|
post-autistic
economics review |
|
issue 30
contents
PAE
Review index
home page |
|
Finding a Critical Pragmatism in Reorienting Economics Bruce R. McFarling ( © Copyright 2005 Bruce R. McFarling There
is an "ology" that pervades the essay in
which Tony Lawson (2003) launches Reorienting Economics. The preface would lead us to believe it
should be ontology. However, while
ontology receives starring credit, it is epistemology that plays the starring
role. This
first essay is structured into four theses.
In his first thesis (2003: 3-8), the focus is on the mode of
explanation of modern economics. This
is argued to be deductivism, defined as explanation
in terms of event regularities. Lawson
refers to systems exhibiting event regularities as "closed", which
can make his work difficult to read for someone with a General Systems background. For someone more accustomed to
thermodynamic or causal closure, it is helpful to mentally translate
"closed" as "event-regular" everywhere Lawson uses the
term. The argument proceeds that this
mode of explanation in terms of event-regular systems leads to the peculiar
types of mathematical formalisms with which we are all familiar. In
his second thesis, Lawson points out the ill-health of the "mainstream
project" (2003: 8). This consists
in large part of remarks taken from mainstream economists that reflect upon
this poor state of health. It is
conceded that they rarely lay the blame on the peculiar type of mathematical
formalism that forms the touchstone of mainstream economics. However, with respect to "ology" sighting, the essential point is Lawson's
conclusion: " ... there is quite widespread agreement
that the modern discipline is not in too healthy a condition, and that
whatever explains the fact that the formalistic mainstream project has risen
to such dominance ..., it has little to do with this project's record so far
in explaining the social world in which we live. (2003: 11)" In
the first two theses, Lawson has been laying the foundation for the critique
presented as the third thesis. Yet the
foundation for his ontological turn in economics is epistemological. He argues that it is not the content of the
theory in the mainstream content that is stable over time, but rather its
mode of explanation. To justify an
interest in the ontology to follow, he presents a picture of the mainstream
project in ill health. Yet the symptom
of ill health is that it is not succeeding in explaining. In
the third thesis Lawson enquires what
ontology is implied by event-regular systems, and how closely does this match
the ontology of social systems? Lawson
argues (2003: 13-15) that there is an extremely strong bias (although not an
ironclad necessity) toward an atomistic view of such systems, where the
individual agents are simple and react in at least stochastically
deterministic ways in response to given conditions. There is also argued (2003: 15-16) to be a
strong bias toward viewing systems as isolated systems. This argument is easily followed, since a
system composed of nothing but deterministic, atomistic agents will not be
homeostatic, so that the state of such a system becomes indeterminate if the
system is exposed to indeterminate external influences. Lawson
(2003: 16-17) then claims a variety of characteristics of social
reality. For example, positions in
social reality are internally related, the social realm is structured, and it
contains value and meaning. As none of
the features in this list can be exhibited by a purely atomistic and
isolated system, it is concluded that much of what economics needs to be
explained is incompatible with its implicit ontology. Lawson
(2003: 18-20) argues that this incompatibility is responsible for the
constant appearance of the central fictions of mainstream economics. These central fictions are a familiar
fixture. They bear a surface
similarity to the isolating fictions of scientific theory. For example, a natural scientist will adopt
the fiction of a weight dropping through a vacuum to eliminate the real
additional influence of wind resistance.
However, because deductivism is constructing
a theory in a fictitious world that is supposed to correspond to events in
the real world, it seems that its fictions cannot be restricted to the
absence of forces that are in fact likely to be present. They must also include the presence of
fictitious forces to take the place of real world influences that cannot be
expressed in an atomistic and isolated system. While
ontology is brought on stage here, it is certainly not appearing solo. The essence of the argument here is that
the mainstream mode of explanation is not capable of explaining what Lawson
wishes to explain. Certainly, it may
be granted that Lawson's wish to explain particular aspects of reality is an
ontological concern. However, the
capabilities of a mode of explanation is an epistemological concern, and that
is certainly the crux of the argument.
Without the limited capabilities of the mainstream mode of
explanation, the critique falls over.
