi (1955).
233
who had trouble with arithmetic in primary school can par-

tici pate.
4. We could allow any transactor able to acquire the trust of
the others to issue 1.0.U.'s (in "Berkeley's") and have them
handed around (or transferred between agents on his books)
instead. 10 If experience tells us that people sometimes mis-
place their trust, we might intervene to force the "bank"
regularly to extinguish its I.0.U.'s or stand ready to do so in
either "our B 's" or real goods.
5. Sorne transactor might be designated as a "credit card com-
pany" which allows others that it trusts to register the debts
and claims arising from resource transfers between them on
this company rather than on one another. The method or
methods for extinguishing these debts and claims, we might
leave to the company unless it proves prone to misplace its
trust.
In a system where sorne mix of these (and perhaps other) ar-
rangements is in operation, it is quite possible that we might
find an empirically stable demand function for a suitably de-
fined "M." 11 Securing its micro-theoretical foundations <loes
not appear an easy task, however. Putting "real M" in utility-
functions, for example, leaves one with a residue of fearful
doubts; and proposals to reduce the marginal utility of M to
zero seem of uncertain import.
The above sketch has not provided conditions assuring the
stability over time of the relationship between "Berkeley's" and
the imaginary accounting unit (LA. U.). Changes in the relation
of B to I.A. U. would , however, be of relatively limited concern
as long as we <leal with systems where the accounts receivable
and payable carried over through time are small or zero, as as-
sumed above. 12
10. Cf. Hicks (1967), Chapter 2.
11. At least as long as arrangements (1) and (2) or variants and permutations thereof
do not come to dominate the oth ers entirely.
12. Horrible penalties for those found feigning the bishop's graven image on the
counters or otherwise manufacturing "money" have been historically helpful. Sover-
eigns and legislators usually end up exempt, however, leaving them with the capability
of appropriating resources from the prívate sector by "money" issue. The consequences
of such "inflation taxes" should not be very serious however-as long as they do not
also succeed in enforcing a fixed relation bet:ween the unit of legal tender and the unit
of account in general use.
234
Nor is a stable relation between the LA.U. and sorne "com-

posite basket" assured by the sketch as far as we have carried it.
The "I.A.U.-value" of the basket could be any positive number.
Jt is interesting, however, that between Charlemagne and the
French Revolution the drift of the libra was rather slow 13 and,
more to the point, without dramatic discontinuous jumps. Com-
parative static models, defined to exclude "money illusion," 14
will provide no reasons to expect this. Yet, it is possible for an
"imaginary money," without secure real anchorage, to drift
slowly enough so as to preserve its usefulness for economic
calculation of the advantages of alternative courses of action
(and, apparently, retain some-ill-understood-superiority
over "composite basket" contracting units). But this, it would
seem, could only be the case if agents faced with the task of set-
ting prices toda y seek help in the memory of yesterday's prices;
i.e., find value in "preceden t." 15
A sketch of this sort will have to lea ve many loase ends. Here,
they are beyond counting. But we resolutely turn our backs to
ali that, hanging on to but one strand-that our various social
bookkeeping devices have not been shown capable of "keeping
track" in systems that a llow nominally denominated debts and
claims to remain outstanding from one "period" to the next.
111. INFLATION ANO THE LAW

Mankind presumably has put more intellectual effort and ethi-
cal reflection over the centuries into the creation of the law
13. Cf. , Einaudi (1953) passim.
14. Toe term "money illusion" is used here with apologies. Recent changes in profes-
sional usage have made it virtually useless as a tech nical term. Originally, it referred to
individuals with a tendency to be fooled by cunency reforms shifting the decimal point
on ali nominaUy denominated contracts or misers with an irrational passion for nominal
money. This concept is trivial but clear-cut and useful. Late1·, in the Keynesian debate,
the term carne to be used with reference to the behaviour of transactors lacking com-
plete in formation on their alternatives of choice. (Cf. Leijonhufvud, 1968, especiaUy
pp. 384-5). More recently still, in the literature on neodassical monetary equilibrium
growth models, sorne writers h ave used it to refer to agents who fail accurately toforesee
the rate of inflation. This last step should signal general abandonment of the term.
15. Cf. Hicks (1970), p. 19: "In imperfect markets prices have to be 'made'; they are
not just 'determined' by demand and su pply. lt is much easier to make them, in a way
that seems satisfactory (because it seems fair) to the parties concerned, if substantial use
can be made of precedent; if one can start with the supposition that what was accept-
able before will be acceptable again."
235
than has yet gone into social benefit-cost analysis. If, then, re-
peated rounds of gradually improved social cost calculations for
inflation keep repeating the answer that it is relatively trivial, it
gives one pause to note that the law is helpless to assure justice
in inflations.
Because of this impotence of the law, inflations tend to accel-
erate the secular tendency of most Western countries to rnove
away from the Rule of Law toward Rule by Men. Associated
therewith, we expect to observe a tendency for the dominant
popular conception of social justice in democratic societies to
shift from Equality under the Law towards Income Equality.
The first of these conceptions focuses on the evenhanded appli-
cation of the rules governing social and economic activities irre-
spective of the identities of individuals and of the social status
they occupy, etc. The second focuses on the ex post real outcome
of individual economic activity. 16
The two linked tendencies are, of course, subject to divergent
value j udgments. Sorne would cheer them on, others wish that
they could be braked, halted, or even reversed. Here we are
concerned to argue only that the strength of these tendencies
will be associated with inflation and, consequently, that this as-
sociation should be considered in assessing the consequences of
inflation. One note might be added to this: namely, that infla-
tions, even as they speed up the process, are likely to make or-
derly and coherent evolution in the d irections indicated more
difficult to achieve.
The law is help less to assure that a just real outcome is re-
stored to contracts concluded in nominal terms. That is so for
rather simple reasons. The expectations about the rate of infla-
tion in prospect 17 that th e two parties originally held cannot be
16. I n o rder to get on with the topic, this paragraph had better be left as is--patently
inadequate. Two references to cover my escape: Hayek (1973) and Rawls (1971). Toe
basic o pposition between Hayek's emphasis on "spontaneous orders" and Rawls's
equally evident "constmctivism" need not, as fa r as 1 can see, p roduce a clash in the
context of th is section (Chap ter 10 would be another matter-but there I will avoid the
issue).
17. He1·e and elsewhere we make use of "the rate of inAation" as if bo th parties to a con-
tract would define inAation, with regard to their own best economic interests, in terms
of money-price changes of sh-ictly identical composite baskets.
This fudge seems unavoidable if we are to go ahead with the arg ument. But-<:ould
this condition ever be exactly fulfilled (while leaving room for gains from trade between
236
objectively ascertained after the fact. 18 The only "evidence" for

what they then were would be what the two contending parties
now allege and it, of course, is useless to the courts.
No independently defined measuring rod suggests itself as a
standard of justice. Measures of the inflation that has taken
place over the term of a contract cannot be imposed as a stan-
dard ex post. 19 If both parties initially expected 5 percent infla-
tion (in the price of sorne agreed-upon composite basket) and
the actual rate was I O percent, a court using the actual rate to
recompute a contract would fix a debtor loss of 5 percent as
the legally enforced outcome. By simply enforcing the contract
in nominal terms as written, the result would be a creditor loss
of 5 percent relative to the original intentions of the contract.
If, on the other hand, both parties had expected a 15 percent
inflation, a 10 percent actual rate means that the debtor loses 5
percent if the contract is settled without dispute; if a court
were to adjust the contract by adding on the actual infl ation the
resulting debtor loss would be 15 percent.
T he parties may have had discrepant expectations about the
rate of inflation in prospect. In that case, it will be impossible in
pure principie to find an adjustment coefficient such that, when
applied to what the contract says, one succeeds in realising the
expected real outcome for both parties.
Finally, there will be a class of contracts in existence of which
it is true that the parties would n ever have been able to come to
terms- i.e., would not have found any mutual gains from trade
in prospect-had their expectations (correct or not) about the
future inflation rate originally been in agreement.
the two)? Assume two agents with identical, homothetic consumption Lastes. If they are
to trade, there must be division of labour (or differential end owments) between the
two. No price will be more significant to their respective economic interests than that of
the good that is the object of their specialisation of labour. The prices of what they
sell must be included in the respective welfare calculations . . . and we are in trouble.
18. Note that the economist studying the "distribution effect5" of inflation on the basis
of data on the n et monetary creditor or debtor position of transactors or groups will be
in the same boat. T h e work of Armen Alchian and Reuben Kessel (reported in nu-
merous articles) of sorne fifteen years ago is subject to this uncertainty. T h e solidity of
the inferences drawn d epend s on that of the assumption that bo th parties expected
price stability at the time their contracts were negotiated.
19. Sorne such standard may, in effect, be imposed via. legislated price controls or in-
comes policies--but the courts would and could never do it (which is a sidelight of sorts
on what incomes policies imply).
237
Consequently, the law refuses to recognise inflations as a

source of "unjust" outcomes. If suit were brought claiming that
the legitimate expectations of one party to a contract (e.g., a
United States Savings Bond) have been defeated by inflation,
such a suit would be thrown out of court. The price-stability fic-
tion-"a dollar is a dollar is a dollar"-is as ingrained in our
laws as if it were a constitutional principle. Indeed, it may be
that no "real" constitutional principie permeates the law as
completely as does this manifest fiction . lnflations (or defla-
tions) end up being ranged with those Acts of God for which
parties are not held accountable. But this is not because jurists
have mis-identified the potentate responsible. It is because the
law cannot tangle with "him," whoever he is.
To see this in proper perspective, one should realise how
very wide is the range of contingencies with regard to which the
law will adjudícate. The outcomes of any individual's efforts are
contingent upon the present and future behaviour of others.
The law seeks to provide a stable framework of social interac-
tion within which people can form expectations about the out-
comes of their actions sufficiently firm, if not precise, to allow
them to plan their conduct accordingly . It does so, in the first
place, by making certain broad classes of behaviour permitted
or forbidden, in the penal code, to everyone. For a socio-
economic system dependent u pon a very high degree of special-
isation of Iabour this will not suffice. The "rules of the eco-
nomic game" (in the game-theoretic sense) must be given a
more detailed, consistent structure or else the "positive sum"
capable of being realised will be very modest and less than reli-
able. One system of design to accomplish this is to co.astrain the
"strategies" of individual players or groups by restrictions of
the type referred to by Mitchell. Individuals whose economic
effort depends for its result on the behaviour of "the shoe-
maker" are provided the assurance that he will have to "stick to
his last" . . . and his son after him, etc. The other system of
design, of course, is that which provides the legal frame of
"monetary exchange systems." One of its principal features will
be provision for "free" contracting between parties. If your wel-
fare is . significantly dependent u pon the behaviour of shoe-
makers, you contract with a shoemaker-depending upon the
238
potential cornpetition of other shoernakers to prevent hirn frorn

holding you over the barre!. To work reasonably well, there-
fore, this systern of legal design requires cornpetition as a "gov-
erning principie of contract." It also requires dependable
"money" if people are to be "ernancipated frorn restrictions on
what they do and what they get" and be !et loase to do as they
please. The vast, overwhelming rnajority of contracts will spec-
ify receipt of "general purchasing power" as the rnain right of
at least one of the parties.
One of the dorninant concerns of the law in an exchange sys-
tem rnust then be to ensure the dependability of contracts.
How is this to be done? It is a tempting but most naive notion
to envisage a systern of law that guarantees (in sorne sense) to ev-
erybody the realisation of the expectations held when the con-
tract was concluded. 20 This is impossible even as justa general
model of approach (e.g., with "scaled-down" guarantees-"90
percent as a mínimum," or whatever). Sorne of the reas0ns are
obvious-in particular, the omnipresence of a class of contin-
gencies outside the control of the community as a whole: "Acts
of God" and the behaviour of people outside the law's jurisdic-
tion (OPEC). And, of course, people may and will sornetimes
expect more than they can get in any case. But the problem
with outcome-guarantees is more fundamental than that.
It would not work even in a "closed system"-i.e., a system
"closed" off from the wars, pestilences, and natural disasters of
a wrathful Deity and the greed of foreigners alike. For the ex-
pectations of parties can never be made either to mesh per-
fectly with one another or to match ali conceivable contin-
gencies-putting aside the inconceivable ones that none the less
materialise. 21
20. Proponents of guaranteed real income schemes for everybody had better give sorne
thought to the underlying rationale of the structure of inherited law in this respect. In
Britain, during the autumn of 1974, there was sorne public debate of universal real in-
come guarantees (by "indexing") as a notional device for snapping out of the "cost-
push" syndrome. Sorne commentators envisaged guaranteeing the present real living-
standards of the population-at a time when the United Kingdom trade-deficit
amounted to 10 per cent of national consumption. This might be the simplest recipe
for h yperinftation and unreconcilable social strife ever invented.
2 1. I am, of course, denying any ''.jurisdiction" to Arrow-Debreu contingency market
models in the present realm of discourse. Hopefully, it is superfluous to elaborate on
this. My indebtedness to the works of Ronald Coase and Steven N. S. Cheung will, on
the other hand , be evident in what immediately follows.
239
The recorded terms of a contract will never reveal the original

