Content uploaded by D. Wade Hands
Author content
All content in this area was uploaded by D. Wade Hands on Sep 27, 2020
Content may be subject to copyright.
P1: JZP
9780521883504c10 CUFX202/Hausman 978 0 521 88350 4 July 11, 2007 0:54
TEN
Popper and Lakatos in Economic Methodology1
D. Wade Hands
D. Wade Hands (1951– ) was educated at the University of Houston and then
Indiana University and has taught at the University of Puget Sound since 1980. He
is one of the leading figures in contemporary economic methodology. Hands was
President of the History of Science Society in 2005–2006 and is currently the editor of
The Journal of Economic Methodology. His most important book, Reflection without
Rules, won the Spengler Book Prize from the History of Economic Society in 2004.
This essay provides a brief introduction to the ideas of Karl Popper and Imre Lakatos
and to the issues that arise in applying them to the philosophical understanding of
economics.
Overview
The purpose of this chapter is to critically reappraise the methodologicalAu: Reset
abstract. OK as
set? advice offered to economists by Popperian philosophy, in particular Pop-
perian falsificationism and Lakatos’s ‘methodology of scientific research
programmes’. These two philosophical positions and the difficulties they
raise for economic methodology are carefully considered in the chapter. It is
argued that while economists have benefited from the influence of Poppe-
rian philosophy in a number of ways, neither falsificationism nor Lakatos’s
methodology provide an appropriate guide to the acceptance or rejection of
economic theories. The implications and caveats surrounding this argument
are considered in the conclusion.
Pages 61–75 of Rationality, Institutions and Economic Methodology, edited by Uskali M¨
aki, Bo
Gustafsson, and Christian Knudsen. London: Routledge. Copyright c
1993 by Uskali M¨
aki,
Bo Gustafsson and Christian Knudsen. Reproduced by permission of Taylor and Francis
Books UK.
188
P1: JZP
9780521883504c10 CUFX202/Hausman 978 0 521 88350 4 July 11, 2007 0:54
Popper and Lakatos in Economic Methodology 189
Introduction
Popperian philosophy of science has been extremely influential in economic
methodology. Popperian ‘falsificationism’, first introduced into economics
by Hutchison (1938), remains one of the dominant approaches to economic
methodology. In addition to this direct influence, Popperian philosophy has
also affected economic methodology through the work of Imre Lakatos. A
fairly extensive literature has developed around the question of the applica-
bility of Lakatos’s ‘methodology of scientific research programmes’ (MSRP)
to economics.2
The purpose of this chapter is to critically reappraise the methodological
advice given by Popperian philosophy. In this reappraisal both Popperian
falsificationism and Lakatos’s MSRP will be examined. Neo-institutionalist
economics will not be explicitly discussed; instead the focus will be the
general standards for economic theory choice which influence every eco-
nomic theory (including neo-institutionalism). Throughout the discussion
the philsophical positions will be appraised only with respect to economic
methodology: ‘economic’ in that only economics and not other fields of
enquiry will be discussed, and ‘methodology’ in that only questions of the-
ory choice and theory appraisal (not more general philosophical considera-
tions) will be examined. In particular, questions such as whether ‘economic
methodology’ should be pursued at all (recently raised by McCloskey(1985))
will not be examined here.
Falsificationism
No doubt Karl Popper is best known for his falsificationist approach to the
philosophy of science: a theory first presented in Logik der Forschung in
1934 (English translation, Popper 1959). Falsificationism represents Pop-
per’s view of the growth of scientific knowledge as well as his solution to
(or dissolution of) the problem of induction. It is for falsificationism that
Popper claims responsibility for the death of logical positivism (Popper
1976b: 88).
Popperian falsificationism is actually composed of two separate theses:
one on demarcation (demarcating science from non-science) and one on
methodology (how science should be practised). The demarcation thesis is
that for a theory to be ‘scientific’ it must be at least potentially falsifiable by
empirical observation, that is, there must exist at least one empirical basic
statement which is in conflict with the theory.3This potential falsifiability
P1: JZP
9780521883504c10 CUFX202/Hausman 978 0 521 88350 4 July 11, 2007 0:54
190 D. Wade Hands
is a logical relationship between the theory and a basic statement; in par-
ticular, the demarcation criterion only requires that it be logically possible
to falsify the theory, not that such a falsification has ever been attempted.4
While Popper’s demarcation criterion has been the subject of an exten-
sive debate in the philosophical literature, demarcation is seldom the issue
in economics. For economists the more important issue is methodology
(choosing between/among theories not merely labelling them scientific or
unscientific) and Popperian methodology requires the practical (not just
logical) falsifiability of scientific theories.
