In the last few weeks I have  read Dava Salyor's Gallileo biography, reread
two great science classics, Gamov's Birth and Death of the Sun (orginally
published in 1939), David Bohm's great Quantum Mechanics textbook (1954) and
took a look at my son's current cell biology textbook.

I must say, if one criteria for science is a field that produces a
continuous (but dialetic) accretion of knowledge throught theoretical
development and empirical testing, the track record of economics is pretty
pathetic compared to physics and biology.  For example, comparing my son's
cell biology textbook to the one that I used 30 years ago, there is an order
of magnitude advance in the detail, comprehensiveness, theoretical
foundation and empirical evidence of the knowledge presented.  And this is
an undergraduate class!  Can anything like this be said of undergraduate
textbooks in economics.

Both Gallileo and Gamov have some very unflattering things to say about the
"philosophers" of there time, as did Marx.  I think the economics of the
last 50 years is a lot closer to this kind of philosophy than to anything
that looks like science. 

-----Original Message-----
From: ravi [mailto:[EMAIL PROTECTED]]
Sent: Monday, November 26, 2001 1:17 PM
To: Pen-L Mailing List
Subject: [PEN-L:19959] If Economics isn't Science, What is it?



recently someone forwarded a bit to the list about the Nobel prize in
economics and whether it was meaningful, etc. reading through klemke,
hollinger, rudge: introductory readings on the philosophy of science, i
came across a piece that might be of interest. i have reproduced
sections from the piece below. my apologies if this is old or
uninteresting stuff:

If Economics isn't Science, What is it? -Alexander Rosenberg

In a number of papers, and in "Microeconomic Laws", I argued that
economic theory is a conceptually coherent body of causal general
claims that stand a chance of being laws. My arguments elicited no
great sigh of relief among economists, for they are not anxious about
the scientific respectability of their discipline. But others eager to
adopt or adapt microeconomic theory to their own uses have appealed to
these and other arguments which attempt to defend economic theory from
a litany of charges that are as old as the theory itself. Among these
charges, the perennial ones were those that denied to economic theory
the status of a contingent empirical discipline because it failed to
meet one or another fashionable positivist of Popperian criterion of
scientific respectability. With the waning of positivism these charges
have seemed less and less serious to philosophers, although they have
retained their force for the few economists still distracted by
methodology. But among philosophers charges that economics does not
measure up to standards for being a science have run afoul of the
general consensus that we have no notion of science good enough to
measure candidates against. This makes it difficult to raise the
question of whether economics is a science, and tends to leave
economists, and their erstwhile apologists like me, satisfied with the
conclusion that since there is nothing logically or conceptually
incoherent about economics, it must be a respectable empirical theory
of human behavior and/or its aggregate consequences.

The trouble with this attitude is that it is unwarrantably complacent.
It is all well and good to say that economics is conceptually coherent,
and that there are no uncontroversial standards against which economics
may be found wanting, but this attitude will not make the serious
anomalies and puzzles about economic theory go away. These puzzles
surround its thoroughgoing predictive weakness. the ability to predict
and control may be neither necessary nor sufficient criteria for
cognitively respectable scientific theories. But the fact is that
microeconomic theory has made no advances in the management of economic
processes since its current formalism was first elaborated in the
nineteenth century. And this surely undermines a complacent conviction
that the credentials of economics as a science are entirely in order.
For a long time after 1945 it might confidently have been said that
Keynesian macroeconomics was a theory moving in the right direction:
although a macro theory, it would ultimately provide the sort of
explanatory and predictive satisfaction characteristic of science. But
the simultaneous inflation and unemployment levels of the last decade
and the economy's imperviousness to fiscal policy have eroded the
layman's and the economist's confidence in the theory. Moreover the
profession's reaction to the failures of Keynesian theory is even more
disquieting to those who view economic theory as unimpeachably a
scientific enterprise. For a large part of the response to its failures
has been a return the microeconomic theories which it was sometimes
claimed to supersede. The diagnosis offered for the failure of the
Keynesian theory has been that it does not accord individual agents the
kind of rationality in the use of information and the satisfaction of
preferences that neoclassical microeconomic theory accords them. The
alternative offered to Keynesian theory in the light of this result is
nothing more nor less than a return to the status quo ante, to the
neoclassical theory of Walrus, Marshall and the early Hicks, that
Keynesianism had preempted.  This cycle brings economic theory right
back to where it was before 1937, and it should seriously undermine the
confidence of anyone's beliefs that economics is an empirical science,
with aims and standards roughly identical to other empirical sciences.
For the twentieth-century history of economic theory certainly does not
appear to be that of an empirical science.