expectations of the parties "in their entirety" (whatever that
might be made to mean); nor will they ever anticípate all rele-
vant contingencies and specify outcomes preagreed upon for
each. In part, the expectations held will be left unstated for the
simple reason that the parties will often wish not to revea! to
each other how they intend to "profit from the deal." But, more
fundamentally, their expectations will in general not be com-
pletely structured; innumerable contingencies will be unan-
ticipated, and not in the sense of being assigned a low or zero
probability, but in the sense of not envisaging the situation that
would arise, if and when they materialise, in the specifics of its
behavioral structure. Expectations with regard to such contin-
gencies are left "unformed." Contracts fail to state them not
because of their "unspeakable avarice" (though that might
often be a decent reason) but for reasons of a more Wittgen-
steinian profundity: "Whereof one cannot speak, thereof one
must be silent."
T he contingencies capable of significantly affecting the out-
come to contracting parties will never be exhaustively enumer-
ated. Again, one may explain this by reference to the "cost" of
letting the fine print run on indefinitely. And this would be a
true statement--no contract will explicitly cover all those con-
tingencies that can be envisaged, for it <loes not pay to do so.
But, beyond that, the conditions of human understanding will
not allow for the anticipation of every relevant contingency. 22
Economists, I firm ly believe, need to do a great deal of fur-
ther work in this direction. lf we are ever going to get a firm
grasp of what isomorphisms we may claim to obtain between
our models and the real system, we need to understand much
better than most of us now do how the law seeks to reduce the
22. T he actual economic "game" that people find themselves "playing" has vast arrays
of the pay-off matrix" blan k ex ante. Entire dimensions of the ouLcome space are left
unspecified (also in the probabilistic sense). WhaL Lhe parties will know about most of
the "blan ks" of the matrix, however, is that, if that is where they find Lhemselves ending
up, Lhe courts will adjudicate: i.e., will provide an ex post definition of what Lhe rules
should have been understood to have been. They will expect, moreover, thaL sucl1 a
ruling will most often, tho ugh not invariably, "make sense" to them. More importantly,
they know that they will not end up deadlocked in an irreconcilable conflict.
One of Lhe d imensions of the matrix should be reserved for drnnges in the value of
money. Along that d imension, the above observations do noL hold.
240
uncertainties of the human condition to (literally speaking)

"manageable proportions" and, more importan ti y, why its solu-
tions to this are structured in a particular way. But here we
must leave off without attempts to transcend the naiveté with
which the problem has been sketched above.
The point for present purposes is this: T he set of contrac-
tually unspecified contingencies where the law will step in to ad-
judicate the outcomes to parties is almost infinite. But it is
not exhaustive. "Changes in the Value of Money" are left out.
In adjudicating d isputes, 23 the courts will, in effect, make a
determination of what expectations the parties could legi,timately
entertain. T he case will be settled so as to satisfy everybody's le-
gitimate expectations, in this sense. Most often, this will be done
by reference to precedents. A court will not hesitate to invoke
precedents of which each party is and was manifestly totally ig-
norant. And it will make new law where no precedents are to be
found. In so doing, it may argue from consistency with existing
law, advancing the particular d ecision, as it were, as a novel
"lemma" to long-established laws. More significantly, for our
purposes, it may adjudicate a case without precedent by refer-
ence to general communal conceptions of what is and is not
"fair ," and hold the parties responsible for understanding and
sharing these social conceptions. The law will have the most
difficulty with those unprecedented cases with regard to which
the public does not hold certain "Truths to be self-evident."
Yet , inflations-apart from hardly being unprecedented- are
not like that. They are "unfair"- "everyone knows that." No
social convention could be stronger and more universally
shared. But the law is impotent. The next section attempts a
preliminary analysis of the behavioural implications of this fact.
One subject has been ignored: "indexation." It is potentially a

large one. I have little to say on it, except that I do not believe it
gives us a way out.
The law is utterly permissive with regard to indexed con-
tracts, escalator clauses and the like. 24 It will recognise and en-
23. My indebtedness to Hayek ( 1973,passim) will be evident here.
24. Many countries do, however, prohibi t index contracts. lt may be that in most cases
such prohibitions are of old standing, going back to an age when sovereigns were
241
force nominally defined debts and claims, it is true, but it will

allow the parties very wide latitude indeed in specifying mutu-
ally agreeable formulae whereby this nominal sum is to be com-
puted.
Having emphasised , first, the impotence of Justice in infla-
tions and, now, the permissiveness of the Law with regard to
stable purchasing power clauses, one can only go on to suggest
that there are deeper problems to indexation than is revealed
by recent discussion.
For "indexation" persists, of course, in failing the market
test 25 long after the force of any initially prevalent social con-
vention of "money illusion" type must have been dissolved.
Even the most ardent proponents of indexing schemes are
usually looking for government to take the lead and put it into
effect. But why are not governments, saddled with the borrow-
ing requirements common today and given their record of
printing money to "redeem" debt, forced by the competition of
the private sector to rely on index-bonds? We have seen sorne
spread of escalator-clauses in labour contracts. That only makes
the situation more odd, however, since these are of short
term 26-short enough, generally, for models of "foreseen infla-
tion" to possess sorne measure of putative relevance.
The fact that the system <loes not spread by itself, one must
suppose , probably contains a few lessons for macroeconomists
habituated to index-deflated "real magnitudes" as variables of
scientific analysis. If transactors found no problem in finding a
mutually agreeable composite basket, and saw no novel and po-
tentially serious risks from using it, is it at all plausible that the
system should not spread rapidly in the present age?
The mutually agreeable basket is not necessarily a problem so
struggling to establish their own coinage as a dominant money. Lending the powers of
the law to the enforcement of prívate agreements concluded in comracling units that
do not correspond to the payment unit of government issued legal tender would entail
a self-imposed constraint on the sovereign's ability to rely on inílationary finance in a
pinch. But in Finland the prohibition is recen!, having been imposed following the
abandonment of the celebrated Finnish experiment with indexation.
25. Cf. B. Klein ( 1976).
26. Such short-term employment contract.s will not be affected by the capital gains
provisions of tax-law. It may be lhat it is chieíly the tax law that inhibits the develop-
ment of longer term index contract markets. I doubl, however, that this could be the
whole story.
242
trivial as to be swamped by perception of the uncertainty of

inflation rates. Even in the simplest case of the "pure consump-
tion-loan" between two parties of identical, homothetic, time-in-
dependent tastes, we might expect to find sorne wrangling over
the virtues of Laspeyre versus Paasche and over "the" rate of
interest which should go with one or the other. Where the
specialisation in production of at least one of the parties is part
of the raison d'étre of contracts, things get murkier. Suppose,
both "shoes" and "apples" are in the composite basket used in
comprehensive indexation of contracts. If the apple harvest
fails badly, the shoe-producer finds himself obliged to increase
wages. The apple harvest would not normally be a business risk
that much concerned him. If demand shifts from shoes to
apples and apple prices promptly go up, the shoe-producer
might have to raise his own price in face of falling demand.27
And so on.
With regard to the use of indexation to provide not just pre-
dictable prices and wages but predictable incomes, the work of
S. N . Afriat shows that use of one common index number to
scale up nominal income proportionally will not leave the real
income distribution among income classes unaffected. In gen-
eral, a different "marginal price index" should-in fairness-be
used for each income-cbss and even that will fail to take care of
individuals with atypical tastes in a given income-class. 28
When the law draws a line between legitimate and "illegiti-
mate" expectations of contracting parties, the result is, as we
have indicated, a line between contingencies for which a party
can and cannot seek redress at court. Inherited law thus em-
bodies a "choice" of the adverse contingencies that parties must
accept without recourse as well as of profitable outcomes that
they need not share. Since the system as a whole <loes not
possess "certainty in the aggregate," the law must necessarily
contain sorne set of rules allocating risks in this manner. The
particular rules that we have inherited might have a functional
basis. If so, it is one ill-understood by the economics profession
at the present time. In any case, it is clear that prívate parties
27. The point is Klein's.

28. S. N . Afriat ( 1978).
243
contracting on an index basis will thereby (a) redefine the sets

of adverse and favourable contingencies for themselves, and (b)
within the former set give novel definition to the sub-set for
which sorne measure of redress can be sought. And, to repeat,
they are not doing it.
There remains the question: Suppose everybody did, what
would be the systemic consequences? Until we gain a better un-
derstanding of the considerations sketched above, we cannot
hope to get a foil answer to this one. But the point forcefully
made in a recent paper by Davidson and Kregel suffices, in my
opinion, to settle the question of the desirability of trying to
bring it about. It would, they argue, "institutionalise" and give
legal force to unitary elasticity of price-expectations. A system
where expectations generally had this property would, as Hicks
pointed out long ago, be on a knife-edge at best. Any small dis-
turbance increasing one price could set "the price leve!" going
up without end. And monetary restriction, Davidson and
Kregel add, could then only serve to break virtually every
index-contract in existen ce. 29
IV. THE SOCIAL AND POLITICAL

CONSEQUENCES OF INFLATION
In 1919, Keynes began a short piece on inflation by paraphras-
ing Lenin as having declared that "the best way to destroy the
Capitalist System was to debauch the currency." And Keynes
agreed: "Lenin was certainly right. There is no subtler , no
surer means of overturning the existing basis of Society . . . . " 30 .
So, we have two thinkers with sorne influence on our times con-
curring that inflation is not to be trifled with. This sweeping
judgement that they shared obviously differs not just in degree
but in kind from that of those latter-day students of the prob-
Iem who seek the social cost of inflation in the effects of a pre-
dictable tax on money balances.
But appeal to "authority" <loes, of course, exactly nothing to
elucidate the issues for us. Indeed, to the extent that these are
scarecrow authorities to sorne people, it may confound the is-
29. Paul Davidson and Jan A. Kregel (1975).
30. Cf. Keynes,EssaysinPersuasion (1972), pp. 57- 58.
244
sues. Besides, neither man has a spotless record as a social sci-

entist. We are obliged to ask whether they knew what they were
talking about. And if at the time they <lid, <loes it still apply to
the world of the twentieth century's last quarter? Keynes, for
example, was much preoccupied with the effect of inflation on
the saving habits of the Victorian middle and upper classes.
The bourgeoisie of the nineteenth century is no longer with us.
So it is not at all obvious that Keynes's and Lenin's obiter dicta
have any bearing on how the social consequences of inflation in
the "mixed economies" of our age are to be assessed.
Keynes, moreover, can be pretty discouraging: "The process
[of inflation] engages ali the hidden forces of economic law on
the side of destruction, and <loes it in a manner which not one
man in a million is able to diagnose." 31 Any individual is en-
titled to the claim of being one in a million-in sorne respect.
But not in this one. This famous line is quoted here only to
lodge the complaint that the United States is short of the 200-
odd experts on the "Social Consequences of Changes in the
Valu e of Money" that, on Keynes's reckoning, we are entitled
to.
What may be attempted at this stage, given how the whole
problem area has been neglected in recent decades, can be little
more than to state sorne of the questions that need to be at-
tacked.
The social cost calculations of the output-loss attributable to
inflation have had the dominant share of economists' attention
in this area in recent years. It seems natural to start from them,
therefore. Two sets of questions suggest themselves. First, have
they the "strictly economic" effects of inflation right? Second,
are the redistributive consequences of inflation correctly
derived and are they then appropriately weighed on an accept-
able scale of redistributive justice?
From the given state of the debate, these are the "natural"
questions to pursue. They are questions that certainly may not
be avoided in any attem pt to assess the social consequences of
inflation. But natural or not, 1 submit that they do not now
belong on top of our agenda. T he assumption that, once the
31. !bid.
245
output-loss (if any) attributable to inflation has been estimated

and taken into account, the Social Consequences of Inflation
end with its redistributive incidence may be the single most
serious stupidity to which economists are prone when dis-
cussing inflation.
In trying to think analytically about the question, we would
do well to concentrate, to begin with, on a thought experiment
that puts all the problems of the ex post redistributive incidence
of inflation to one side. There will be an incidental benefit in so
doing for, once those problems are brought on to the agenda,
emotive political and ideological considerations inescapably im-
pinge on our thinking. It is important that we direct our atten-
tion away from such divertissements, for as long as this can legit-
imate ly be done, and on to questions of the behavioural
implications that flow from the experience of inflation. Its re-
distributive consequences are not the "final outcomes" of infla-
tion; there are the further questions of how people experience
them, of how their perceptions of society are thereby affected,
and of how they adapt their behaviour in society as a conse-
quence. And these may be the most important questions of
them all; whether that is so or not, they are the questions that
can put us on the trail of what Lenin and Keynes were talking
about.
In order to set aside the immediate redistributive conse-
quences, therefore, let us proceed "as if" we were dealing only
with a set of individuals that are "representative" in the limited
(and somewhat peculiar) sense that their ex post redistributive
gains and losses cancel each other out in approximately the
same way as for the economy as a whole.=i2 To illustrate: For ali
I know, I may be such a " representative" individual. I am being
swindled on my life insurance and my pension but am getting a
sizable stream of ill-gotten gains on my home mortgage. Sup-
pose these things cancel.
Does that mean that for people in this "representative" posi-
tion inflation <loes not matter? O f course not. How silly ever to
think so. That ex post real net worth may happen to be unaf-
32. For reasons already given in the last section, it is ver)' doubtful indeed that we
would be able to ascertain who exactly belongs to th is set and who <loes not. But-no
matter. . ..
246
fected does not mean that such an individual is living in the

"same world" as provided by a regime of price stability. H is
socio-political attitudes will not be unaffected, unless he is u n-
commonly obtuse; his behaviour will change and adapt, unless
he is "irrational."
What are for such an individual the most salient facts about
inflation sum up to the sadly trite cliché: Two wrongs do not make
a right. You may happen to come out even, as the dice fa ll, but
the game is not inherently fair. At no point in time do its rules
make sense. Besid es, "th e H ouse" will switch them on you with-
out warning. (Th at in a society with progressive income taxes,
the House a lso takes a cut we h ere ignore.)
We can see that substitutions among patterns of socio-
economic activities in two broad directions are ind icated:
IV A
Being efficient and competitive at the production a nd distribu-
tion of "real" goods and services becomes less important to the
real outcome of socio-economic activity. Forecasting inflation
and coping with its consequences become more important.
People will r eallocate their effort and ingenuity accordingly.
The relative significance of two types of capacity for adapta-
tion to changing conditions has changed. The product de-
signer who can come up with a marginally improved or more
attractive product, the p roductio n manager who in a good year
is capable of increasing the product per man hour by a per cent
or two, the vice president of sales who might reduce the real
cost of distribution by sorne similar amount, etc., have ali be-
come less important to the stable functioning and/or survival of
the organisations to whi ch they belong. Other functions requir-
ing different talents h ave increased in importance: the vice
president of finance with a tale nt for so adjusting the balance
sheet as to minimise the real incidence of an unpredictable
inflation ra te is an example. But the "wise gu y" who can do a
good job at second-guessing th e m onetary authorities sorne
moves ahead is the one wh o really counts. Smart assessment of
the risks generated by the política! game comes to outweigh
sound judgement of "ordinary" business risks. Other roles will
247
gain in importance also (for reasons that we will come to).

