<<

. 12
( 25 .)



>>

it is easy to see that, if one wants to explain the properties of a dynamic of lagged endogeneity since learning takes time and must precede decision-
economy, it is methodologically much more interesting to do this within a making. On the other hand, a concept of ˜static endogeneity™, that is, the
framework where knowledge is variable. Moreover, there is little reason proposition that knowledge and economic choices are simultaneously
why knowledge needs to be fixed in neoclassical theory. Traditional determined, is not relevant here since this would leave Austrian theory, let
arguments that deny the potential variability of knowledge in neoclassical alone neoclassical theory, undetermined and thus make both theories
© LAWRENCE A. BOLAND
98 Principles of economics Knowledge in neoclassical economic theory 99
equally incomplete or circular. data which is accumulating daily? If this were a problem in price theory,
Lachmann is correct in arguing that the Austrian assumption of exogen- and we were to ask for the impact of a price change on an individual™s
ously-variable knowledge is preferable to the short-run neoclassical consumption, it would be evident that we could produce a satisfactory
assumption of exogenously-fixed knowledge. Moreover, the Austrian answer only if we had a satisfactory characterization of the individual™s
assumption provides a vital starting point for examining the question of underlying tastes. The same point applies to the problem at hand. If we
how the stability of knowledge (viz. expectations) affects the stability of wish to determine the impact of a change in ˜the news™ on an individual™s
the aggregate economy. However, one might implicitly disagree with expectations and thus on the pattern of his or her decisions, we must
Lachmann over the calculation of the net benefits associated with this satisfactorily specify the theory of knowledge through which the changed
specificational improvement. No matter how satisfactory the methodologi- news is fed. Lachmann and Shackle would appear to assume that all
cal role one assigns to knowledge in economic models, the benefits of this individuals respond to the news through a common theory of knowledge. It
will be nullified if one happens to pick, at the same time, an unsatisfactory is, however, easy to conceive of at least three different epistemological
characterization of the views of knowledge held by economic agents. viewpoints on the same change in the news and thus three different
decision-making responses.
Consider first individuals who follow Mises and thereby hold an
THE EPISTEMOLOGICAL PROBLEM OF KNOWLEDGE
Apriorist theory of knowledge. Since, by definition, their expectations are
So far nothing I have said would require a commitment to any particular formed quite independently of the ˜data™ revealed by the news, their
views of knowledge which might be held by economic decision-makers. expectations and the pattern of their decisions can only be invariant to
My concern has only been with the methodological role that knowledge, changes in the news. Consider next people who, like many of us in the past,
however defined, might play in the structure of economic explanations. In hold a Positivist view of knowledge. These individuals will normally only
this section the central concern is explicitly with these alternative views look for new positive evidence to empirically support their guesses (viz.
and their implications for neoclassical and Austrian theory. In this light, it their inductive inferences). If a change in the news reveals predominantly
is interesting that nowhere in Lachmann™s essay does he tell us exactly ˜verifying facts™, then these individuals™ expectations (which are being
what constitutes ˜knowledge™. Implicitly Lachmann must hold that there verified) will change little and thus the pattern of their decisions will
exists some ˜accepted™ theory of knowledge upon which all economic change little. It is only in the extreme case where the changed news reveals
agents inevitably base their decisions, a theory which apparently is so well predominantly ˜refuting facts™ that these decision-makers™ expectations will
accepted that it need not be stated. The implied claims about the be appreciably affected. Consider finally people who hold a Scepticist
theory of knowledge.6 These are people who are always looking to the
homogeneity of viewpoint shared by all economic decision-makers are
indeed comforting. However, they do not sit well alongside the facts that, news for indications that they should change their expectations. Except in
in spite of Lachmann™s grouping of them, Mises is usually considered an the case where the changing news consistently reveals only ˜verifying
Apriorist3 while both Hayek and Shackle based their views of knowledge facts™ (which are irrelevant to the Sceptic), it is evident that these people
on Inductivism4 (albeit of the sceptical kind). Lachmann is in fact prepared rarely have uniform patterns of behaviour and the effect of new ˜refuting
to play down Mises™ Apriorism [p. 56] on the grounds that epistemological facts™ can often have a devastating impact on their expectations and
differences are of little consequence to the matters at hand. My point is decisions.