In a nutshell, falsificationist scientific practice proceeds as follows. The
scientist starts with a scientific problem situation (something requiring a
scientific explanation) and proposes a bold conjecture which might offer a
solution to the problem. Next the conjecture is severely tested by comparing
its least likely consequences with the relevant empirical data. Popper’s argu-
ment for severe testing is that a test will be more severe the more prima facie
unlikely the consequence that is being tested; the theory should be forced to
‘stick its neck out’, to ‘offer the enemy, namely nature, the most exposed and
extended surface’ (Gellner 1974: 171). The final step in the falsificationist
game depends on how the theory has performed during the testing stage. If
the implications of the theory are not consistent with the evidence, then the
conjecture is falsified and it should be replaced by a new conjecture which is
not ad hoc relative to the original, that is, the new conjecture should not be
contrived solely to avoid this empirical anomaly.5If the theory is not falsified
by the evidence then it is considered corroborated and it is accepted pro-
visionally. Given Popper’s fallibilism this acceptance is provisional forever;
the method does not necessarily result in true theories, only ones that have
faced a tough empirical opponent and won.
Now while there are a number of reasons why economists have felt that
Popperian falsificationism would be a desirable methodology, the fact is that
falsificationism is seldom if ever practised in economics. This seems to be the
one point generally agreed upon by recent methodological commentators.
In fact, this (empirical) claim is supported at length by the case studies
in Blaug (1980), a book which consistently advocates falsificationism as a
normative ideal. The disagreement between critics and defenders of falsifi-
cationism is not whether it has been practised, basically it has not, but rather
whether it should be practised. The real questions are whether the profession
should ‘try harder’ to practise falsificationism though it has failed to do
so in the past, and the related question of whether the discipline of eco-
nomics would be substantially improved by a conscientious falsificationist
practice.
P1: JZP
9780521883504c10 CUFX202/Hausman 978 0 521 88350 4 July 11, 2007 0:54
Popper and Lakatos in Economic Methodology 191
One approach to the question of the appropriateness of falsificationism
in economics would be to directly address the question of the adequacy of
Popper’s falsificationist methodology as a general approach to the growth
of scientific knowledge; this is not the approach that will be followed here.
Rather than delving into this general question, the following discussion will
simply survey some of the criticisms which falsificationism has received
explicitly as an economic methodology. This list of criticisms is not exhaus-
tive, but it does capture the major concerns which have been raised regard-
ing the falsificationism in economics. The list is not necessarily in order of
importance.6
1. For a number of reasons, the so-called Duhemian problem (or Duhem–
Quine problem) presents a great difficulty in economics.7First, the
complexity of human behaviour requires the use of numerous ini-
tial conditions and strong simplifying assumptions. Some of these
restrictions may actually be false (such as the infinite divisibility of
commodities), some of these assumptions may be logically unfalsi-
fiable (such as the assumptions of eventually diminishing returns),
while still others may be logically falsifiable but practically unfalsifi-
able (such as the completeness assumption in consumer choicetheor y).
Even where assumptions and restrictions can be tested, such testing is
very difficult because of the absence of a suitably controlled laboratory
environment.8In the presence of such a variety of restrictions it is vir-
tually impossible to ‘aim the arrow of modus tollens’ at one particular
problematic element of the set auxiliary hypotheses when contrary evi-
dence is found. Second, there are many questions and disagreements
about the empirical basis in economics. It is always possible to argue
that what was observed was ‘not really’ involuntary unemployment or
‘not really’ economic profit, etc. Although it is fundamental to Pop-
perian philosophy that the empirical basis need not be incorrigible,
it is necessary that there be a generally accepted convention regard-
ing the empirical basis,9and in economics even such conventions are
often not available. Third, even if these first two problems have some-
how been eliminated it is still possible for the social sciences to have
feedback effects that do not exist in the physical sciences. The test of
an economic theory may itself alter the initial conditions for the test.
Conducting a test of the relationship between the money supply and
the price level may alter expectations in such a way that the initial
conditions (which were true ‘initially’) are not true after the test (or if
the ‘same’ test were conducted again).