Among them is the lawyer capable of finding ways to minimise
the impact of sudden new governmental interventions and that
of the "operator" who is quick to spot ways of making profit (or
avoiding loss) from new subsidy, quota, or price control
schemes.
In short, being good at "real" productive activities-being
competitive in the ordinary sense-no longer has the same•pri-
ority. Playing the inflation right is vital.
Perhaps, we had better consider these to be primarily "eco-
nomic" rather than "social" consequences. One had better not
presume that their social aspects are negligible. But phi-
losophising on what effects on the "quality of life" in society
may follow from changing the relative rewards of "hard work"
and "huckstering" seems neither inviting nor promising. If we
postpone these considerations until we come to the Economic
Consequences of Inflation we will at least find the jargon in
which to talk about them more comfortable.
One exception has to be made, however. The most important
of the effects of this type will straddle the boundary between
"economic" and "socio-political" consequences no matter how
we choose to draw that line. It concerns the great majority of
workers. They, too, are put in a situa:tion where individual ef-
fort and performance at work have become a less effective way
of augmenting or just maintaining family real income. The
increases in wages that an individual could hope to gain in any
given year through bonuses or upgrading of his job classifica-
tion, etc., are of little consequence in a double-digit inflation.
Collective action becomes correspondingly more important. He
will have to put increasing reliance on his union.
Since the United States has a lower proportion of workers
unionised than most Western countries, the "theory" that puts
the "blame" for inflation on union "militancy" has gained less
currency in the United States than elsewhere. This should be to
the country's advantage in trying to address its problems ra-
tionally, since this "cost-push theory" basically misidentifies the
forces at work, making the "cause" for inflation out of what is a
predictable consequence o[ inflation; namely, observably increas-
248
ing union activism. 33 In any case, we should note that the asso-
ciation between high inflation and union activism, out of which
has been conjured the inflation theory most "popular" in sorne
other countries, is observable also in the United States. Unions
will not only bargain harder and more frequently, they will also
Jobby more energetically and continuously in Washington and
¡0 State capitals. This brings us to our second set of observa-
tions about the behavioural adaptations that we expect to find.
IV B
People will rely relatively less on private contracts and relatively
more on political compacts in trying to ensure for themselves a
reliable frame for their economic lives.
Inflation, and particularly a ragged inflation , renders prívate
agreements less reliable in their outcome. Inflation also renders
prívate agreements less "agreeable"-shall we call it?- in the
simple sense that the fact that both parties initially entered into
an agreement "voluntary" carries much less of a guarantee that
it can be carried out amicably and without rancour than is the
case in a regime of stable prices. 34
The "economic interest" of individuals goes beyond consum-
ing food, clothing, shelter, health care, entertainment, and so
on. We all strive to control our fates, to shape our lives, and to
gain sorne sphere of relative autonomy in the midst of a world
which "in the large" is quite beyond our control. Most of us are
conscious that the trouble with unemployment and with pov-
erty lies less in the reduced size of the "consumption basket"-
which at other times and in other places has allowed people to
live contentedly and with dignity-than in the loss of control
and autonomy in this sense that individuals experience. Were
it otherwise, a programme of adequate hand-outs could eradi-
cate the social problem-a barbarous presumption.
33. While this should be to our advantage in trying to understand the processes in
which we are caught up, it is one that we squander by simply going witch-hunting
among big business and food-chain middlemen, etc., instead. The natural sciences have
got rid of "animism" ali the way clown through primary school but "social animism" still
is a far, far way from fall ing into general disrepute.
34. Cf. Hicks (1970): . .. direct economic loss and (very often) loss oftemper as well."
249
In a regime of unstable money, it is not rational for people to

rely on private contracts and agreements to the same extent as
in a stable money regime. The substitute instrumentality is po-
litical. 35 We expect people to use their votes and lobbies increas-
ingly to help ensure for themselves a predictable real income.
Such activity may take the form of demands on the government
itself for adjustment of taxes, for transfer payments, for "free"
or subsidised government-provided services. Less obviously
perhaps-but more importantly, probably-we expect our
"representative" individual to rely less on competition and con-
tractual agreements and more on legislated or administered
regulation to control and constrain the activities of those other
groups and agents in society on whose present and future behav-
iour the outcome of his own efforts most significantly depends.
The following observations seem pertinent in relation to this
substitution of public political for private economic ways of
goal-seeking:
i. Consider the polity as a feedback regulated machinery.
If our political institutions allow unemployment to grow, the
feedback will be in unmistakable clear text: You'd better do
something about unemployment or else. . . ! If they err on the
side of inflation, there will be widespread and general com-
plaining about rising prices to be sure, but that diffuse message
is quite drowned in the rising babble of specijic demands and
concrete proposals from ident~fiable interest groups-to compen-
sate me, to regulate him, to control X's prices, and to tax Y's
"ex cess p rofi ts," etc. , etc.
The political demands triggered by unemployment are to
reduce unemployment; those triggered by inf-1.ation are for the
most part not obviously identifiable as "instru ctions" to stop
inf-1.ating. There is an informational bias to the process.
ii. Inflation-induced política! activities are pot likely to be
"neutral" in their budgetary implications. The "representative"
35. In less developed countries, a slowing down or reve1·sal of the movement out of the
"subsistence sector" and into the "market economy" may be the more feasible adapta-
tion. In highly developed industrial economies, to withdraw into economic activities the
outcomes of which are largely not contingent u pon what others do will not be a relevant
option for any significant number of people. It is ignored here. The process of eco-
nomic development is not reversible-which is not to say that a developed economy
could not unravel and come apart at the seams.
250
individuals whose undeserved losses are balanced by ill-gotten

gains might be expected to lobby rather earlier and rather
harder for compensation for their losses than for taxation of
tbeir gains. There is then a bias towards deficits to the political
game of trying to re-redistribute the redistributions via govern-
mental budgets. Growing deficits will make it harder to brake
the inflation clown even as the realisation that it <loes after ali
have deleterious social consequences spreads. And the economy
generating the taxes is not going to get better at it from the
proliferation of regulations and controls-even if these were
not often half-baked as such interventions go, but fully studied,
carefully considered, and intelligently implemented.
iii. The efficiency of the polity as a "productive organisation"
should also be considered, however. Is it, perhaps, subject to
laws of diminishing marginal returns to input of "issues"? It
seems more than likely that inflation-induced politicking is
overloading our political institutions. There are limits to what
they can handle intelligently and wisely in any given session.
Inflations create more "wrongs" than legislatures can put
"right." 36
Much has been made in American media of the legacies of
Vietnam and Watergate as explaining the obviously mounting
ill-temper of public debate, and impatience with "the system."
How big a part of the story these events make is impossible to
tell. But it is simply foolish not to note that the same phenom-
ena are prominent in other countries, such as Britain, who were
not involved in the Vietnam War and have had no Watergate
but who have also failed to control inflation.
iv. The overloading of política! institutions is exacerbated by
another factor. Inflation will unsettle a number of political com-
pacts and compromises reached in the past. 37 Consider mini-
36. In early 1975, President Ford attempted to get action on his own proposals b y por-
traying the present Congress as a "do-nothing" Congress. He was rebuked by a Con-
gressional leader who pointed out that the 94th Congress had already at that time
passed a far greater amount of "significant legislation" than was passed by any of the
Congresses where Gerald Ford was Mino1;ty House Leader. The number of "signifi-
cant changes" per year in the laws governing a country would be an odd index to
choose for either "wise" government or "health" of the polity. That number in any case
is rising. But is there any indication whatsoever that our political institutions are
thereby catching up with the demands for "Justice, Now!"?
37. Again, cf. Hicks (1970).
251
mum wages, for example. Economists are apt to think of the

erosion of minimum-wage barriers to the employment of the
young and of minority groups as a reminder that "there are
good things about inflation too." But our professional disap-
proval of minimum-wage laws is not really to the point as long
as the basic distribution of economic-political interests and the
ways in which we have constitutionally agreed to let them take
expression are as they are. Ali it means is that the lobbying, log-
rolling, and so on will have to be done over again. With mini-
mum wages we expect this to happen regularly, predictably,
and in short order. But presumably this is not always the case.
Issues regarded as long settled may be irrelevant in elections;
politicians make no promises relating to them, and groups with
a significant interest in them decide how to vote on other
grounds. When such compacts come unstuck, the political
"equilibrium" of which they were part will not necessarily be
quickly reformed. Rights and privileges won in constitutionally
fair política! contests become more impermanent. Thus, the
polity too becomes less reliable in d elivering the goods.
Prívate economic contracts, we know, will be concluded for
shorter contract terms, and, even so, be more uncertain as to
their real outcomes. Both statements can be made also for polit-
ical agreements.
v. The law and the political agreements in force embody the
rights and privileges, immunities, duties and obligations that
constitute the framework of social order within which individ-
uals live their social lives and pursue their economic goals. A to-
tally inflexible framework prevents such necessary adaptations
to the social order in a changing world and will ultimately
break. A totally "flexible" one is not a social order at ali. Sorne
measure of basic continuity must be present, must be main-
tained. One cannot treat all the laws and political compacts as
perpetually "fresh" issues, up for renegotiation or open to fun-
damental reform in every season. This is so not so much be-
cause "change" will thwart particular individuals or groups in
achieving their goals (whatever they may be and whatever we
may think of them). It is rather because sorne continuity is nec-
essary for any individual to "make sense" of h is social setting, to
be able simply to set goals for himself and his family and to
252
formulate plans to work towards them. The rights, immunities,

7
and obligations with which one goes to bed at night must be
there in the morning and not found unpredictably reshuffled
or a meaningful social existence becomes impossible.
Any society must strike and maintain a balance between con-
servatism (in the literal sense) and reformism.
There is a third bias to the inflationary process viewed in
terms of its socio-political rather than "purely" economic conse-
quences that should be pointed out in this connection. Consider
once again the hypothetical individual who is "representative"
of society at large in that his gains and losses from inflation bal-
ance. As long as the economic machine continues to turn out
the goods in roughly the same volume, his consumption stan-
dard, etc., will not be impaired. He is suffering undeserved
losses and will identify certain institutional arrangements as the
instrumentalities whereby this has occurred, certain groups or
organisations as the "privileged" recipients of the corre-
sponding gains, and certain immunities of the law as barring
restitution. He will side with others seeking reform of one or
more of these features of the inherited social order which he
sees as having combined to produce a manifestly unjust out-
come. In the nature of the case, the set of institutional arrange-
ments that produce his ill-gotten gains will not be (completely)
the same. Those members of society that directly or indirectly
are paying for his inflationary gains will be out to reform a dif-
ferent set of laws and political compacts.
When inflation gets into double digits by a good margin, one
thus has to expect that virtually all the institutions providing
the framework of economic arder will in this way come under
attack. To sorne extent, of course, they always are-there will
always be critics with sorne following among dissatisfied groups.
But normally most such "movements" will be ineffective; at any
rate, we expect only a very few of them to make significant
headway at any one time. Great inflations, however, are capable
of letting loose a social epidemic of effective but uncontrolled
and incoherent pressures for institutional change.
For where could we expect the defenders of continuity to
come from? Whence th e reserves of "countervailing powers"?
Ordinary, decent, honest people will not stand up for the laws
253
and institutions producing the gains they know to be ill-got-

ten. ~8 Conscience forbids it and con science, despite impressions
to the contrary, is a widespread attribute. Our "representative"
individual, who has so far come out even, is not likely to defend
his ill-gotten gains when they come under political attack by
others; he is more likely to respond by redoubling his efforts to
remove the sources of his own losses.
T he "representative" citizen will, on balance, be on the attack
against, not on the side of the defence of, the inherited order.
vi. Ali of the above concerns the "rational," relatively deliber-
ate and unemotional adaptations that people are apt to make to
the experience of a rapid, but ragged inflation. But to assume
that the degree to which they maintain their deliberate ra-
tionality is itself unaffected by the process runs counter to the
most casual observation. T he process is ill-understood by every-
body; it is controlled by nobody; relatively few people will know
themselves to be nefit systematicaUy, predictably, and lastingly
from it. But the notion that "somebody is behind it," somebody
who is in control and who is doing it for profit will be almost
inescapable to a great ma ny p eople. The habit of confusing the
allocation of "blame" with the description and explanation of
historical processes is almost universal. T hus public opinion in-
creasingly acquires paranoid overtones. Opinion-making entre-
preneurs make careers from such suspicions. The legislative
process itself cannot r emain-does not remain-entirely unin-
fected by irrational expressions of social strife.
V. INFLATION ANO RESOURCE ALLOCATION