simple. Such differences in views of knowledge are absolutely central to These alternative characterizations are of course most relevant in
the matters at hand.5 assessing Shackle™s view that volatility and instability in the news imply
We need not be troubled with the many questions of overriding volatility and instability of the aggregate economy in general. Clearly if all
philosophy of science implied here. We only need a few simple decision-makers were Apriorists or the news revealed only ˜verifying
characterizations of theories of knowledge which ˜ordinary™ economic facts™, Shackle™s argument simply could not hold. Perfect volatility in the
decision-makers might hold. The importance of these theories may be news is consistent with perfect stability of decision-making in aggregate. In
directly brought out by considering a problem posed frequently in fact, it is only in the extreme case where most decision-makers are
Shackle™s critique [1972, p. 180]. He asks, in what way do decision-makers Scepticists and the news contains mostly ˜refuting facts™ that Shackle™s
respond to an (exogenous) change in ˜the news™ “ the new information or instability argument may prove interesting. And emerging from this
© LAWRENCE A. BOLAND
100 Principles of economics Knowledge in neoclassical economic theory 101
extreme case are all the criticisms of the neoclassical presumptions of ily-imperfect knowledge.
rationality, economic stability and the fixity of knowledge. This general concern with only successful rational action greatly limits
The question, of course, is whether Shackle™s extreme case is a the role of rationality in economic theories and methodology. It is well
˜satisfactory™ characterization of the epistemological environment faced by known that logic or rationality can only tell us when we can expect our
an aggregate of individuals. If it is not, in the sense that a weighted average rational conclusions (predictions or expectations) to be true. It cannot help
of the theories of knowledge held by all individuals does not reveal them to us predict whether our conclusions are false. This means that true and
be predominantly Scepticist and a weighted average of all new ˜data™ does successful conclusions are logically possible even when one™s assumptions
not reveal a high proportion of ˜refuting facts™, then criticism of the are ˜unrealistic™. This is because it is possible for one to deduce true
stability and fixed-knowledge presumptions of neoclassical economics are conclusions from false assumptions even when the deductions are logically
valid.7 Thus successful actions do not prove that the decision-maker™s
beside the point. Even if the news is perfectly volatile and unstable, this is
of limited interest whenever individuals show little responsiveness to assumptions were true. Moreover, if one follows Shackle by accepting the
changes in the news. Moreover, the methodological comparative advantage view that ˜rational success is only possible with perfect knowledge™, then
of specifying knowledge as a variable must be largely illusory under these one could never explain a decision-maker™s successes as a necessary
circumstances. However, even if Shackle™s extreme case was a outcome of a rational argument. Not only because true assumptions are not
˜satisfactory™ one, there is no way in principle in which this could necessary for successful actions, but because even if they were the
challenge the rationality assumption of neoclassical economics. Scepticism decision-maker could never know for certain whether his or her
is an eminently rational viewpoint for a decision-maker to hold and it assumptions were true “ knowledge is always changing in response to the
would be a mistake to confuse the very high elasticity of response to a news. It is for this latter reason that Shackle would have us replace the
change in the news with non-rational behaviour. rational successes of neoclassical economics (i.e. stable equilibria) with a
second-best kind of success “ his ˜kaleido-statics™ or constantly changing
temporary equilibria. His concept of knowledge, he claims, is adequate for
THE INTERDEPENDENCE OF METHODOLOGY AND
that limited purpose. On this last point he may be correct, but what would
EPISTEMOLOGY
be the cost?
Lachmann™s efforts to stress the importance of knowledge in economic In this light, Shackle™s critique seems to have lost sight of at least one
theory in general are commendable. Nevertheless, neither he, nor Shackle, important methodological virtue of modern neoclassical economics. That
nor any existing economic theory has gone far enough in stressing the virtue is the requirement that we should be explicit concerning all our
epistemological role of knowledge in particular. None of the viewpoints active assumptions. Any long-run equilibrium can be shown to follow
considered here can be taken as both methodologically and epistemologi- logically from some specific set of exogenous givens (resources,
cally complete with respect to knowledge. This is primarily due to all technology, tastes, etc.). But if, as Shackle argues, one of the givens of
parties taking the answer to the question ˜What is knowledge?™ for granted, neoclassical theory is the fixed and exogenous knowledge and it is not
and this presumption may in turn be explained by the common acceptance logically complete itself, then the explicitness of neoclassical theory turns
of a very particular view of the role of ˜rationality™ in all decision-making out to be the source of its alleged downfall. Unfortunately, the insights
and in explaining all decision-making. This view is simply that the gained through Shackle™s focus on change and instability can never be a
˜adequacy™ of the assumptions about the role of knowledge in economic satisfactory compensation for the resultant loss of neoclassical explicitness.