P1: JZP
9780521883504c10 CUFX202/Hausman 978 0 521 88350 4 July 11, 2007 0:54
192 D. Wade Hands
2. Related to, but actually separate from the Duhemian problem, is the
problem that the qualitative comparative statics technique used in eco-
nomics makes severe testing very difficult and cheap corroborational
success ‘too easy’. In economics it is very often the case that the strongest
available prediction is a qualitative comparative statics result which
only specifies that the variable in question increases, decreases, or
remains the same. Since having the correct sign is much easier than
having both the correct sign and magnitude, an emphasis on such
qualitative prediction generates theories which are low in empirical
content, have few potential falsifiers, and are difficult if not impos-
sible to test severely. The result is often economic theories which are
confirmed by the evidence but provide very little information.10
3. Popper’s ‘admitted failure’ (1983: xxxv) to develop an adequate theory
of verisimilitude11 presents a fundamental difficulty for a falsification-
ist methodology in economics. Popper’s theor y of verisimilitude devel-
oped as an attempt to reconcile his falsificationist methodology with
scientific realism. For a realist science aims at ‘true’ theories; according
to falsificationism, scientific theories should be chosen if they have been
corroborated by passing severe tests. If the falsificationist method is to
fulfil the realist aims of science it should be demonstrated that more
corroborated theories are closer to the truth. Such a demonstration
was precisely the goal of Popper’s theory of verisimilitude. Actually a
satisfactory theory of verisimilitude would serve Popperian philosophy
in at least two different ways. One way, as already mentioned, would
be to provide an epistemic justification for playing the game of science
by falsificationist rules. Such a justification is very important for Pop-
perian philosophy since without a theory of verisimilitude it can be
argued that there are philosophically ‘no good reasons’ (Popper 1972:
22) for choosing theories as Popper recommends. The second function
of a theory of verisimilitude is more practical. Verisimilitude would
provide rules for choosing the ‘best’ theory in troublesome cases: like
the situation where both available theories have been falsified. A theory
of verisimilitude would help in such cases because it would provide
a rule for determining which of the two theories in question actually
has more verisimilitude: which is closer to the truth. A similar argu-
ment could be made for cases involving a choice between a falsified
but bold theory and a corroborated but modest theory; having a way
to determine which has more verisimilitude would allow us to choose
a theory which is more consistent with the aims of science, which is
P1: JZP
9780521883504c10 CUFX202/Hausman 978 0 521 88350 4 July 11, 2007 0:54
Popper and Lakatos in Economic Methodology 193
closer to the truth. This second, more practical, function of the theory
of verisimilitude is very important in economic methodology. The rea-
son is that economists are almost always faced with choosing between
two falsified theories, or choosing between a bold falsified theory and
a more modest corroborated one. If Popper’s theory of verisimilitude
had been a success and it could be added to the norms of simple fal-
sificationism (both to justify the norms and to help in making the
practical decisions of theory choice) then falsificationism might have
an important role to play in economic theory choice. Without such a
link between severe testing and truth-likeness, the method is of limited
value in pursuing the realist aim of science.
4. Popper’s rules for progressive theory development (non ad hocness)
are seldom appropriate in economics. Popper argues that if one theory
is to constitute ‘progress’ over a predecessor the new theory must be
‘independently testable’; it must have ‘excess empirical content’, pre-
dict ‘novel facts’.12 This issue will be examined more carefully in the
Lakatos section which follows, but for now it should be noted that
while Popperian progress may sometimes be of interest to economists,
often progress in economics is (and should be) very different to what
Popper prescribes. Economists are often concerned with finding new
explanations for well-known (non novel) facts, or alternatively, with
explaining known phenomena by means of fewer theoretical restric-
tions. What constitutes ‘progress’ in economic theory (or what should
constitute progress) is a complex and ongoing question, but it is appar-
ent that any suitable answer will require a different, and possibly much
more liberal, set of standards than those offered by strict Popperian
falsificationism.
All of these criticisms add up to a negative appraisal of falsificationist eco-
nomic methodology. Despite the fact that preaching falsificationist method-
ology has been very popular among economists, the method fails to provide
a reasonably adequate set of rules for doing economics. Strict adherence to
falsificationist norms would virtually destroy all existing economic theory and
leave economists with a rule book for a game unlike anything the profession
has played in the past. This high cost would be paid without any guarantee
that obeying the new rules would result in theories any closer to the truth
about economic behaviour than those currently available. How this result
should be interpreted will be discussed in the conclusion, for now let us turn
to Lakatos’s MSRP.