Observations about the "purely economic" effects of inflation-
or, more accurately perhaps, a bout the state or our knowledge
r egarding them-are collected in this section. They are col-
lected under three sub-headings; it will be obvious that these
are not exhaustive of the issues. In this section , we attempt to
r etreat in good order-hopeful of avoiding a rout- to within
the boundaries of standa rd economic theory. Constructive dis-
38. Cf. Keynes (1972), especially pp. 68-69.
254
cussion requires that we now obey the neoclassical "rules of the

7
game"-more or less.:rn
Sorne remarks on the relationships of neoclassical construc-
tions to what has gone before may aid in transition:
1. T he standard model treats the economy as a subsystem
whose interactíons with the rest of the socio-political system
may be ignored for the purpose at hand. The definitíon of
protected property rights, permitted and enforceable con-
tract forms , the kinds and extent of political intervention,
are treated as parametric. The good X ; is x; and stays X¡ and
that is that.
2. The model leaves no room for the production manager,
product designer, distribution expert, et al. , to whom we
made reference in the last section. It represents a world
without need for people whose Sisyphean job it is to try to
keep you on the minimum cost curve, judge where the de-
mand-curve is at, keep things "running smoothly" when
somebody falls sick or the coffee-machine breaks down. The
"efficíent loci" are there for anyone to see and you will not
drift off them if nobody pays attentíon.
3. It is at least unclear whether money is needed as a means of
payment on a regular basis. T ransactors apparently hold it
as a buffer-stock against unplanned, temporary deficits in
their balances of payments on current account but the rep-
resentation of the system leaves the possibility open that
most debts incurred might be extinguished by the delivery
of (arbitrary?) baskets of non-monetary goods.
4. "Money" is not needed as an aid to economic calculation.
Convex production sets and convex preferences meet for a
coolly tangential kiss-hygienically separated by the Cello-
phane of a hyperplane-without such mercantile intermedi-
ation. A huge steel corporatíon, say, can be justas efficiently
39. I n my "Maximization and Mar·shall," Leijonhufvud (unpublished) I forswore the

use of the term "neoclassical" arguing that the conceptual differences separating Walra-
sians, Marshallians, and Mengerians are of greater significance to the microfoundations
of macroeconomics debate than are whatever common denominators "neoclassical"
mig ht refer to. So much for New Year's resolutions. Here I need a broad blanket to
cover standard micro-constructions of ali sores and, soggy as it is, "n eoclassical'' will do.
255
run by calculating ali values in terms of apples as the

numéraire (and will, as we have seen , not be embarrassed by
ending up a profitable fiscal year with a rather long position
in apples).
5. Since transactors are good at solving n-dimensional decision-
problems simultaneously under "uncertainty," they make no
use of other devices for simplifying calculation either. In
particular, they have no need for Hicksian "precedents." Of
course not ali the constructions of standard theory represent
worlds in which memory is of no use and the global equilib-
rium is recomputed from freshly gathered information in a
daily before-breakfast tatonnement. Memory, even if limited
to the somewhat non-vettebrate capacity of storing no more
than sorne half-dozen lagged G.N.P. terms, may well be es-
sential to the formation of transactor's "expectations" in
such models. But ex post values of observed variables do not
enter into the decision-rules that agents use to guide their
actions given these perceptions of their opportunities. 40
Fair enough. Now what is there left to say about inflations?
A. Price Adjustment Processes and Prlce Signals

The first thing to say, surely, is that we know very little about
how inflations work their way through the economy. Our em-
pirical knowledge is scant, 41 which becomes less surprising once
one notes that the theoretical work needed to lend it analytical
structure h as been neglected, too. T he n eoclassical monetary
general equilibrium growth model has inflation as "near-neu-
tral" as makes no difference. The Au strian tradition has infla-
40. In "Maximization and Marshall," I interpret the role of the "constant marginal util-
ity of money" assumption in Marshall's theory of consumer behaviour along the lines
hinted at in the text. Last period's MU ,,-an ex post magn itude and hence a "constant"-
is used by the consumer to simplify his n-dimensional decision-problem and achieve
what he hopes to be a good approximation of the optimal outcome. Assuming cardinal,
additive utility, the reliance on MU., makes possible sequential decisions on purchases
following the thumbrule to buy if and as long as:
(MUxlMU.u) =P ~> P x
lnflation will obviously wreak havoc with this decision procedure.
41. So scant that one is more than usually indebted to Phil!ip Cagan for bis recent
pamphlet, Cagan (1974).
256
tion associated with systematic and serious distortions of the

price system and hence of resource allocation. lt is difficult to
see that we have the empirical knowledge that would discrimi-
nate between the two. My own "hunch" with regard to present-
day conditions would be that the price distortions are apt to be
Jess systematic than in the Austrian view but none the less
serious. There is no good evidence for this view either. The
procedure of arriving at indirect measures of "real G.N .P." by
index-deflation of money value data gives us little indication of
how sizable the losses might be. 42 But we might entertain the
hypothesis that, when "everybody" complains of being worse
off in the face of reportedly unchanged real per capita G.N.P.,
they may be right. The more popular hypotheses adducing epi-
demics of "money illusion" or spontaneous outbreaks of men-
dacious greed are not necessarily true.
How does the price-rise process work through the system? lt
depends on what type of markets we are talking about.43
For securities and commodities traded on the organised ex-
changes the usual "auction" model is probably good enough. So
these we pass over with the observation that, in the United
States, the prices of (the not very oil-intensive) basic food-stuffs
have in the last years severed a long, close association with the
other components of C.P.I. and wandered off on their own,
while individual markets-meat, sugar, etc.-show rather un-
commonly severe "hog-cycling," patterns. 44 It is not the case
that everything is well in our "flexprice" markets.
For most manufactured goods, we have "fixprice" markets.
For such markets, "my story"-obviously both impressionistic
and incomplete-would go as follows. 45 An original increase in
42. E.g., if we move people from "real productive activities" into "inAation huckster-
ing" (or price control bureaucracies, etc.) al unchanged salaries, "real G.N.P." might
show no significant change. Suppose, for example, that the new price-controller's besl
efforts are precisely stalemated by the corporate manager newly assigned Lo precisely
this task. The work oí both may, to a first approximation, end up counted as "real ser-
vice output" measured by Lheir G.N .P.-deílated salaries.
43. The followi ng discussion owes obvious debts to Sir John Hicks, particularly his
recent Tlze Crisis in Keynesian Economic [Hicks (1974)). In a fuller treatment, I
would lean more than is here done on P. Davidson (1974).
44. I am indebted to my colleague Larry Kimbell for driving home this point to me-
with striking statistical ilJustrations.
45. The basic "plot" is dueto Armen A. Alchian and William R. Allen (1967, pp. 86ff).
Cf. also P. Cagan (1974) especially pp. 2-7, and 21-26.
257
monetary demand, increases rates of sales and reduces inventa-

ries faster than anticipated. Prices, I assume, are most often not
put up at this stage. Sorne producers may have "sticky" prices
simply because they are wary of the competition; others will
prefer to "stick" because they hope over the medium-run to
cash in on hitherto unexploited increasing returns. Orders to
r estock are passed backwards through the chain of intermedi-
ate goods producers, leading to inventory reductions at these
levels. At various places down the line we finally run into pro-
ducers who find themselves unable to expand output at con-
stant cost. Now, price-increases begin to be passed for ward
throu gh the same maze of interlocking customer-supplier
chains. T he demand-impulse comes back on the rebound as
"cost-push." Cost increases that a supplier can be confident he
h as in common with his competitors will be passed on in fairly
short order-also, I assume, by sellers, who, if assured of a per-
rnanently higher turnover, would find their present prices very
p rofitable. Reservation-wages of labou r will react in the same
way to cost-of-living increases. We observe "mark-up pricing" in
operation.
We know li ttle about the overall lag-tirne of this process. How
rnuch "inflation" is still in train at sorne date following the ter-
mination of the demand-impulse will be almost irnpossible to
predict. T his matter was probably rather badly misjudged
around 1964-5 in the U nited Sta tes and reaction-patterns h ave
undoubtedly adapted to the exp erience since that time.
Presumably, the process of inventory depletions running
backward and price increases passing forward <loes not proceed
at u niforrn speed be tween sectors and industries. In sorne lines
of business, moreover, th e practice will be to adjust prices in
fairly small steps at fairly frequent intervals; in others, to u se a
larger step-size with longer intervals of posted "fixprice."
Consequently, even if the inflation were balahced , it works its
way through jerkily. At double-digit rates (on sorne smoothed
average), one may expect sizable price increases o n sorne sub-
set of goods to be announced every wee k. Wh at are the irnplica-
tions? T hey can h ardly be d iscussed without at least bending
the "rules of the garne" a bit.
First, of course, it becomes a bother to keep up with it ali.
258
Scale-economies will affect who does and who <loes not try hard
to do so. Traders expecting to transact large quantities will in-
vest considerable resources in keeping track of prices. (Still, one
would not expect a 10-20 percent inftation to be "enough" to
call forth inftation-trading specialists in large numbers--die Gu-
laschbaronen are not yet prominent amongst us). Most house-
holds will not try to maintain their stock of price-information at
tbe "quality" they normally desire-even as they spend more ef-
fort at it. If beef prices go up in every odd-numbered week and
potatoes every even week, sensible beef-and-potato eaters will
resign themselves to a constant proportions diet that is non-
optimal every week-and curse the statisticians who assert their
real income is unaffected by it ali.
Perhaps that sort of thing is not important. But another
proposition, I feel, is: Transactors will not be able to sort out the rele-
vant "real" price signals from the relative price changes due to these
iriflationary leads and lags. How could they? Messages of changes
in "real scarcities" come in through a cacaphony of noises sig-
nifying nothing . . . and "sound" no different. To assume that
agents generally possess the independent information required
to filter the significant messages from the noise would, I think,
amount to assuming knowledge so comprehensive that reliance
on market prices for information should have been unneces-
sary in the first place. Sorne adjustments in resource allocation
that are needed will not be made. Sorne will be made that
should not have been. Between the omissions and commissions,
the vector of effective excess demands is distorted and the
"hunt" for the G.E. solution vector goes off on false trails.
Transactors will gradually lose all firm conception of where
the equilibrium neighbourhood for relative prices lies. Setting
prices and determining reservation wages becomes a more dif-
ficu lt problem-and also a problem that no longer "makes
sense" in the way it used to. We may safely assume that, even in
more stable times imposing less pressing short-run information
requirements, agents have not been used to consider the prob-
lem in n dimensions. Rather, your own past price was used as
the main "precedent'' to be revised in the light of new informa-
tion on changes in demand and on developments in a relative
small set of markets-for the main inputs and substitute prod-
259
ucts. With prices "popping all around" and in irregular

sequence, such a partía! "Marshallian" rnethod rnakes less and
less sense-the pot in which all its ceteris paribus presurnptions
have been thrown together is boiling furiously and cannot be
ignored.
Consider the task of sornebody put in charge of price control.
When is it safe to freeze relative prices? Not right now is always
the answer. Could they be regulated by sorne "rule of propor-
tion" relating thern to prices obtaining in a less discoordinated
state at sorne date in the past? What date? Obviously, there
never is a particularly "good" one to pick. Yet, price-controllers
invariably find thernselves rnaking decisions based on changes
frorn sorne past date or dates-although the econornic theory
they learned at school probably never featured decision-making
based on precedents. Economic agents "at large" will have more
and better information than, but possess no secrets of efficient
decision-making not accessible to, price controllers.
What "value"-in sorne "real" sense-is the rest of society
willing to pay for one's marginal product? We lose track of
what can be expected. In the process, conceptions of what is
"fair" also dissolve. 46 In their original choice of specialisation,
producers are guided by expectations of what real rewards soci-
ety accords this role in the overall division of labour, what
frequency of unemployment might be expected in it, how this is
affected by seniority, and so on. The role is voluntarily chosen
and most people are, actually, fairly well acculturated to the un-
derstanding that the real reward is not socially guaranteed if
tas tes change or someone comes u p with a better way to make a
mousetrap. The irregular change in the real purchasing power
of nominal income that occur in a ragged inflation cannot be
traced to such understandable changes in what the rest of soci-
ety will accord you.
We will tend to end up, therefore, with symptomatic strug-
gles over "fair shares." It is not necessary to postulate that peo-
ple's envy is excited by inflations to explain this. It suffi ces to
note that the normal basis for making (reservation) price-
decisions and forming income expectations has badly eroded.
46. Cf., once again, Hicks (1970).
260
People are forced to look around for sorne reasonably simple,

even though inferior, guideline. What one used to earn relative
to others is it.
Beyond this point we cannot go without ending up back in
Section IV. We have bent the neoclassical rules of the game
here but to bring in the further complications to efficient adap-
tation by transactors that political feedback will cause would be
to break them entirely.
B. The Fisher Equation