explanations is to be judged only by whether or not they can ˜rationalize™
successful decision-making. Shackle and the Austrians are correct in noting
CONCLUDING REMARKS ON THE LACHMANN“SHACKLE
that a theory of successful rational action does require some judgement as
EPISTEMOLOGY
to what constitutes adequate knowledge, although one may admit to a
variety of different notions of ˜success™ and ˜adequacy™. Where Mises I have stressed a number of key points in the above sections. It is mislead-
˜examines the elements of a logic of successful action™ [Lachmann 1976, p. ing to claim that neoclassical theory is wrong whenever it does not give an
56], Shackle examines what kind of success (albeit limited) is possible essential role to knowledge. To the contrary, as can be seen in Chapter 4,
when we constrain rational decision-making by the limitations of necessar when examining the long-run equilibrium solution to any neoclassical
© LAWRENCE A. BOLAND
102 Principles of economics Knowledge in neoclassical economic theory 103
model, knowledge does not play a decisive role. Specifically, the solution here, in any model where knowledge or time is supposed to be relevant, the
can be obtained without reference to the knowledge of the decision-makers. two questions cannot be regarded as independent of one another since
Thus any argument for the relevance of knowledge presumes the absence answers to the first must constrain answers to the second and vice versa.
of a long-run equilibrium. But the Marshallian tradition based on the long- Epistemology considerations should not be taken for granted. An
run vs short-run dichotomy is misleading. In the short run the decision- inductivist epistemology will always be an inadequate foundation for
maker™s knowledge is necessarily fixed, but as Shackle also stresses, any discussions of knowledge dynamics. Logically, the complete success of
induction requires an infinity of time. 9 This requirement raises an
knowledge which is fixed is also potentially unstable. It constantly needs
updating. His conclusion regarding the inevitability of instability is based important problem. For a short-run equilibrium to be also a long-run
unfortunately on his inductivist epistemology which presumes that all equilibrium, the appropriate knowledge must have been acquired. But if the
knowledge necessarily ˜rests on inductive inference™ [1972, p. 407]. Such attainment of a long-run equilibrium is to be presumed, successful
an epistemology by its peculiar nature is limited since it can only comment inductive inference must entail a sufficient amount of time. Such
on the successful acquisition of the needed knowledge. sufficiency is at least problematic; and this problem is the keystone of
While Shackle™s views can be criticized, there is much to learn from Shackle™s critique of neoclassical economics. He argues that in order to
them. One can see that almost all relevant issues concerning the role of maintain the relevance of knowledge real time must matter. Specifically,
knowledge in modern economic theory can be reduced to two key the amount of time necessary for the attainment of the long-run equilibrium
questions: must be denied. But this critique works only for an inductivist
epistemology. One could just as easily argue for the irrelevance of
(1) What constitutes a ˜satisfactory™ explanation of any economic
(successful) inductive inference in decision-making, even in the short run.
decision-making process if knowledge is taken to be an exogenous
Thus without an argument for the necessity of inductive inference,
and/or fixed element in this explanation?
Shackle™s inductivist epistemology will necessarily be an inadequate
(2) What is a ˜satisfactory™ characterization of the theory of knowledge 8 explanation of both short-run knowledge and the role of knowledge or
held by any given decision-maker of our economic models? knowledge change.
In summary, it is only by attributing a questionable theory of knowledge
The first question concerns the methodological role of knowledge and calls
(viz. Inductivism) to the economic decision-makers of neoclassical
for judgements as to the explanatory adequacy of alternative specifications
economics that Lachmann and Shackle can successfully reach their critical
of a knowledge variable in economic models. These judgements are to be
conclusions. Once the importance of real time and the resulting
distinguished clearly from those on the second, which concern theories of
interdependence between the methodological and epistemological roles of
what knowledge actually is (as seen by economic decision-makers) or the
knowledge are recognized, in conjunction with alternative views of the
epistemological role of knowledge in economic models. The second
epistemological role in particular, the supposed comparative advantage of
question is irrelevant only for models where long-run equilibria are
Austrian theory over neoclassical economics disappears.
assumed to hold since such an assumption entails the (successful)
In so far as neither neoclassical nor Austrian theory provides a
acquisition of adequate knowledge (however characterized).