P1: JZP
9780521883504c10 CUFX202/Hausman 978 0 521 88350 4 July 11, 2007 0:54
194 D. Wade Hands
Lakatos’s Methodology of Scientific
Research Programmes
Lakatos’s work in the philosophy of science first appeared in the early 1970s
(Lakatos 1970, 1971) and it was endorsed almost immediately by a number
of economists. Numerous papers on Lakatoshave appeared in the economics
literature, many as a result of the Nafplion Colloquium on Research Pro-
grammes in Physics and Economics in 1974 (Latsis 1976a). This literature
on ‘Lakatos and economics’ has basically been of two types. The first type
is historical, it attempts to ‘reconstruct’ some particular episode in the his-
tory of economic thought along Lakatosian lines. The second type is more
philosophical, it attempts to appraise Lakatos’s methodology of scientific
research programmes as an economic methodology and/or compare it to
other philosophies such as Kuhn or Popperian falsificationism.
Lakatos’sMSRP is clearly par t of the Popperian tradition in the philosophy
of science but it was also motivated by philosophically minded historians of
science such as Kuhn (1970). For Lakatos the primary unit of appraisal in sci-
ence is the ‘research programme’ rather than the scientific theory. A research
programme is an ensemble consisting of a hard core, the positive and neg-
ative heuristics, and a protective belt.13 The hard core is composed of the
fundamental metaphysical presuppositions of the programme; it defines the
programme, and its elements are treated as irrefutable by the programme’s
practitioners. To participate in the programme is to accept and be guided by
the programme’s hard core. For example, in Weintraub’s Lakatosian recon-
struction of the Neo-Walrasian research programme in economics, the hard
core consists of propositions such as: agents have preferences over outcomes
and agents act independently and optimize subject to constraints. The pos-
itive and negative heuristics provide instructions about what should and
should not be pursued in the development of the programme. The positive
heuristic guides the researcher toward the right questions and the best tools
to use in answering those questions; the negative heuristic indicates what
questions should not be pursued and what tools are inappropriate. Again
using Weintraub’s analysis of the Neo-Walrasian programme as an example,
the positive heuristic contains injunctions such as: construct theories where
the agents optimize, while the negative heuristic implores researchers to
avoid theories involving disequilibrium. Finally, the protective belt consists
of the programme’s actual theories, auxiliary hypotheses, empirical con-
ventions and the (evolving) ‘body’ of the research programme. The major
activity of the programme occurs in the protective belt, it occurs as a result
of the interaction of the hard core, the heuristics, and the programme’s
P1: JZP
9780521883504c10 CUFX202/Hausman 978 0 521 88350 4 July 11, 2007 0:54
Popper and Lakatos in Economic Methodology 195
empirical record. For Weintraub’s Neo-Walrasian programme the protec-
tive belt includes almost all of applied microeconomics.
A research programme is appraised on the basis of the theoretical and
empirical activity in the protective belt. There is theoretical progress if each
change in the protective belt is empirical content increasing; that is if it
predicts novel facts.14 The research programme exhibits empirical progress
if this excess empirical content actually gets corroborated (Lakatos 1970:
118). Lakatos also requires a third type of progress, heuristic progress (non-
ad hoc3ness), which specifies that the changes be consistent with the hard
core of the programme. Lakatos’s definitions of theoretical and empirical
progress presuppose that the changes in question are consistent with heuris-
tic progress.
One obvious example of the link between Lakatos and Popper is the way
in which Lakatos characterizes empirical content and novel facts. Lakatos,
like Popper, defines the empirical content of a theory to be ‘the set of its
potential falsifiers: the set of those observational propositions which may
disprove it’ (Lakatos 1970: 98, n. 2). Thus, even though Lakatos considers
empirical progress to come through empirical corroboration rather than
falsification, his characterization of the relationship between theory and
fact is still basically falsificationist. There are many other signs of Lakatos’s
Popperian lineage but his definition of empirical content and novel facts
are the most important in the appraisal of Lakatosian economic method-
ology.