In the models, from which it is argued that the cost of inflation
is relatively trivial, the Fisher equation plays a crucial role. A
ful] discussion of the questions surrounding this relation would
ramify into ali corners of monetary theory. 47 Here, I want to
take up only one question. Letting ;tl stand far ( 1/P )(dP(' /dt), the
relation is normally written
( 1)
where i is the observed, nominal market rate of interest and r,

called "the real rate ," is interpreted as the real return facing
savers and the r eal opportunity cost of funds to investors.
It would be much preferable, I believe, if our convention
were to write it instead as follows:
(2)
where i and j denote individual contracting parties. We are
dodging the additional formalism required to distinguish risk-
classes and time-structures of contracts. We should think of (2)
as referring to the market for a particular type of contract.
The first requirement for efficient allocation of a good is
always that a single price should rule in the market. It is such as
analytically trivial proposition that we get in the habit of passing
quickly to more intriguing exercises in welfare theory.
Here, we may assume that competition establishes a unique
value of i. If all individuals (somehow) perceived the same real
47. Touching, for example, on several of the core issues that separate the modern
monetarists from ali the various macro-traditions (e.g., Mises-Hayek; Lindahl-Myrdal ;
Robertson-Keynes-Hicks) that accord Wicksellian themes a prominent role.
261
rate in prospect, then trading in this market would go on until,

at the margin of the positions taken, inflation-expectations were
uniform-to put it very roughly. If inflation-expectations were
uniform to begin with , then competitive trading would go on
until perceived marginal real rates of return were equal. If we
find it difficult to justify one or the other of these two assump-
tion s, we cannot conclude that competition will produce =rj rr
and jij'=f'; as separately holding cond itions. But, presumably,
one would like to establish sorne such proposition as part of
on e's case for the "near-neutrality" of (foreseen) inflations. 4 8
Consider, first, the assumption that a common perception of
real rates of intertemporal transformation is autonomously
given . For a Crusonia world-does the plant still flourish on the
South Side of Chicago?-this makes sense. Only u se of money
and inAation do not. Perhaps, it might be stretched to Fisher's
paradigmatic two-period case, where a homogenous present
good is subject to d iminishing marginal rate of transformation
into a physically identical future good. Accepting the assu mp-
tion in that context amounts, however, to assuming that the
pricing-process works as if "dichotomised." A single inpu t, sin-
gle (but tra nsformed) output case might still do, at least if it is
also point-input, point-output. But multiple stream-inputs,
multiple stream outputs makes computation of "real rates" vir-
tually impossible to conceive of49-unless, of course, fixed rela-
tive prices at a constant rate of depreciation of money were
(somehow) guaranteed. B ut that would be the second case.
I can see no "mechanism" that we could plausibly adduce
which would tend to bring inAation rate expectations into con-
formity. lf we assume a world which h as already been ex-
periencing an unvarying rate of x percent for a generation or
two, ene h as to agree that it is p lausible people will expect it to
continue- unless they learn of developments that migh t
threaten the institutional arrangements of this peculiar "mone-
tary standard." T he an alysis of this possibility is useful for
48. We say "presumably" and "sorne such" here because, intuitively sensible as the no-
tio n seems, it appears almost impossible lO give it precise analytical formulation for the
general case. Sorne of the difficulties are hinted at below.
49. The economic hist0rical literalUre on late medieval, early Renaissance develop-
ments in accounting and the "rationalizalion" of business methods and on the innova-
tions in business organisation that such (necessarily) "monetary calculation" made feasi-
ble is ver y instruclive in this context.
262
various theoretical benchmark purposes. But surely one might

justifiably postpone taking it seriously as a theory of how the
world behaves until such time as somebody actually brings the
trick off? Here, at any rate, it is simply left aside. Without it, it
is still plausible that there will be sorne substantial degree of
conformity with respect to the inflation rate in prospect for the
more immediate future-i.e., that people will share sorne gen-
eral auto-correlation notion: "Things won't change much over-
night." But beyond that, what can we say?
While acknowledging that more theoretical work is needed,
my own tentative position is as follows. Future inflation rates
are not to be drawn from on e of Nature's Urns. Decision-
makers can hardly assume that current observations are drawn
from sorne "normal distribution." What the rate will be five or
ten years down the road is "uncertain," but it is not uncertainty
in that domain of their "natural" expertise where transactors
have learned to make (implicit) probability judgements.
Farmers cope with uncertain harvest outcomes. In speaking
theoretically of "decision-making under uncertainty" as a gen-
eral rather than specific skill we tend to blind ourselves to im-
portant aspects of behaviour. To have learned to manage ra-
tionally despite the vagaries of weather, however, will not leave
much experience applicable to coping with the consequences
compounded from the vagaries of voters in future elections, of
legislatures and governments, and of Central Bank responses to
the contingencies that the polity produces. Nor do "rational ex-
pectations" models provide assurance. They require an un-
derlying, relatively swift and sure "survival of the fittest" pro-
cess anchored in relatively stable conditions of "real scarcities"
for their results to be plausible. 50 Do we have something of the
same sort governing the price leve!?
Benjamín Klein has discussed this matter in terms of the
theory of monetary standards. 51
1. With the old gold standard, it was "rational" to expect
(roughly speaking) reversion of the price leve! back to its old
leve! following a rise or decline.
50. E.g., the type of process we adduce in explaining to students why refraining from
"destabilising speculation" has survival-value in commodity markets.
5 1. Klein (1974).
263
2. From the mid 1930s through to the early 1960s or so, "ra-
tional expectations" (for Americans) might have been to
count on the monetary authorities to revert to a zero rate of
change "as soon as feasible."
3. Klein refers to the situation of recent years as one of a
"purely fiduciary standard." This is a fair description-but
how would one describe the operating "rules" that would
govern the "probability distributions" of future price levels?
I would not even try. 52 My impression is that the interna-
tional monetary "system" has for sorne time been in a period of
unstructured experimentation and "innovation." Whether this
will converge to a stable institutional arrangement and, if so,
what it will be like seems obscure indeed-if for no other rea-
sons than that those doing the innovating do not understand
what they are tampering with or know what their criteria of
design should be.
In the United States, a transactor might listen to those econo-
mists who argue that policy should not be employed to reduce
inflation, but at most to stabilise it. If he believes they rule the
world, he will get unity as the lower bound to the elasticity of
his price-level expectations.53 Another transactor, looking back
over the past ten years, might be more impressed with the fact
that the Fed will still, whenever unemployment is "tolerable,"
listen to Congressional complaints of "high" (nominal) interest
52. The type of "uncertainty" envisaged in most standard economic models of deci-
sion-making under uncertainty may be illustrated by a game of dice. We know the
properties of the mechanism generating the probability distribulion of outcomes. If the
agent does not know it-tbe dice may be biased, say-a Bayesian learning model may
still be used to model his adaptive behaviour. For most economic decisions, the game of
chess may, however, be the better source of appropriate metaphors. Here we cannot
exhaustively specify ali tbe possible alternalive future positions in a game. Con-
sequently, the "actuaria! calculus" cannot be applied to the decision-problems of the
game (cf. also p. 240 and footnote 22 above).
Consider then major business decisions, the outcomes of which are crucially depen-
dent u pon the future rate of inflation, and which have to be made in a setting where no
rules, ultimately constraining the rate of money creation, are accepted as "constitu-
tionally binding" by the legislature and monetary authorities. Observed inflation rates
are not "drawn" from a probability distribution generated by a law-abiding mechanism.
T he appropriate metaphor fo r this case, 1 suggest, is that of playing "chess" in Lhe pres-
ence of an official who has and uses tl1e power arbitrarily to change tl1e rules-i.e., a
man who may interrupt at move 14 with the announcement: "From now on bishops
move like rooks and vice versa . . . and 1'11 be back with more later."
53. Cf. again, Davidson a nd Kregel ( 1975).
264
rates and take the chance to deflate. If we surveyed people's ex-

pectations about "the price level" in 1980--assuming that they
are tolerant enough to answer such a "dam"-fool question-and
found them bimodally distributed, who is "irrational"?
The most plausible conjecture, I submit, is that perceived
"real rates" are not brought into line so that "capital" is being
misallocated all over. Question: Would this be favourable to the
employment of labour?
Integrating the analysis of "ragged" price-rise processes (spot
and forward) from Sub-section A with that of intertemporal
allocation under conditions of non-uniform inflation-expecta-
tions is left as "an exercise for the reader."
C. The Demand for Flexibility

With a tax on "money," we expect people to substitute into
longer placements and to reduce non-interest earning accounts
receivable. With nominal contracts more uncertain, we expect
people to substitute into "real" assets. The first-mentioned ten-
dency would operate endogenously to accelerate inflations. If
this has been happening, increasing "velocity" has had less to
do with it than expected. With regard to the secondary ten-
dency, stock-markets have not been noticeably firmed up by
inflation.
When the future becomes more "uncertain," but the risk that
increases is not a simple "actuaria!" one, we expect people to
avoid long-term commitments in favour of more "flexible' posi-
tions. 54 You steam slow waiting for the fog to lift (and sound
your bullhorn a lot). The demand for flexibility is expressed by
going "short and nominal." 55 Thus, this tendency will tend to
counteract the two mentioned earlier.
That resource allocation will be affected is obvious. We will
not elaborate on it. Flexibility is brought up here because I
54. For the concept of "flexibility," cf. A. G. Hart (1951). Long neglected in macro-
theory, the concept is brought to prominence and the necessity of its inclusion in our
tool-box driven home in Hicks ( 1974) Chapte1· II. ·
With a "simple actuaria! risk," I mean in the text to refer to cases where Hart's "com-
pounding of probabilities" is not needed.
55. This is what J aneway has been talking abou t in commcrcials for savings and loan
institutions that have much upset American economists.
265
believe it ranks in significance with the two tapies already dis-

cussed, not because I have anything new to say in general
terms. Instead, two pieces of "casual empiricism" plus a com-
ment:
a. In the United States, short rates have been plummeting

since the summer of 1974. Long rates are staying up. We ex-
pect short rates to move with greater cyclical amplitude than
long r ates. Yet, this time there may be a bit more to it. First,
the weakness of long markets is properly appreciated only
when the uncommonly short average d uration of the mas-
sive Federal debt is recognised. Secondly, the fall in the
short rate is to sorne extent deceptive. Many corporations
(and New York City) have had their credit-ratings written
down (Aaa to Baa, etc.). Reports in the press indicate that
underwriters are hardly to be found for floating Baa bonds.
Sorne borrowing demands are being rationed out. T hese
prospective borrowers are missing from the supply side of
bond markets. Lenders are going for short and safe place-
me nts in this kind of market. The fa)] in interest rates gives
an exaggerated impression of all-around "credit ease."
b. In Britain, during th e autumn of 1974, the inflation rate was
close to 20 percent. Yet, much of the banking system was at
or beyond the "prudential limits" conventionally deemed
saf e. T h e corporate manufacturing sector and much of agri-
culture were in bad liquidity straits with serious immediate
cash-flow problems. Banks were unable to render further
h elp which would require additional long lending against
short borrowing. Meanwhile, the government was running a
deficit such as to give a borrowing requirement corre-
sponding to 10 perce nt of G.N.P. while, at the same time,
the "fiscal drag" from inflation was proving negative (and siz-
able). With a rate of investment lower than desirable, the
country was running a balance of trade deficit equal to 10
percent of national consumption, mostly financed by "petro-
money" inflows so short as to increase the strain on banks.
Money and liquid assets were piling up in the "personal sec-
tor" and in the portfolios of such institutions as Oxbridge
colleges and insurance companies-"earning" their holders
266
obviomly negative real rates. No positive and safe real rates

were perceived. Alternative placements would include lend-
ing to transactors to whom banks would not lend or pur-
chase of shares in ~orporations whose equity might be ex-
propriated by government as a condition for assistance with
ready cash.
A rather different picture from the Quantity Theory of bal-
anced inflations where one expects to find "dollars burning
holes in every pocket"! An economist, ignorant of the rate of
inflation , taking a look at the "real" situation by sectors of the
British economy would see it as in dire need of "reflation. "
In an earlier article 56 I outlined what was there termed a
"corridor hypothesis" of the adjustment capabilities of (mone-
tary) market economies. In brief, I proposed that within sorne
range around its "equilibrium" time-path, such systems will
tend to exhibit predominantly self-stabilising properties of the
basic type that neoclassical models presume. Outside the corri-
dor, on the other hand, (Keynesian) "effective demand fail-
ures" would increasingly impair the ability of market homeos-
tats to get the system back on course. Two subsidiary
hypotheses, proposed in this paper, about system behaviour
outside the corridor seem relevant here:
1. we should expect to observe the emergence of distribution
effects loosening the normal empírica! relationships among
monetary aggregates and between them and aggregate de-
mand; this would be associated with increasing spreads be-
tween interest rates on safe and risky claims and with in-
creasingly prevalent rationing of borrowers with low or
deteriorating credit ratings;
u. in such situations, monetary policy action should be ex-
pected to be less effective than normally-and particularly if
operated against the current of a contrary fiscal policy.
This 1973 paper was written, out of ivory-tower mental habit,
with prolonged large-scale unemployment as the problem fore-
most in mind. I would now like to add to it the claim that the
Gestalt of the theory sketched there is one that will accommo-
56. Leijonhufvud [1973].
267
date discussion of double-digit inflation-including that stage

of it where unemployment still stays safely within the "single
digits."
VI. CONCLUDING REMARKS