With the limited capabilities there is something of substance to the
critique, even if one differs with Lawson's ontology. The
final thesis is the conclusion of the critique. Adherence to a mode of explanation in terms
of event-regular systems is the reason for the lack of health, and is indeed
the constraint preventing mainstream economics from being scientific. It is argued that even where science
deliberately constructs event-regular systems in an experimental setting, it
does so to form a cause and effect theory applicable to non-event-regular
systems. (Lawson 2003: 22-26) The
argument may be seen as claiming that the mainstream mode of explanation is
not a scientific mode of explanation.
Here, too, the argument seems as much epistemological as ontological. It
should be stressed that this argument has substantial merit. Constructing a theory in an abstract
event-regular system is not, in fact, the same thing as constructing an
experiment – an artificial event-regular system in the real world. In the former case, the event-regular
system has no necessary connection with a real world non-event-regular system. In the latter case, the fact that it is
constructed in the real world of real components provides the connection with
the same components interacting in a non-event-regular system. While
the starring role of epistemology is the strength of the argument, the
anonymity of the star is a weakness.
Possessing a mode of explanation that implies an inadequate ontology
is not necessarily a critical flaw. As
long as the explanatory process allows the worst mismatches to be replaced by
less severe mismatches, one can hope that the mode of explanation may evolve
toward one that implies an adequate ontology. The
central question, therefore, has not changed in a century, because while
theoretical stances within the mainstream project have proliferated and shifted
ground, essential aspects of its mode of explanation have not changed in a
century. That question is, why do
mainstream economists not engage in evolutionary science? (Veblen, 1898). Of
particular interest for the past 60 years is how a determined effort to
subject theories to statistical analysis have left the mainstream project
every bit as stalled as it was a century ago. Lanis and McFarling
(2004) point in the direction of one flaw.
We construct a highly artificial, highly regular scenario in the
context of explaining the degree of disclosure of accountants in the context
of different national accounting and economic institutions. However, the scenario we have constructed
is non-functional. That is, legal,
professional and commercial institutions will establish norms for what must
be disclosed to provide an adequate report, what may be reported, and
what ought to be reported in exceptional circumstances, and what ought not to
be reported. Any
mathematical relation connects elements from its domain to elements from its
range. In our artificial scenario, we
relate a score on a synthetic index of social attitudes to the degree of
disclosure within the accounting institutions of a nation. A functional relation connects either one
or many elements from the domain to one element in the range. Any relation that connects either one or
many elements from the domain to multiple elements in the domain is therefore
non-functional. Regression analysis,
where it is used correctly, overcomes the problem of relating to multiple
scattered observations by constructing the function in terms of a probability
distribution, so that the scattered observations are interpreted as different
samples drawn from the same distribution. However,
the regularity in our scenario is on the bounds on reporting. Even if one
viewed this range of discretion as a probability distribution, each shift in
the bounds would result in a new probability distribution. The degree of
discretion, which will vary from one institutional context to another,
implies that the regularity is not functional – it is one collection of
institutional norms to many possible degrees of disclosure. Therefore, any effort to fit the best
function will necessarily fail. It
appears that functional regularity is equivalent to Lawson's
event-regularity. We point out several
statistical techniques that may be used with better effect – techniques which
will only be picked up outside the mainstream project, if Lawson's argument
regarding the event-regular epistemology of the mainstream mode of
explanation is correct. In
any event, here is part of the answer to the puzzle of why sixty years of
determined empirical testing has left the mainstream project stalled. If your tool for finding and correcting
mismatches with the real world fits functions to data, you will be left blind
when the problem involves a regular relationship that is not a function. A
more specific epistemological query than the mode of explanation is the unit
of analysis. Any analytical explanation
will involve one or more units of analysis, so that the phenomena to be
explained are explained in terms of the unit of analysis. In McFarling
(2004), I find that the unit of analysis in a particular corner of the
mainstream project to be selection from alternatives, followed by
performance. It may be noted in
passing that this is essentially the same neoclassical unit of analysis that
is ably dissected and subjected to acid critique by Veblen
(1898) as part of an increasingly obsolete natural law approach, which he
contrasts with a modern, evolutionary approach. However, this unit of analysis was
originally unearthed in the work of Posner (1995). Thus, even if much of the theory of the
mainstream project is in a state of flux, this unit of analysis clearly
exhibits greater longevity. I
then pose the question whether it is possible to arrive at a theory of
culture with such a unit of analysis, and argue it is not. If culture acts in part as a restraint on
action, then situating culture in the selection will erode those
restraints. And if culture is simply
embedded in the constraints on decisions, then culture is a deus ex machina,
affecting the outcome but not explained by the unit of analysis. Culture cannot be analysed with selection
followed by performance as the unit of analysis – it can at best be taken as
a given. What
is the point of identifying the unit of analysis? Being informed that the mainstream project
is deductivist, based on presumed
event-regularities, may make it easier to identify a mainstream
economist. However, it does not go
very far in explaining how the mainstream economist is reproduced. The unit of analysis, on the other hand,
goes a long way toward explaining the reproductive process. The first thing a nascent researcher needs
to learn is what type of questions to ask.