On the other hand, there are many aspects of the MSRP which are fun-
damentally at odds with Popperian falsificationism. The most significant of
these is the immunity of the hard core to empirical criticism; immunizing
any part of scientific theory would be in conflict with Popper’s falsifica-
tionist method of bold conjecture and severe test. Popper clearly recognized
that science has experienced periods of Kuhnian ‘normal science’ where
the critical spirit seems to be temporarily arrested, but for Popper these
episodes are something to lament not praise (Popper 1970). Another point
of disagreement is the question of corroboration versus falsification. While
Lakatos defines empirical content in a thoroughly Popperian way, he has
no respect for the role of falsification in science. For Lakatos all theories are
‘born refuted’ (1970: 120–1) and the task of philosophy of science should be
to develop a methodology which starts from this fact. For Lakatos progress
comes from the corroboration not falsification of novelfacts. Finally, Lakatos
clearly embraces a historical meta-methodology whereby the actual history
of science is used to appraise various methodological proposals.15 This is
very different from Popper where methodology is purely a normative affair
P1: JZP
9780521883504c10 CUFX202/Hausman 978 0 521 88350 4 July 11, 2007 0:54
196 D. Wade Hands
and where there is no pathway open for the actual history of science to help
evaluate methodologies.
These places where Lakatos differs from Popper are exactly the places
where Lakatos is likely to win the favour of economists since these hap-
pen to be areas where there is substantial tension between falsificationism
and the actual practice of economics. Certainly economics is replete with
metaphysical ‘hard cores’; there is not much consensus on what these hard
core propositions should be, but there seems to be a consensusthat such hard
core presuppositions exist and that they often define alternative research
programmes in economics. A philosophical programme such as Popperian
falsificationism which requires practitioners to be willing to give up almost
any part of their research programme at any time will not provide as
adequate a guide for economists as Lakatos’s methodology which allows
for such pervasive hard cores. This economic preference for Lakatos over
Popper also extends to the issue of corroboration versus falsification. It is
clear that falsificationism has not been practised in economics and there is
good reason to believe that enforcement of such strict standards would all
but eliminate the discipline as it currently exists. On the other hand, there is
a great amount of empirical activity in economics, the facts do matter, but
they matter in a much more subtle and complex way than falsificationism
allows.
Finally, economists would prefer Lakatos to Popper on the question of
the role of the history of science in supporting particular methodological
proposals. The general question of the relationship between the history of
science and the philosophy of science is an unsettled question which contin-
ues to be debated in the literature, but economists have recently been very
sympathetic to methodological proposals that are sensitive to the actual his-
tory of their discipline. Economists have produced an extensive literature
using the Lakatosian categories to reconstruct various parts of the history of
economic thought. Most of this literature focuses on a particular research
programme in economic theory (past or present) and tries to isolate the
hard core, the positive and negative heuristics, and the type of theoretical
activity occurring in the protective belt. Such work usually results in a pos-
itive or negative Lakatosian appraisal of the ‘progressivity’ of the particular
economic research programme. Examples of these reconstructions range
widely over various topics in the history of economic thought.
An overall assessment of this Lakatosian historical literature is very diffi-
cult because many of the economists writing in the field have taken very little
care in the way they use the Lakatosian terminology. This lack of fidelity to
Lakatosian terminology has resulted in ‘hard cores’, ‘heuristics’ and (particu-
larly) ‘novel facts’ whichbear little resemblance to their Lakatosian analogues
P1: JZP
9780521883504c10 CUFX202/Hausman 978 0 521 88350 4 July 11, 2007 0:54
Popper and Lakatos in Economic Methodology 197
or how these terms have been used in reconstructions in the physical sciences.
Much of this literature has provided valuable and independently interesting
history of economic thought, but it sheds little light on the methodological
adequacy of the MSRP. The only general conclusion that can be reached
from this historical literature is that in the case studies where the relevant
language is consistent with Lakatos, ‘progress’, and the prediction of novel facts
it necessarily implies, has been a rare occurrence. There have been some well-
researched cases where novel facts actually seem to have been uncovered;16
but these cases correspond to only a very small portion of what the economics
profession would consider its major theoretical ‘advances’. Lakatos’s crite-
rion for ‘theoretical progress’, the prediction of novel facts, may have been
sufficient for what economists have considered to be theoretical progress in
certain special cases, but it does not seem to be generally necessary. Just as
‘the development of economic analysis would look a dismal affair through
falsificationist spectacles’ (Latsis 1976b: 8), it seems that economics would
look almost as dismal on a strictly Lakatosian view.