I have attempted to point out a number of issues that appear to
me germane to the task of providing microeconomic founda-
tions for macro-theory. Still other issues are implicit above. An
attempt at systematic summary and assessment would seem to
little purpose here. Readers who have actually survived to this
point might, I hope, agree.
Sorne concluding remarks on the "attitude" of the writer may
save time in discussion. It will h ave emerged that I am (again)
critica! of general equilibrium theory and "neoclassical" models
more generally and on severa! counts. Amon g those others who
share my critica! view (and would add to them), sorne will ask
why one should bother with these branches of theory at ali.
When faced with m ethodologically profoundly difficult prob-
lems of "relating"-never mind "integrating"-branches of eco-
nomics that for long periods have developed along separate
and independent lines, the easiest posture to take is outright
and wholesale rejection of one approach or the other. Almost
always, I strongly believe, this will prove too easy a way out.
Epistemologically sophisticated and convincing cases why this
or that aspect of reality can, in pure principie, not be captured
via sorne particular approach will not often be much to the
point. Such philosophical "impossibility theorems" have a bad
track record in the history of science. All too often, "sorne
damn fool" will go ahead and do it anyway and clean up his
methods, or h ave others do it for him, afterwards.
In any case, this writer h as never come clase to considering
'junking" neo-Walrasian constructions. If I h ave been more
harpingly critica! of this branch of theory than of any other, it is
because in its highly developed modern form it gives us some-
thing precise to refer to. Although my own "beliefs" about how
real world economies behave cannot be adequately represented
by current neo-Walrasian models, I find that-for my "personal
use"-th ey provide, as it were, clear benchmark reference mo-
266
tions that I would not do without. I do not expect other critics

to share this mental habit, nor is there any point in attempting
to convince them that they should.
The result, of course, of trying to hang on to achievements
gained by as yet methodologically incompatible approaches will
be a bit of a muddle. lt is easily productive of sundry analytical
tangles that will be merely tiresome to others. There is no
wonder at all that many economists will see the incentives to
plump for one exclusive approach.
This paper is a good muddle. lt will have been evident to the
reader that it draws on Marshallian, Austrian, and l nstitu-
tionalist as well as Neo-Walrasian sources. The predominantly
critica! tone towards the last mentioned branch of economics is
due, in the author's mind, simply to disproportionate reliance-
with attendant diminishing marginal returns symptoms-on
this branch in recent discussion of the paper's topic.
The time for deciding what approach to economics should be
it, I believe, is not yet. Probably pretty far off, in fact. Mean-
while, we need all the help that we can get. Drawing from
disparate traditions for "insight" means that one still accords le-
gitimacy to "intuitionism" in economics, even as sorne of its
branches develop so as to increasingly resemble sorne sort of
science.
Hence, there is still in my view an important element of "art"
in economics. With regard to the very broad problems in partic-
ular, one is obliged to "play it by ear." Whether the "chord" of
Marshallian, Austrian, lnstitutionalist, and Neo-Walrasian
"notes" struck here makes acceptable "harmony" to others, I do
not know. Yet, in trying to understand the consequences of
inflation, one should, I believe, search for sorne such balance.
269
CHAPTER TEN
lnflation and the Economists:
Critique
About ten years ago, our collective confidence in what econo-

mists could accomplish in the area of stabilisation policy crested
and we were not reluctant to tell anybody who would listen
what we could do. The policy record (in the United States)
since that time has been thoroughly lamentable, featuring
mounting inflation and a "stop-go" pattern of policy response
of increasing severity.
11
Five times the American public has been promised a campaign
to end inflation:
l. the "Art of Central Banking" credit crunch of 1966;
2. the "Keynesian" tax surcharge;
3. the "monetarist" crunch of 1969-70;
4. the price freeze of autumn 1971 plus the price control
"stages";
5. the 1974 biggest crunch of them all.
Rounds (2) and (4), in one sense, should not count. They were
interludes during which the monetary system was stoked up for
a resumption of worse inflation. In another sense, they do
count, namely, as parts of the pattern showing our policy-insti-
tutions consistently failing to deliver on ballyho'ed prom-
ises. The public has come increasingly to doubt that policy-
271
makers will persevere with their stated policy-intentions and

that standard fisca l and monetary policy instruments can do the
job. Quite apart from exogenous complications (OPEC, etc.)
therefore, our situation has become steadily more difficult to
manage:
1. anti-inflationary policy becomes more difficult and costly to
conduct, and the Iags in its effects more tricky to predict,
when you are playing it "against" a public that does not
believe its goals will be realised. Stabilisation policy is easier
to conduct when the prívate sector regards the stated policy-
intentions as good , strong predictors of th e future state of
affairs.
11. The mistakes of past years have constantly buffeted the sys-
tem every which way. The economy is today (spring 197 5) in
a more disorganized state than at any time since 1950 or so.
We know a fair amount about how the economy behaves in
the neighbourhood of "full" employment and with reason-
ably stable prices. We cannot have at all the same confidence
in our knowledge about how it will behave and will r espond
to policy actions in the p resent situation. O ur accumulated
store of quantitative information is less reliable for purposes
of extrapolative forecasting.
Round (5)-the h arsh monetary restraint of Iast year
(1974)-is now regarded by sorne media commentators as hav-
ing "licked inflation" at last. They cite the sh arply reduced rate
of increase particularly of the wholesale price index in the Iast
couple of months. Not many economists share the view, even as
we look forward to more months of the same as efforts to
reduce inventories and weak commodity markets continue. T h e
last downward kick of the whip-saw h as simp ly been the hardest
kick so far-plunging us into serious recession. The policy ma-
chine has already been put into reverse and is picking up max-
imum steam in the opposite direction. We have yet to see h ow a
$80-billion deficit will be financed. The second half of 1976 and
1977 should tell whether 1974-5 was when we snapped out of
inflation or was "merely" another phase in a time-pattern of
divergent oscillations. I think the odds are on the latter.
272
INFLATION ANO THE ECONOMISTS: CRITIQUE
111
Economists have not controlled events, of course. One can tell
chis deplorable story as a sequence of hard-to-handle ex-
ogenous disturbances combining with abnormal obstacles to the
formulation and execution of a consistent, coherent stabilisa-
tion policy-Peruvian sardines conspiring with Arabian sheikhs
to make things difficult; first a President intent on "guns and
butter," then one inattentive because of Watergate; a Congress
too preoccupied with Vietnam or Watergate to produce the
right fiscal policy with short enough lags; and, of course, the
always accursed Fed. The "foil" story of the last ten years would
be a very complex tale indeed; obviously, having that tale told
right would be useful. But the trouble with a complex tale is
that one cannot draw a simple Moral from it.
Could it be that through this tangled web of events there
runs a skein of systematic error in policy-response? If so, why-
and wh y not earlier? And, if we systematically fail to do things
right, <loes the economics profession have any part of the re-
sponsibili ty?
It may be that simple Morals follow only from outright Fa-
bles. Perhaps my impressions add up to no more than a Fable.
IV
It has become a widespread view among American professional
economists that the economic costs and social dangers of infla-
tion tend to be grossly overestimated by the general public and
among policy-makers, particularly in relation to the social costs
and dangers of unemployment.
I disagree with this "New View." Indeed, I am apprehensive
that the undesirability of inflation is, if anything, underes-
timated by politicians, media commentators, and the public.
Hence, as an economist, I am quite untypically fearful of infla-
tion . I n any scientific field, the untypical view is most likely to
be quite wrong.
But the econ omics profession as a whole has not done its
homework on inflation. We have little in the way of well vali-
273
dated knowledge about inflationary processes, such as the one

of the last decade, and their econornic, social, and political con-
sequences. Theoretical apalysis and ernpirical research alike
have been neglected-presumably because of the attitude that
inflation is not such a serious social problem. The New View
just is not on solid ground. Where science is ignorant, one <loes
not get at Truth by attitudinal surveys arnong scientists.
The general type of staternent with which I want to take issue
rnay be exernplified as follows:
a. "For the purpose of abating inflation it will alrnost never be
worth incurring any non-trivial increase in aggregate unern-
ployrnent."
b. "If the action required to reduce the inflation rate by 10 per
cent will increase the unernployrnent rate by 1 per cent (far t
quarters), it ought not to be done."
c. "It is always better policy to stabilise the ongoing inflation
rate (whatever it happens to be) than to reduce it, since the
latter alternative will always create sorne unemployrnent."
Sorne economists hold opinions adequately paraphrased in this
manner. A probably far greater nurnber see an obvious,
serious, known social cost to unernployrnent while recognising
that, in terrns of present-day econornics teaching, the costs and
dangers of inflation appear uncertain, intangible and possibly
trivial. The feeling that it would be irresponsible to counte-
nance incurring known costs for benefits considered "specula-
tive" in nature and unknown in extent rnakes the policy pro-
nouncernents of this latter group for ali practica] purposes of
the same irnport as the advice of those who express opinions,
such as those paraphrased, with conviction.
V
T he policy-rnaking institutions are endogenous to the systern
the behaviour with which we are concerned. A change in the
perceived ratio of the costs associated with inflation relative to
those associated with unernployrnent will change the response-
pattern of the policy-rnaking "sector" and thereby the dynamic
behaviour of the systern as a whole. A reduction in the per-
274
INFLATION AND THE ECONOMISTS: CRITIQUE
ceived ratio of inflation to unemployment "damages" will imply

a tendency for the historical "stop-go" pattern to change to-
wards longer, harder "go"-phases and shorter, more hesitant
"stop"-phases; it would make you more prone to use your
majar, proven policy instruments to keep employment high
and to try doubtful, ad hoc measures to hold inflation clown,
"hoping for the best." It also brings with it a tendency toward
more myopic, short-horizon decision-making on policy. The
cost of unemployment that comes first to mind is that of the
output irrevocably lost right now, whereas the benefits of price
stability are those of a lasting regime. The politick "Short View"
tends to take precedence over the statesmanlike "Long View."
You go hard for the best feasible policy-outcome this year and
cross next year's bridges when you come to them.
Changes of this sort in the pattern of policy response can, of
course, suffice to change significantly the dynamic behaviour of
the system. T hey could account for the emergence of gradually
divergent policy-oscillations around an underlying trend of
mounting inflation in a system previously showing much more
"favourable" behaviour. Is this what has happened? To make a
convincing case that it is would admittedly be very difficult.
A change of the ratio of perceived costs would in any case not
be the wh ole story. Lack of policy co-ordination and an inap-
propriate allocation of responsibilities for "national goals"
among Congress, the Executive and the Federal Reserve has
also been part of it. There was no significant alteration in these
institutional arrangements in the early 1960s, however. Yet,
their weaknesses did not show up in such a serious way earlier.
The new attitude towards the costs of inflation, on the other
hand , was gaining ground in the economics profession from
the early 1960s on.
VI
Now, I "feel" that the New View on inflation is "unsound" and
would use the same word for th~ various statements about the
"ratio of social costs" or the "social marginal rate of substitu-
tion" b etween the two ills of unemployment and inflation. I say
"unsound" rather than "wrong" because it is unclear whether
275
the categories "true" or "false" are pertinent to them. It is no

less unclear, moreover, whether categories of ethical judgment
or of political preference, etc., apply to their appraisal.
They are statements of a sort that is difficult to debate. What
kind of propositions are they? What are their basis? Where do
they come from? The last of these questions looks easiest.
VII
The notion of a stable Phillips Curve is gone. By now, every-
body's Phillips Curve shifts and tilts and loops, now clockwise,
now counterclockwise-and goes north by east when the Gods
are against you . The original idea has evaporated. But it has
left us a curious legacy-the empty space where it used to be.
And we stay there, spinning perilous confusions in it.
The original problem was to explain the rate of change of
money wages. Suppose, like other prices, they move in response
to "excess demand." How measure it? Unemployment must
surely reflect "excess supply" of labour. Suppose observed un-
employment to be a stable proxy for the theorist's concept of
"excess supply." A reasonable hypothesis that deserves a try.
Phillips tried it and thought the results encouraging enough to
warrant further pursuit of the general approach. But the hy-
pothesis was falsified in the same paper where it was advanced.
The "loops" in the data and the vertical scatter at low unem-
ployment showed that the two variables were not related by a
(single-valued) function.
In a famous paper, Samuelson and Solow 1 used a Phillips
Curve regression as the basis for a discussion of the policy-
maker's "Dilemma"-he cannot have price-stability and a toler-
able leve! of unemployment at the same time. This Dilemma
discussion set the context in which the Phillips Curve became
popularised, quickly gaining entry to the textbooks and from
there into the financia! pages.
The change in the perception of the Phillips-curve construct
that this carne to entail was of considerable significance. The
Dilemma discussion tentatively treated Phillips Curve regres-
l. Samuelson, P. A. and R. H. Solow (1960).
276
sion results, in effect, as information about an "opportunity set"