Of course, eighty years is not a long time in the life of a science, or
even a theory, so the fact that economics has not substantially
changed, either in its form or in its degree of confirmation, since
Walras, or arguably since Adam Smith, is no reason to deny its
scientific respectability.  But it is reason to ask why economics has
not moved away from the theoretical strategies that have characterized
it at least since 1874, in spite of their practical inapplicability to
crucial matters like the business cycle, economic development, or
stagflation. On some views of proper scientific method, of course,
economists have been doing just what they should be doing. Since the
19th century they have been pursuing a single research strategy, acting
in accordance with a ubiquitous and powerful paradigm. For, economists
have been steadily elaborating a theory whose form is identical to that
of the great theoretical breakthroughs in science since the 16th
century.  Accordingly it may be argued that it would be irrational for
economists to surrender this strategy short of a conclusive
demonstration that it is inappropriate to the explanation of economic
activity. The strategy is that of viewing the behavior economists seek
to explain as reflecting forces which always move toward stable
equilibria that maximize or minimize some theoretically crucial
variable. In the case of microeconomics, this crucial variable is
utility (or its latter-day surrogates), and the equilibrium is given by
a level of price in all markets that maximizes this variable. This
strategy is most impressively exemplified in Newtonian mechanics and in
the Darwinian theory of natural selection. It is no surprise that a
strategy which serves so well in these two signal accomplishments of
science should have as strong a grip in other domains to which it seems
applicable.  Moreover, the constraints on theoretical and empirical
developments that this strategy imposes can explain many of the
greatest successes of Newtonian and Darwinian science, and much of the
puzzling character of developments in economic theory.

I call this strategy the extremal strategy, because it is especially
apparent in Newtonian mechanics when that theory is expressed in
so-called extremal principles, according to which a system's behavior
always minimizes or maximizes variables reflecting the mechanically
possible states of the system. In the theory of natural selection this
strategy assumes that the environment acts so as to maximize fitness.
This strategy is crucial to the success of these theories because of
the way it directs and shapes the research motivated by them.

<...>

These theories are all committed to explain everything in their domains
because of their extremal character. <...> An extremal theory cannot be
treated as only a partial account of the behavior of objects in its
domain, or as enumerating just some of the many determinants of its
subject's states; for any behavior that actually fails to maximize or
minimize the value of the privileged variable simply refutes the theory
tout court. In fact, the pervasive character of extremal theories
insulates them from falsification to a degree absent from non-extremal
theories. <...> The axioms of theories like Newton's, or Darwin's, or
Walras' do not embody even implicit ceteris paribus clauses. With these
theories the choice is always between rejecting the auxiliary
hypotheses - the description of test conditions - or reject the theory
altogether.

<..>

Extremal theories are an important methodological strategy because they
are so well insulated from falsification. This has enabled them to
function at the core of research programs, turning what otherwise might
be anomalies and counter-instances into new predictions and new
opportunities  for extending their domains and deepening their
precision. Accordingly, it may be argued, economists' attachment to
their extremal theory represents not complacency, but a well-grounded
methodological conservatism.