And the unit of analysis provides a trio of questions that can be
asked about individuals in a wide variety of settings. What selections are faced by this
individual? Which one is likely to be
selected? And what performance is
likely to follow that selection? The
attraction of this unit of analysis is that there are always more puzzles to
solve. If you try to provide a
complete theory of the economy with a unit of analysis that is blind to
important aspects of the economy, each new trial solution will prove to be a
misfit when it encounters the affects of one or more excluded aspect. And if the reaction to a misfit is to start
over with the same trio of questions, there will always be a permutation of
available selections and likelihood of occurrence that has not been tried
before. Given
this unit of analysis, a question that arises is whether ontology can
be used to reorient mainstream economics.
Supposed this is your unit of analysis in developing new explanations,
and suppose econometrics is your tool for finding out what the problem is
with your explanations. How will you
react when being told that there are certain features of the social world
that do not fit the ontology implicit in your method? You will interpret these features in terms
of your unit of analysis. Indeed,
you may devise econometric tests to determine whether the features as you
have interpreted them are present in the data you have available. If you get statistically significant
results, you may even publish the outcome in a respectable mainstream
journal. Yet you are not likely to
have made a step toward evolutionary science.
When interpreting the features with your unit of analysis, you will
omit what is incompatible with your unit of analysis. Indeed,
in McFarling (2004), I find hope in New
Institutional analysis. This comes
from the argument that the New Institutional unit of analysis is the
transaction, followed by performance.
Yet there is not necessarily an ontological advance here. Indeed, it may be that New Institutional
analysis maintains its credibility within the mainstream project in part by
adhering to the same flawed ontology as follows from the neoclassical unit of
analysis (though see David Dequech 2002). However, by placing the selection in the
context of a transaction between two individuals, the unit of analysis admits
questions regarding relationships between individuals that the neoclassical
unit of analysis does not admit. The
conclusion is not that New Institutional economics is an evolutionary
science, but rather that it is not prevented from being an evolutionary
science by its unit of analysis. The
dangerous face of epistemology is the invitation to focus on the ways that we
understand the ways that we understand things. And it is when I consider the mode of
explanation in this essay that I come upon a concern. Event-regular systems are classified as
closed systems. Everything else are
classified as open systems. One
thing this blinds us to is any other kind of regularity. Suppose that an individual has a regular
response to a cluster of events, where a response consists of one of a range
of actions that are meaningful in that context. Response-regular systems are not
necessarily event-regular, but it is a form of regularity. Suppose that an individual has a response
that is within regular bounds, with clear discretion within those
bounds. Boundary-regular systems are
not necessarily event-regular, but it is a form of regularity. Suppose that a system is homeostatic, so
that the response to an event is contingent on the discrepancy between the
current internal state of the system and the target internal state. A homeostatic system would only approach event
regularity in a perfectly homogenous environment where events are
sufficiently infrequent so that the state prior to the event closely
approximates the target state. Yet not
only is homeostasis a form of regularity, but a collection of homeostatic systems
can create regularity in a wider system. In
other words, in this system of classification, we are to label event-regular
systems as closed and all other systems as open, whether or not they exhibit
one or several other forms of regularity.