facing policy-makers. Although this seems a natural enough ex-
tension of Phillips's attempt to predict the rate of wage-inflation
from unemployment data, the opportunity set notion turned
developments onto a completely new track. To non-economists
the notion had tremendous appeal-the Phillips Curve in this
version, promises to dispense with the need to learn a lot of
"technicalities" of inflation and unemployment theory, wrap-
ping up what you need to know about both subjects in one neat
package. But economists too were influenced-much of sub-
sequent research and discussion has been in pursuit of the
opportunity set Phillips Curve rather than "merely" wage-
inflation prediction.
The change in the perception of the Phillips Curve has had
two effects:
a. It entirely changes the research question. Finding what vari-
ables will give a good proxy for the excess supply of labour
and thus provide an equation predicting wage-inflation is
one research task. Finding a stable reduced form relating
inflation and unemployment is a completely different one.
One task may be feasible and promising and the other a
fool's quest. In any case, they are not the same. Much of the
later Phillips Curve literature strikes one as confused in this
regard.
b. It recast the theory of stabilisation policy as a "choice prob-
lem" exercise in the conceptual space given by the two axes
of the Phillips Curve.
With (a), we will not concern ourselves. T he entire tangle of
problems referred to as the "Phillips Curve controversy" is ir-
relevant to what follows. We will be concerned with (b) only.
VIII
The inflation rate and the unemployment rate are considered
as "outcomes" of policy. To alternative policy programmés
under consideration there will correspond combinations of the
two forecast ("for next year") with more or less accuracy. The
locus of these combinations is thought of as an opportunity set
277
boundary for policy-makers. It may shift, tilt, etc., but at any

given date, there it is.
T he habit of thinking of any consciously undertaken action
as requiring, if it is to be intelligible, a preference ordering over
the alternative "ou tcomes" now takes over. Where there is an
"opportunity set" there must be "tastes." Otherwise, how could
one decide at ali ? So a preference ordering with the "outcomes"
of alternative policy actions as its arguments must exist. Except
for being defined over "bads" rather than "goods," why should
it not have ali the same general properties that give stability,
convexity, etc. , to a consumer's utility-function (for, say, apples
and oranges)? Except, of course, that this one ought, in a dem-
ocratic society, to be a "social welfare function,"-i.e., a hypo-
thetical preference ordering over alternative "states of society"
that <loes not represent the policy-maker's own interests and
sympathies but is derived, somehow, from similar preference
orderings held by individual members of society.
Voilct. We have managed to squeeze a very complex question
of what is a "wise" course of policy for a nation into a two-
dimensional conceptual space (with, !et u s say, the unemploy-
ment percentage on one axis and the rate of CPI inflation on
the other). And we have partitioned the problem "neatly" into
questions of feasible "opportunities" and of appropriate
"tas tes."
What are the consequences of accepting this conception of
the problem and of purveying it in public places?
The practica) consequences we have experienced and are still
experiencing. But leave that aside. What twist will acceptance of
the conception give to the work and discussions of economists?
The partitioning of the problem complex into questions of
"opportunities" and of "tastes" appears very nearly to be a par-
titioning between questions about which one makes, respec-
tively, "positive" and "normative" statements. Suppose we take
it that way.
Then the economist will tend to think of his strictly profes-
sional responsibilities as confined to the determination of the
"policy options"; i.e., to the tasks of forecasting. On this side of
the partition, where positive statements rule, the disciplined
Popperian process of Conjectures and Refutations will operate.
278
Having defined the rest of the problem for himself as "a mat-
ter of preferences" he will tend to ignore the factual consequences
of inflation and unemployment as subjects of research. Value
judgements he knows to be statements irreducible within eco-
nomics itself. Economic enquiry halts where it runs up against
normative propositions and <loes not trespass on the ground
beyond them.
Since the política] process <loes not in fact grind out a social
welfare function, no one knows what it is. For an economist
who looks at this as "a matter of preferences," it is by that token
also a matter on which "everyone is entitled to his say" includ-
ing, of course, he himself. At the same time, however, to such
statements about what "should be" done will apply that part of
the professional credo which runs: De gustibus non est cli.1putan-
dum. Here, then, we do not necessarily expect to see a Pop-
perian process in operation. So when economists earnestly lec-
ture students, newspaper readers or members of Congress on
what the public good dictates with regard to the inflation-
unemployment trade-off to be made next, what they say may
not have been through any crucible of Popperian criticism. But
they are likely to get a serious and attentive hearing anyway.
Many people will defer to sorne degree to our opinions on the
rather natural assumption that, selectively filtered through per-
sonal value judgements as these normative recommendations
may be, what has been thus filtered must still be a much more
detailed, objective knowledge of what inflation and/or unem-
ployment "means" than laymen would possess. But is not such
deference quite misplaced?
"Values" and "knowledge" will be conflated in what they hear
and read, all right. But how they are fused and how they might
be disentangled is obscure. And when their expression takes
the form of an indifference map in Phillips Curve space, who is
to distinguish shoddy ethics and sketchy knowledge from their
genuine, warranted counterparts?
IX
The unclear fusion of values and knowledge poses a nasty pre-
dicament for whoever thinks he sees "unsound" views gaining
279
ground. To join debate means to get oneself entangled in the

"rules of the game" associated with this entire conception.
There seems to be no avenue by which the "real issues" can be
reached that <loes not lead first through a quagmire of "ideo-
logy" and what not. Fastidious aversion to mud on your face
will put you on the sidelines. So one wades in.
For example: In 1973 (say), someone who had retired on a
prívate pension in 1966 or thereabouts had already been taken
for one-third of his life's savings. Stabilising the inflation rate as
it was going would mean th at h e could only look forward to
more of the same. At the same time, the unemployment rate
was high but the average duration of unemployment was not
yet such as to cause a substantial fraction of the unemployed to
exhaust their rights to unemployment compensation. "Cyclical"
or "non-structural" unemployment is to the individual a tempo-
rary status; he is partially compensated for it through transfer
payments; he may possibly have options for investing in human
capital that are profitable during the period when forgone cur-
rent earnings are reduced; he may by his own subsequent ef-
forts "undo" sorne of the loss of his lifetime earnings, and so
on. N one of which-one is quick to add- makes his unemploy-
ment a matter of social indiffe rence. The man on a prívate pen-
sion shrinking in real purchasing power will not see his current
real income loss reversed; he is not compensated for it; he is
beyond the age where learning additional skills or working
harder will get him back "nearly" to his pre-inflation wealth-
positions.
And so on. Hopeless, isn 't it? These two h ypothetical individ-
uals are not th e only ones affected. They are not "typical." Even
if they were, such arguments could not lead to conclusions with
which any decent person will be compelled to agree. T hey can-
not settle what our social value judgements "ought to be."
Still, as matters stand, it is not pointless to pursue such discus-
sion. On the contrary, reluctant as we will be to get into such a
compromising, unscientific tangle, such debate cannot be dis-
pensed with. For it will revea! to others (and remind us of) two
things. First, it will revea] the immense, tangled complex of fac-
tual consideration-meaning the fates of individuals- relevant
to any responsible judgement and how little we know about
280
that. Second, it will make clear to everyone that there is no

simple, coherent, widely acceptable e thic-or, indeed, party
platform-such as to enable us, once the factual consequences
are taken into account, to derive general (time-independent)
guidelines of the type "one unemployment percentage point
is as bad as x inflation points."
In an older tradition of scientific enquiry, that posed Wertfrei.-
heit as an approachable even if unreachable ideal, the economist
was obliged to keep his social values to himself and out of his
work. This conception no longer rules even in the natural
science fields from which it was at one time presumed to have
been imported. The "right" to state and argue for one's social
value judgements is now no longer challenged. But this "right"
may be on the way to something more-to becoming a privilege
with which to cloak sketchy analysis, casual empiricism, and
shallow thinking on questions of gravity to the common weal-
and with which to shield the basis for judgements from critica!
scrutiny.
X
Since disputing over tastes is so fruitless, one is tempted to try
another tack. Professional training may "pervert" an econo-
mist's attitudes to social questions to sorne degree, but they are
naturally shaped very largely by the same influences that
operate on everybody else. lf disputing these attitudes is point-
less and/or illegitimate, those critica! of them are tempted into
sociological reflection to "explain" them instead. For example:
1. In sorne not too sharply defined sense most economists are

egalitarian at heart. Do we think, implicitly, of the pen-
sioner (again) as someone who, if h e "really" has a lot to lose
from inflation , must by that token be "pretty well-to-do"?
And of the "working classes," who risk unemployment, as
prima facie poor? Does the reluctance to put a brake on
inflation stem-in-part from a vague feeling that anti-infla-
tionary measures amount to "regressive" economic policy?
lf so, are the generalisations about the groups affected be-
hind that presumption sound ones? Or, are we concerned,
281
in this illustration, with groups that, since they differ in age

differ also in their respective ratios of inflation-taxed ne;
worth to relatively inflation-proof human capital? And, if it
were the case that letting inflation rip is indeed the "pro-
gressive" thing to do, is it also the constitutionally and politi-
cally sound way to go about the redistribution of wealth ?
11 . The memory of th e horrors of the Great Depression of the
l 930s still runs deep in the American polity-which, until
now, h as n ot experienced serious inflation. But the eco-
nomics profession is probably the particular repository of
this tribal memory of la rge-scale un employment-it is in-
grained in assistant professor s that were not born then in
ways they are hardly aware of. (Meanwhile the intellectual
immigrants from Continental Europe who so greatly con-
tributed to the flowering of economics in the United States,
and who had experienced serious inflations first-hand,
h ave moved out of influence in Academia into re-
tirement- p romptly to be swindled out of their retirement
income by inflation. Do their still active colleagues sorne-
times send them a grateful thought?) T he view that unem-
ployment is the worst of ali social ills is the lesson American
economists h ave drawn from the 1930s. T heir advice on the
unemployment-inflation trade-off is heavily influenced by
it. And that advice is contributing to impoverishing the ret-
irement years of that very generation which suffered
through the unemployment years of the 1930s (and then
went to war).
m. We used to assume that equities and other real assets were
inflation-proof. One lesson of the last severa) years is that
human capital is virtually the only reasonably reliable store
of value in periods like the present. Is it just coincidence
that active acadernics (not the emeriti) , m~dia commentators
and "intellectuals" in general think the rest of society makes
too much fuss over inflation as a social problem ? Or is it
perchance the case that these groups- who conduct the
public debate on social, economic, and política) issu es-are
composed disproportionally of individuals on whose per-
sonal experien ce the consequences of inflation are not
brought home with full force? The real value of top-grade
282
"intellectual" human capital is insensitive to inflations and,

for that matter, unless complemented by strong political
convictions, to changes in political regime. Over the last
century, the "intellectual classes" have a sorry record of toy-
ing with revolutionary notions. On occasion, the "man in
the street" will show a ready appetite for them. Ordinary
people who would, on such occasions, rather keep the hell
out of the street will feel differently.
XI
The last ten years have brought a spreading realisation among
economists that subjective value judgements on the relative so-
cial undesirability of inflation and unemployment are not good
enough and that the presumed objective components of the
policy recommendations made need be brought out in the
open. One result of this has been a number of simple social
benefit-cost (or, rather, comparative social cost) calculations on
inflation and unemployment. T he early examples of this brand
of Political Arithmetick have produced numerical results that,
on the face of it, strongly support the most complacent attitude
about inflation: it takes an inflation rate well into double digits
or even near triple digits to equal the social cost of 1 (addi-
tional) per cent of unemployment. Such "objectification" of the
issue compels adoption of a common unit of social cost-
measurement; this, of course, is unnecessary when the problem
is left in "preference space." The results obtained depend over-
whelmingly on the choice of measuring-rod for social cost that
has been made. For unemployment, the choice has been the na-
tional product loss attributable to the market inactivity of the
unemployed. At first sight, this may seem a "natural" measure
for the social cost of unemployment. A second look is less reas-
suring. In any case, this choice necessitates measuring the cost
of inflation also in terms of "output loss"- a less obvious notion.
What is thus quantified as the cost of inflation is sorne estímate
of the productive and transactional inefficiencies associated
with attempts to economise on the holding of money balances
that will be induced by the negative real rate on money during
inflations. This, of course, turns out to be a modest number. To
283
the extent that inflation does not affect the size of the "G.N.P.
pie" annually turned out by the economy, it is considered to
have zero social cost.
This social cost concept, then, is drawn from a social welfare
function into which "distributive justice" arguments do not
enter in the sketchy, implicit, and haphazard way in which they
tend to be present in expressions of "social preferences." On
the unemployment side, the "output-cost" is the same whether
the unemployed receive compensation or not. It is similarly in-
variant to the time that individuals spend in an unemployment
pool of given size. If one individual is "voluntarily" unem-
ployed and foil of hope that he can do better than the employ-
ment opportunities immediately open to him while another in-
dividual in despair and resentment sees no better alternative
than criminal activity-the measure of "social cost" is the same.
On the infiation side, the neglect of ali other consequences than
aggregate output-loss may be given a semblance of respect-
ability by assuming (a) that redistributive losers can be compen-
sated, or (b) that they will be compensated, or (c) that, as the
inflation is or becomes foreseen, transactors will be able to safe-
guard themselves so as to make compensation irrelevant. None
of these (including the first one) is sound on the face of it. What
particu lar mix of the three is relied upon in these cost-calcula-
tions is not aJways clear. The questions left dangling are with-
out end: How is the probability of receiving or the ability of ob-
taining compensation distributed among losers? What is the
probability that compensation will be paid by gainers? What in-
stitutional mechanisms exist or can be conceived to carry out re-
redistribution? By what methods are the gains and losses to be
ascertained and accurately measured? How are the skills
required to conduct one's affairs successfully in an inflationary
regime distributed in society? And assuming ali these things to
be known and settled, can we evaluate the results as if they
were "final outcomes"? Or do, perhaps, further social conse-
quences flow and economic implications follow from these
"given" results?
Economists wh o have tried their hand at this sort of PoliticaJ
Arithmetick have not claimed anything more than rou gh -and-
ready first approximations. T he suggestion is rather that this
284
type of "objectification" of "social preferences" gives us a