˜satisfactory™ characterization of such epistemological foundations, both
As Lachmann suggests, historically it is trivially true that Austrian
theories share a common defect. A most important part of this failure is the
theorists have answered the first question better than those classical and
common assumption that the objective of any economic theory is to explain
neoclassical theorists concerned only with long-run equilibrium solutions.
only rationally-successful action, which then constrains all epistemological
However, like most writers on the role of knowledge, Lachmann appears to
theories to explain universal (rational) success or universal (non-rational)
assume that the second question is of little importance or, at least, that there
lack of success. While Lachmann wishes to establish a clear-cut
exists an answer to it which can be taken for granted in any case. Such a
comparative advantage for Shackle™s Austrian economic theory over
view would be acceptable only if (i) the first question could be answered
neoclassical theory, there would seem to be little point in elevating the
independently of the second, or (ii) the assumed answer to the second was
former simply on the grounds that it is less optimistic about rational
satisfactory on its own grounds. Fortunately, Lachmann argues neither
success or that it suggests an extreme characterization of an
point (i) nor point (ii). Moreover, both points are false. As I have argued
epistemological environment which leads to this conclusion. Any proof of
© LAWRENCE A. BOLAND
104 Principles of economics
the ultimate superiority of Shackle™s foundations must be clearly
7 A naive theory of technology
demonstrated for any and all epistemological specifications and it is indeed
and change
fortunate that such global proofs are themselves largely precluded by their
own scepticism.

NOTES
1 The remainder of this chapter is adapted from Boland and Newman [1979]
which I co-authored with my friend and former student Geoffrey Newman. I
thank the editors of the Australian Economic Papers for giving me permission
to do so.
2 I have discussed the role of exogeneity in Boland [1989, Chapter 6]. Elsewhere
[in Boland 1982a, Chapters 2 and 6], I explained that in long-run models, where
appropriate knowledge is simply assumed to exist, any question about the time Sometimes it may be natural to think of ˜technology™ as a separate
needed to acquire that knowledge is beyond consideration. input element, an extra variable in the production function... Suppose
3 Apriorism is usually of the form that knowledge is based on introspection and we have a change that could be described, roughly, as an increase in
in particular all knowledge is founded first on a priori assumptions such that the productivity of labor. This could mean that there has been a change
observations always are secondary. in certain conditions upon which the productive effect of a certain
4 Recall that Inductivism is a methodological doctrine that limits all claims to specified standard input of labor depends. But it could also mean that
knowledge to inductive proofs. I have explained this doctrine more fully in note the units of labor have changed their quality in a way which is not
5 of Chapter 1. reflected in the kind of measure we use for labor input. The same
5 It is thus ironic that Lachmann defends the view that expectations may be could be true for the input of capital... Changes in the kind of capital
divergent among economic decision-makers [1976, p. 58] without admitting used would be a time-requiring process. A change in technology
that epistemologies that decision-makers endorse could also be divergent by the would permit a larger potential of capital accumulation.
same arguments. Trygve Haavelmo [1960, pp. 147“8]
6 Scepticism is the rationalist view of knowledge that says that all attempts to
prove knowledge to be true will lead to an infinite regress. Thus, according to
There is one aspect of knowledge that has always been explicitly
Scepticism, all knowledge that is claimed to be true will always be
questionable. recognized in Marshallian economics, namely technical knowledge.
7 Philosophers call this the ˜fallacy of denying the antecedent™ [see further Bear However, the knowledge recognized in Marshallian economics is about
and Orr 1967; Boland 1982a, Chapter 9].
production technology. While technology is thereby not a neglected
8 For example, is it Apriorism, Positivism, Inductivism, etc.?
element in neoclassical models, it can be argued that changes in technology
9 One might say that this is the only way to ˜solve™ the problem of induction “ of
have been ignored. Recall that Marshall defines the long run as the period
course, if by ˜infinity™ we mean an impossibly long period of time, then the
problem is not solved [see further, Boland 1982a, Chapters 2 and 6]. during which knowledge is fixed [1920/49, pp. 291 and 315]. Since long-
run prices are determined mostly by production costs, I think Marshall is
saying that technical knowledge is fixed. In this regard, consideration of
the problem of knowledge dynamics discussed in Chapter 6 might lead us
to question the adequacy of neoclassical models to deal with questions
involving changing technology and particularly their adequacy when it
comes to questions of economic history.