<...>

But this conservative rationale for the attachment of economists to
extremal theories is vitiated by a crucial disanology between
microeconomics and mechanics or evolution. Economists would indeed be
well advised not to surrender their extremal research program, if only
they could boast even a small part of the startling success that other
extremal research programs have achieved. But two hundred years of work
in the same direction have produced nothing comparable to the
physicists' discovery of new planets, or of new technologies by which
to control the mechanical phenomena that Newton's laws systemized.
<...> There has been no signal success of economic theory akin to these
advances of extremal theory. <...> There is, of course, a vast
literature on why economics has so little in the way of predictive
content, and on how a theory so dependent on idealizations and
factually false assumptions as microeconomics can nevertheless
constitute a respectable scientific enterprise. This literature goes
back to John Stuart Mill and forward to, for example, Hal Varian. The
only two things clear about this literature are that economists have
found it almost universally satisfying and legitimating, and
non-economists have consistently been left unsatisfied, insisting that
methodological excuses are no substitute for attempting to do what
economics has hitherto not done: improve its predictive content.

Having shaken free from the complacent attitude toward economic theory
evinced in Microeconomic Laws, I have come to think that the failure of
economics is not methodological, or conceptual, but very broadly
empirical.  Despite its conceptual integrity, microeconomics, together
with all the sciences of human action and its aggregation, rests on a
false but central conviction that vitiates its axioms and so bedevils
the theorems deduced from them. Economic theory assumes that the
categories of preference and expectation are the classes in which
economic causes are to be systematized, and that the events to be
explained are properly classified as actions like buying, selling, and
the movements of markets, industries, and economies that these actions
aggregate to. The theory has made this assumption, because of course it
is an assumption we all make about human behavior; our behavior
constitutes action and is caused by the joint operation of our desires
and beliefs.

<...>

The real trouble with economics, the real source of its failure to find
improvable laws of economic behavior, is something that has only become
clear in philosophy's recent attempts to understand and improve the
foundations of another science in trouble: psychology, and particularly
behavioral and cognitive psychology. Philosophers have shown that the
terms in which ordinary thought and the behavioral sciences describe
the causes and effects of human action do not describe "natural kinds",
they do not divide nature at the joints. They do not label categories
of states that share the same manageably small set of causes and
effects, and so cannot be brought together in causal generalizations
that improve on our ordinary level of prediction and control of human
actions, let alone attain the sort of continuing improvement
characteristic of science. <..> This work has been devoted to
understanding "intentional" terms like "belief", desire", "action", and
their vast hoard of cognates. <...> The trouble with beliefs and
desires is that when people have them the propositions they "contain"
need not be true or false and the objects they are directed at need not
exist or even be possible objects. So we can't decide whether a person
is in a given mental state by determining either the truth or falsity
of any statement open to our confirmation, or the existence or
attainment of any object or end of human action. <section on problems
of behaviorism and identity theory>

<...>

Applying the new orthodoxy in the philosophy of psychology, it becomes
clear that economics' predictive weakness hinges on the intentional
typology of the phenomena it explains and the causes it identifies. Its
failure to uncover laws of human behavior is due to its wrongly
assuming that these laws will trade in desires, beliefs, or their
cognates. <...> Thus the failure of economics is traced not to a
conceptual mistake, or to the inappropriateness of extremal theories
and their elegant mathematical apparatus to human action, but to a
false assumption economists share with all other social scientists,
indeed with everyone who has ever explained their own or others'
behavior by appeal to the operation of desires and beliefs.

<...>

A better comparison for economic theory than Mendelian genetics is
phlogiston theory, whose failure is traceable to its incommensurability
with the oxygen theory that superseded it. Phlogiston theory is a
scientific dead end, because there is no such thing as phlogiston,
because the notion of phlogiston does not divide nature at any joint.
Phlogiston is not a natural kind. This is an empirical fact about
nature, and the claim that intentional notions are not natural kinds is
equally a contingent claim. Thus, economics and phlogiston theory are
not methodologically defective. They are simply false.