If you need event-regular systems to be a reputable mainstream
economist, this may suffice to tell us whether a person is pursuing that
status or rejecting it in favour of status with some other peer group. However, suppose we accept Lawson's
argument that adequate economic theory will normally have to be compatible
with an "open" system. We
have a simple dichotomy here, and the positive category is the one to be
avoided. Reorientation is required
because the pursuit of theories of closed systems is and will continue to be
fruitless in generating effective explanations. Yet saying that the reorientation will take
open systems as the object of theory is to say that it will take
"not-event-regular" systems as the ultimate object of theory. It
is, in short, simply a restatement of the "though shalt
not" dictum, except that this time the "not" has been located
inside the term "open". Thus
in trying for a positive statement, Lawson must elaborate on what kind of
open system he means. That elaboration
appears to be whatever kind of open system is compatible with social ontology
and his realist transformational model of social activity. And it is here that the argument appears to
become controversial. As Vromen (2004) points out, there is a substantial
inconsistency between the qualifications with which Lawson wraps this model,
and the ultimate authority that Lawson grants it as a final arbiter between
properly and improperly oriented economic theory. This
is the crux of the question of whether ontology or epistemology should have
the starring role in this play. If we
have the "right" ontology, how did we discover that it is the right
ontology? And if we have a way of
discovering the right ontology, which is more fruitful to convey: the method
of discovery; or the ontology itself as received wisdom? Of
course, this is a counterfactual. As
strongly suggested by the qualifications that Lawson places on his ontology –
that it is "… practically conditioned, historical and fallible"
(Lawson 2003, 61) – it would appear that at most we can say that our ontology
seems to be the best we can do at the moment.
As Vromen appears to be arguing, this is a
weak basis for launching a revolution. The
implicit recognition of this is built into the structure of Lawson's first
essay. Accepting that the mainstream
project is generating a flurry of explanations without succeeding in
explaining anything is supposed to generate interest in considering the
suitability of the underlying ontology.
Epistemology is providing the wedge intended to create an
opening. Lawson's particular social
ontology is then supposed to enter the gap that is created. Yet
how are we to discriminate between different explanations, once we have
reoriented ourselves to open systems?
I skip past the ontology of the second essay and the realism of the
third to the essay on explanations in social science. Lawson describes a method of forming
hypotheses in terms of relative contrasts that are to be expected if a
hypothesis is correct. He then argues
that "The hypothesis that performs best in terms of empirical adequacy
in this sense over the widest range of relevant conditions can, with reason,
be accepted as better grounded. (2003: 97)" What
we have here, of course, is a pragmatic criterion for judging the
epistemological fitness of a mode of explanation and unit of analysis – or
units of analysis, since the criterion accepts successful eclecticism as
readily as successful and rigorous modes of explanation. Note that this is a basis for a pluralism
that extends beyond those approaches that we agree with. We can accept that a mode of explanation is
progressing under this criterion even if we think its ontology is flawed and
that its conclusions are fallacious.
Indeed, one can hope that if it continues to pursue a broader range of
successful explanation, it will either eliminate the source of the fallacies,
or it will develop an explanation that shows why we have misunderstood the
question all along. This, then, is Lawson's
critical pragmatism. I naturally
refrain from systematising it, since in that case it would be my pragmatism
rather than his. Its core is the
epistemology that Lawson works out as a side-effect of bridging the gap
between his critical realism and a potentially scientific practice of
economics. I conjecture that it is
narrower in scope than the virtual blank slate offered by the concept of the
open system, but broader in scope than the social ontology constructed in
terms of his realist transformational model of social activity. References Dequech, David. (2002). "The demarcation between the "old" and
the "new" institutional economics: Recent Complications". Journal of
Economic Issues; June 2002, vol. 36, pp. 565-572. Lanis, Roman and Bruce R. McFarling. (2004) " Healthy Economics Healing
Autistic Accounting Theory: Visiting a Neglected Area of Institutional
Economics." Journal of Economic Issues. Vol. 38 (1). March 2004.
pp. 59-83. Lawson,
Tony. (2003). Reorienting Economics. Title in Economics as Social
Theory, Tony Lawson, ed. London: Routledge.
383pp. McFarling, Bruce R. (2004)
"The Clarence Ayres Memorial Lecture: An Institutionalist
Reconstruction of Culture." Journal of Economic Issues.
Vol. 38 (2), June 2004. pp. 339-52. Posner, Richard A.
(1995) "The New Institutional Economics Meets Law and Economics." Overcoming
Law. Harvard University Press: Cambridge, Mass. pp. 426-443. Veblen, Thorsten.
(1898) "Why is Economics Not an Evolutionary Science" The
Quarterly Journal of Economics. Vol. 12 (4), July 1898. pp. 373-397. Vromen, Jack. (2004)
“Conjectural Revisionary Ontology ”, post-autistic economics review,
No. 29, 6 December 2004. article 4, http://www.btinternet.com/~pae_news/review/issue29.htm
|