In Marshall™s day, technical knowledge was variable only over very
long periods such as between generations. Of course, historians are more
concerned with the big picture which involves inter-generational
comparisons. To the extent that history does involve inter-generational
comparisons historians must deal with changes in technology. But while
© LAWRENCE A. BOLAND
106 Principles of economics A naive theory of technology and change 107
knowledge can be acquired by anyone, technology is usually embodied in be seen to be embodied in capital and the accumulation of capital takes
machinery. Without the capital or machinery which embodies the current time. This would seem to mean that there can be no significant change in
technology, one cannot benefit from the current state of knowledge. So it our productive capabilities (i.e. our economic technology) until we have
would seem that to study economic history one ought to examine changed our capital stock (physical or human). It also would seem that
neoclassical capital theory. But unfortunately, neoclassical capital theory is changes in capital cannot be explained by changes in technology and that
more concerned with adjustments of capital within the long-run period changes in output of an economy can only be explained by changes in
where technology is fixed. When a firm changes from one industry to capital (physical or human).
another in the long run, it changes the quality or type of capital from what
was appropriate in one industry to that which is appropriate in the next. The
NON-AUTONOMY OF TECHNOLOGY
choice of industry is based on a given menu of technologies in Marshall™s
long run. I now present my argument for why I think economic technology does not
When, as a new PhD, I first considered the matter of technology and exist by itself but must be manifested in the capital used by a firm or
change I had an idea which I naively thought would be rather easy to work industry. It is common knowledge that we have on many occasions known
out. That idea was that since a firm (i.e. a business enterprise) is how to do something but it was not until much later that we acquired the
fundamentally a social institution, it would be possible to apply the technical capability to do it. In other words, knowledge is not the same
standard theory of social change to the analysis of the technology of a firm thing as technology. For example, technology may depend on knowledge
with respect to changes in an economy. That was a nice idea, but it would but knowledge need not depend on technology [see Agassi 1966]. What is
have failed for the lack of a standard theory of social change that could be it that stands between our knowing how and our acquiring the ability?
applied. This forced me to deal with social change as I dealt with When after acquiring knowledge are we able? This would seem to be a
technology. It is my view now that this can be done by viewing them both fundamental question for the historian of economic technology. Today in
as interrelated aspects of social learning. This chapter presents my early this age of specialization I think the answer is that we are able to
attempts to demonstrate just this view. 1 In the last section of this chapter I accomplish a technically difficult feat, or to produce a new product, only
will briefly outline my conclusions regarding the simple theory of social after we have accumulated the specialized capital in terms of either
institutions developed between the lines, so to speak. In the next chapter I sophisticated machinery or technical personnel. By itself, all the research
will further develop this simple theory to deal with more general questions and development of a firm adds little to the economic technology of that
of institutions in economic history. firm. The fruits of the research or development become part of the firm™s
The evidence of learning by an economic institution, such as a firm, is technology only when the firm invests in (or acquires) the necessary
any accumulation of (new) technology and any improvements in efficiency. machinery and personnel that are specifically designed or trained to do the
This suggests a view of social learning whereby there are two ways a social new job. In other words, all technology exists only by being built into the
institution can learn: (1) through changes in the institution, and (2) through economic institutions, that is, by embodying it in the capital either through
specialized design or through specialized training. 2
institutional reforms. And, while an institution might not learn through a
revolution, a society can learn by overturning some institutions. By To say we cannot obtain new technology without a change in capital
institutional changes in the case of a firm I mean the acquisition of new or (mechanical or human) is to say that technology is not autonomous. Since
different machinery or personnel. By institutional reforms I mean technology may not be autonomous, we need not expect the growth of
improvements in the methods of using existing machinery and personnel. technology to appear to be continuous. Technology necessarily grows in
In each case the learning process takes time. The major learning process, discrete jumps because before a new technological capability appears,
the accumulation of technology, is limited both by the discovery of new institutional changes are required, that is, new specialized machinery must
ideas and by the implementation of them. One can stoically accept the lack be designed and constructed, and/or (new or old) personnel must trained.
of discovery, but the lack of implementation of a known improvement can In either case, the process of introducing a new technology takes time; the
be very frustrating. There is necessarily a disparity between the growth of greater the change, the longer the time. We should not, of course, rule out
knowledge and the growth of technology, that is, between availability and improvements in productive capabilities that arise through improving
implementation. This disparity exists because all economic technology can efficiency. These improvements also take time but they would not account
© LAWRENCE A. BOLAND
108 Principles of economics A naive theory of technology and change 109
for the substantial changes in productive capabilities. Parenthetically, I CAPITAL AND CHANGE
should point out that this latter kind of improvement, or institutional

<<

. 12
( 25 .)



>>