<...>

Bringing what little utility there is in economic theory into harmony
with my diagnosis of its ills is a task for which there are many
solutions. The first thing to note is this: although the laws of supply
and demand and other market-level general statements are deduced from
claims about the intentional determinants of individual actions, they
are logically separable from such claims, and, more important, they can
be shown to follow from assumptions which are the direct denial of
these general claims about rational action. From the assumption that
individuals behave in purely habitual ways, always purchasing the same
or the most nearly similar bundle of commodities available, no matter
what the price, the law of downward sloping demand follows, as it does
from the assumption that their purchases are all impulsively random.
The same can be shown for the choices of entrepreneurs. So surrendering
the extremal intentional approach to human behavior does not logically
or even theoretically oblige us to surrender these "laws". <...> Now
the fact that we can usefully employ false or vacuous general
statements, up to certain limits, is no mystery in the philosophy of
science at all. The clearest instance of such restrictedly useful
though false or vacuous general statements is Euclidean geometry.

<...>

Of course, economic theory has attained nothing like the success of
Euclidean geometry. But the apparent applicability of some of its
claims is to be explained by appeal to the same factors which explain
why we can employ, e.g., the Pythagorean theorem, even though there are
no Euclidean triangles and no Euclidean straight lines. We can employ
the laws of supply and demand, even though human beings are not
economically rational agents; that is, we can employ these "laws" even
though individuals do not make choices reflecting any empirical
regularity governing their expectations and their intentions.  We can
employ them all right, but the laws of supply and demand cannot be
applied with the usefulness and exactitude of the Pythagorean theorem,
just because the kind terms of economic theory are different from the
real kinds in which human behavior is correctly classified.  And this
difference is comparatively much greater than the difference between
the kind terms of applied geometry and those of physics.  There are no
Euclidean triangles, but we know why, and we can calculate the amount
of the divergence between any physical triangle and the Euclidean
claims about it, because we have a physical theory to make these
corrections, the very one which showed Euclidean geometry to be
factually false.  We can make no such improvements in the application
of the laws of supply and demand; we can never do any better than apply
them retrospectively or generically; we cannot specify their
parameters, or their exceptions, because the axiomatic system in which
they figure diverges from the facts very greatly, and because we have
no associated theory that enables us to measure this divergence and
make appropriate corrections for it.

<...>

Euclidean geometry was once styled the science of space, but calling it
a science did not make it one, and we have come to view advances in the
axiomatization and extension of geometry as events not in science, but
in mathematics. Economics is often defined as the science of the
distribution of scarce resources, but calling it a science does not
make it one.

<...>

Much of the mystery surrounding the actual development of economic
theory - its shifts in formalism, its insulation from empirical
assessment, its interest in proving purely formal, abstract
possibilities, its unchanged character over a period of centuries, the
controversies about its cognitive status - can be comprehended and
properly appreciated if we give up the notion that economics any longer
has the aims or makes the claims of an empirical science of human
behavior. Rather we should view it as a branch of mathematics, one
devoted to examining the formal properties of a set of assumptions
about the transitivity of abstract relations: axioms that implicitly
define a technical notion of "rationality", just as geometry examines
the formal properties of abstract points and lines. This abstract term
"rationality" may have far more potential interpretations than
economists themselves realize, but rather less bearing on human
behavior and its consequences than we have unreasonably demanded
economists to reveal.

<...>

We should neither attach much confidence to predictions made on its
basis nor condemn it severely when these predictions fail. For it can
no more be relied on or faulted than Euclidean geometry should be in
the context of astrophysics. Admittedly this attitude leaves a vacuum
in the foundations of public policy. For without economics we lose even
the illusion that we understand the probable, or potential, long-term
or merely possible consequences of choices that policy makers are
forced to make. Of course, the caution that loss of illusions may
foster is certain to be salubrious. On the other hand the vacuum may
attract a really useful foundation for decisions about the economy and
its improvement.

Reply via email to