starting-point for a Popperian Growth of Knowledge process of
successive rounds of improved conjectures and more sophis-
ticated refutations. Sorne work along this line, starting from
one or another specific criticism, has been attempted. It has led
to no more than trivial adjustments in the "numbers." What it
has demonstrated is that the earliest such calculations were
rough-ready-and-robust with regard to model specifications and
estimating methods-which is to say, it has demonstrated that
the initial choice of "output-loss" as the commensurable unit
completely dominates the results.
For societies in which "Man lives by G.N.P. alone," is mo-
tivated in his conduct only by the collective total of G.N .P. , and
where it <loes not matter who gets it, by what rules, or through
what institutional mechanisms, we can take it as firmly es-
tablished that inflation is a trivial social problem. Further ef-
forts to refine and adjust these estimates are superfluous. One
may accept the result mentioned as an "arithmetical certainty."
Given what we have thus learned, it strikes one as odd-
funny, even-that historians have ali but invariably been so
very harsh in their judgement of those statesmen and poten-
tates who have presided over major inflations through the ages
and that they have given such "inflated grades" to currency re-
formers and restorers of monetary stability. One infers that, as
usual, historians have not had the right model. Having now
learned better, thanks to modern quantitive methods, one must
earnestly hope that our leaders will not be afflicted with an old-
fashioned concern for those posthumous reputations that histo-
rians administer but will let their conduct be soundly guided by
nought else than their prospects in the next election.
XII
I know I have not lived up to customary standards of rational,
objective discussion so far. I feel compelled to take up these "is-
sues," but I cannot rnake sense of them. Not being able to make
sense of what I am talking about cramps my style.
The benefit-cost arithmetic is not my basic problem. lt is the
more general, underlying Social Welfare Function concep-
285
tion-of which these "output-loss" calculations represent a class

of crudely "objectified" special cases-that <loes not make sense.
Two axes metered in apples and oranges; their relative market
price; the budget of an individual; his fruity tastes; a budget-
line, a convex indifference curve, a tangency point-and on to
a "rational" solution to an economic problem. These are the no-
tions that we have, by sorne analogy, transferred lock, stock and
barrel into the conceptual space of the Phillips diagram. Infla-
tion and unemployment are "bads" rather than "goods" but,
mutatis mutandis, one naturally expects a "rational" solution to
stabilisation policy to pop out of this thing, if it is only handled
right.
It is then disturbing that the policy-record <loes not strike one
as more "rational" since this conception took hold than it was
before. Perhaps the analogy needs checking? Consider:
i. The polity is not "like" an individual consumer. The Cru-
soe metaphor, always far-fetched, fails us totally, for example.
Or should we write a New Chapter? Wherein Robinson takes a
house-cure for idlen ess, gets that sinking numéraire feeling and
imposes a ceiling price on apples in terms of oranges and fol-
lows it up with an excess profits tax on oranges?
ii. Inflation and unemployment are not "like" apples and
oranges-unless, perhaps, we are . thinking of Discordia's
golden apple (which may have been an orange-a tomato?-
also a matter of dispute).
iii. Are inflation and unemployment in the "social welfare
function" (S.W.F.) as indices of social discord and political
unrest, perhaps? At least, such a version reminds us of the un-
even incide nce of costs and benefits and hence of the presence
of conflicting individual interests. If A and B have directly op-
posing interests on a given issue, we do not ordinarily proceed
by supposing that sundry conditions for the aggregation of
th eir "tastes" are fulfilled so that a collective utility function for
the social group AB can be formed, which is then optimised to
resolve the conflict.
Aggregation <loes not make sense. But, then , neither <loes the
notion that individuals have preference functions that, in addi-
tion to the u sual arguments, include the rate of change of sorne
price index and the national average unemployment rate.
There is nothing to aggregate in the fi rst place. But wh at kind
286
of S.W.F. is it that is not built up by sorne specified aggregation

procedure frorn the valuations of individuals? How do we "le-
gitirnise" it?
iv. Leave legitirnacy aside and consider whether this could be
sorne dictator's utility-function. It had better be a dictator con-
fident in his own caprice-for transitivity, convexity, etc., will
not go very far towards putting together a guiding Principie of
Justice. But no rnatter-the thought experirnent allows us to
ask whether there might exist an underlying general welfare
theory, too cornplex and costly in its inforrnation requirements
to be irnplernented but in the operational version of which un-
ernployrnent and inflation serve as "proxies" for the "real argu-
ments"-regrettably poor proxies, perhaps, but the best that
can be done. Our dictator should be "knowledgeable," there-
fore, and able to keep track of every subject's fate.
We should start from detailed state-descriptions of the sys-
tem. T he dictator's utility-function is defined over such States.
The elernents of a state-description would reflect "how individ-
ual subjects are doing." Not in terrns of their own utility, how-
ever, but in terms reflecting their unernployrnent and inflation
experience separately. Otherwise, the notion of a S.W.F. defined
over unernployment and inflation "proxies" is lost frorn the
start.
Consider unemployment first. Imagine a vector of sorne mil-
lions of elernents, one for each working-age, able-bodied,
sound-of-mind subject. Put a " I " for employed and a "O" for
unemployed. We might suppose our dictator to be ranking ali
such vectors and his ranking to be transitive and ali that. We
now notice, however, that he is neglecting the duration of un-
ernployment, the probability of re-employment tomorrow, the
distinction between "voluntary" and "involuntary" unemploy-
ment-and many other things. So we should proceed to re-
rnedy these errors and omissions. Unfortunately, every step we
take in this direction will carry us further away from a state-
description for which a count of the unemployed could be a
"proxy." It becomes clear, in fact, that the unemployment rate
is if anything worse as a proxy for the r elevant welfare conse-
quences than it is as a predictor of money wage changes. So !et
us drop it and turn to inflation.
Here we rnight, as a first step, imagine a matrix where each
287
column-vector gives the balance sheet of a subject household.

Again, let the dictator know "how much he likes" any state thus
described. Consider the matrix as our operand. Sorne certain
inflation-rate-say, 10 per cent on C.P.I.-will be our operator.
Applying operator to operand, we obtain a state transformation
resulting in a new matrix. (Use of a price idex, rather than in-
dividual prices, fudges the transformation, of course.) The new
state has a different value to the dictator. Note, however, that
we do not and cannot assign "utility" to the operator-the infla-
tion rate; it is associated with changes in the value of the S.W.F.
Note also that this association is nota stable one. Apply the same
operator repeatedly and the successive transformations ob-
tained are not the same. Nor can we have any guarantee that
the process will settle clown, after a limited number of steps, to
repeated identity-transformations in such a way that evaluation
of such steady states will serve as an acceptable approximation
to the "utility" of the entire process. It may not settle clown
ever.
v. In the single consumer example, the apples and oranges
are "final" and "ultimate" consumer goods. He eats them and
they are finished with. Bygone fruits are bygones and leave no
lasting rot in the system. Tomorrow we start all over again with
essentially the same decision-problem.
Inflation and unemployment are not "like" apples and
oranges in this respect either. With them you do not "step into
the same river twice". Troy was never the same after Discordia's
apple had been "redistributed."
They have further consequences. The behaviour of our dic-
tator's subjects adapts to the experience. He needs to keep track
also of another matrix (of individual "behaviour coefficients")
and to evaluate the transformations that it undergoes as well. As
a particular historical process unfolds, he will find, moreover,
that the original matrix of balance sheets needs to be supple-
mented with additional information-for the property rights
and contract forms that underlie its definition are themselves
being transformed.
But here we may as well cease and desist, for it is clear that
wherever such a search for the ultimate arguments of a S.W.F.
of "true generality" might end up, the observed inflation rate
will be utterly and totally hopeless as an "intermediate variable"
288
in any reasonable procedure for evaluating such irreversible

historical processes.
XIII
"The social welfare function is a conceptas broad and empty as
Ianguage itself-and as necessary." 2 Perhaps. With any given
Janguage, though, we need sorne set of injunctions: "Whereof
one cannot speak . . . " etc.
The concepts of the "New" welfare economics at one time did
useful service in identifying the dangers lurking in the "Old."
Yet, in contexts such as the present one, what makes the S.W.F.
notion survive-except fascination with its inexhaustible short-
comings, so many of which will look potentially remediable. ?
T he Samuelsonian "necessity" of imagining a S.W.F. follows
only from prior acceptance of the "necessity" of conceptualising
any problem of policy in choice-theoretical terms. O f the con-
cepts of choice-th eory we may also say that they are "as broad
and empty as language itself . . . " etc. Again, misuse of the lan-
guage needs to be guarded against. Naiveté about the defini-
tion of the "outcomes" of ch oice may, as we have seen, set this
engine of analysis to producing the most appalling muddles.
But there are more serious questions to consider beyond such
abuses. Is a choice-th eoretical formulation always th e sine qua
non far "rational" conduct of policy? T he "choice" may be be-
tween irreversible historical processes that we ill understand
and which we can control only to the extent that a rodeo rider
controls the Brahma bull. Squeezed into the apples and oranges
frame, the world is portrayed "as if' unde rstood and subject to
precalculated control. When the "as if' clause hides more in the
way of unrecognized distortion than of probabi listic approxi-
mation of the situation, the "constructivist error" is afoot. And
that way líes the "Collectivisation of H ubris." 3
We should be on guard against the type of mentality we are
cultivating, or we will end up with students trained to translate
human drama into jargon: Assume a young man n amed Oe-
dipus X; h is utility-function is quadratic, his opportunities
strictly convex, and so, natu rally, he.
2. Quoted from Samuelson (1952), p. 37.
3. Cf. Spengler ( 1972).
289

Leijonhufvud, A. Information and Coordination Cap. 9 y 10

Uploaded by

Copyright:

Available Formats

Leijonhufvud, A. Information and Coordination Cap. 9 y 10

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Leijonhufvud, A. Information and Coordination Cap. 9 y 10

Uploaded by

Copyright:

Available Formats

INFORMATION

New York Oxford

teen years. It trivialises the cost of inflation. This new view is

theoretically defined "pure" concept but as referring to infla-

11. AN INSTITUTIONALIST SKETCH

Sorne of the questions on the present theoretical agenda are

the activities of individuals-devised to ensure that the commu-

In largely non-monetary economies, important economic rights

Neoclassical theories rest on a set of abstractions that separate

fact endogenous to the social system, we mis-identify the conse-

Mitchell uses the term "money" in a sense so broad as to cover

transactions or the running of financia! surplus and deficits, a

who had trouble with arithmetic in primary school can par-

Nor is a stable relation between the LA.U. and sorne "com-

111. INFLATION ANO THE LAW

objectively ascertained after the fact. 18 The only "evidence" for

Consequently, the law refuses to recognise inflations as a

potential cornpetition of other shoernakers to prevent hirn frorn

The recorded terms of a contract will never reveal the original

uncertainties of the human condition to (literally speaking)

One subject has been ignored: "indexation." It is potentially a

force nominally defined debts and claims, it is true, but it will

trivial as to be swamped by perception of the uncertainty of

27. The point is Klein's.

contracting on an index basis will thereby (a) redefine the sets

IV. THE SOCIAL AND POLITICAL

sues. Besides, neither man has a spotless record as a social sci-

output-loss (if any) attributable to inflation has been estimated

fected does not mean that such an individual is living in the

gain in importance also (for reasons that we will come to).

In a regime of unstable money, it is not rational for people to

individuals whose undeserved losses are balanced by ill-gotten

mum wages, for example. Economists are apt to think of the

formulate plans to work towards them. The rights, immunities,

and institutions producing the gains they know to be ill-got-

V. INFLATION ANO RESOURCE ALLOCATION

38. Cf. Keynes (1972), especially pp. 68-69.

cussion requires that we now obey the neoclassical "rules of the

39. I n my "Maximization and Mar·shall," Leijonhufvud (unpublished) I forswore the

run by calculating ali values in terms of apples as the

A. Price Adjustment Processes and Prlce Signals

tion associated with systematic and serious distortions of the

monetary demand, increases rates of sales and reduces inventa-

ucts. With prices "popping all around" and in irregular

46. Cf., once again, Hicks (1970).

People are forced to look around for sorne reasonably simple,

B. The Fisher Equation

where i is the observed, nominal market rate of interest and r,

rate in prospect, then trading in this market would go on until,

various theoretical benchmark purposes. But surely one might

rates and take the chance to deflate. If we surveyed people's ex-

C. The Demand for Flexibility

believe it ranks in significance with the two tapies already dis-

a. In the United States, short rates have been plummeting

obviomly negative real rates. No positive and safe real rates

date discussion of double-digit inflation-including that stage

VI. CONCLUDING REMARKS

tions that I would not do without. I do not expect other critics

About ten years ago, our collective confidence in what econo-