The argument that empirical and theoretical advances in economics occur
(and should occur) in ways other than Lakatos specified in the MSRP, reflects
very poorly (again) on Popper. The reason is that where economics is most
likely to part ways with Lakatos is precisely where Lakatos borrowed most heav-
ilyfromPopper. In certain respects, Lakatos’s work is much better suited
to economics than Popper’s; it seems that looking for the types of things
which Lakatos suggests one should look for in the history of economics
has helped guide a number of important historical studies. Certainly this
historical research has drawn attention to the metaphysical hard core of cer-
tain economic research programmes and it has motivated enquiry into the
important methodological question of the relationship between empirical
and theoretical work in economics, that is, between econometrics and eco-
nomic theory. What the MSRP does not provide is an appropriate model
for the acceptance or rejection of economic theories. Lakatos’s MSRP may
constitute methodological progress over falsificationism, but it still fails to
provide economists with an acceptable criterion for theory choice (or pro-
gressive problem shifts). This is particularly telling for Popper since the
Lakatosian fit seems to be poorest where older Popperian parts were used
with the least modification.
Conclusion
In the final evaluation it seems that ‘Popperian’economic methodology must
be given low marks. Falsificationism, Popper’s fundamental programme for
the growth of scientific knowledge, is particularly ill-suited to economics
P1: JZP
9780521883504c10 CUFX202/Hausman 978 0 521 88350 4 July 11, 2007 0:54
198 D. Wade Hands
and while the interest in Lakatos has produced some valuable historical
studies,17 the overall fit between economics and the MSRP is not good: and
not good precisely where Lakatos is the most Popperian.
This evaluation should not be too harshly interpreted though. It has been
argued that Popperian methodology, both in its falsificationist and MSRP
forms, does not provide a very good standard for judging the adequacy
of economic theories; this does not mean that Popperian philosophy has
not provided any insight at all into economic theorizing. In particular, the
above argument does not say that testing should be unimportant in eco-
nomics, that Lakatosian reconstructions in the history of economic thought
have not provided valuable contributions to the historical literature, or that
economists would have gained more by listening to some other particular
school of philosophy.
In addition to the above disclaimers it should also be noted that the
discussion has entirely neglected Popper’s writings on the philosophy of
social science: his so-called ‘situational analysis’ approach to social science.18
This method, the method of explaining the behaviour of a social agent on
the basis of the logic of the agent’s situation and the ‘rationality principle’,
was proposed by Popper as a result of ‘the logical investigation of economics’
and it provides a method ‘which can be applied to all social science’ (Popper
1976a: 102). It is often argued that the rationality principle is in conflict
with Popper’s falsificationist standards,19 but regardless of how one views
this controversy, the point here is simply to note that none of the above
criticisms automatically transfer to Popper’s work on situational analysis.
The task of this chapter was narrowly defined: to evaluate falsificationism
and the MSRP as a methodology – as a tool for choosing between/among
economic theories/research programmes. It has been argued that Popperian
philosophy should be negatively appraised in this respect, it does not say
that economists have nothing to learn from the Popperian tradition.
Notes
1. Helpful comments on an earlier draft were received from a number of people;
in particular I would like to mention Bruce Caldwell, Christian Knudsen, Uskali
Maki, and Jorma Sappinen. Partial support for the research was provided by
University of Puget Sound Martin Nelson Award MNSA-4489 and portions of
the argument also appear in Hands (1992). The recent article by Caldwell (1991)
also provides an excellent discussion of these issues.
2. Blaug (1976, 1991), Cross (1982), de Marchi (1976), Diamond (1988), Fisher
(1986), Fulton (1984), Glass and Johnson (1988), Hands (1985b), Latsis (1972,
1976b), Maddock (1984), and Weintraub (1985a,b, 1988), is a partial list of the
P1: JZP
9780521883504c10 CUFX202/Hausman 978 0 521 88350 4 July 11, 2007 0:54
Popper and Lakatos in Economic Methodology 199
work on Lakatosian economics. A more complete list is contained in Hands
(1985b) and (1992).
3. The expression ‘basic statement’ has a rather narrow interpretation in Popperian
philosophy. The concept was introduced in chapter V of Popper (1959) and it
is nicely summarized in Watkins (1984: 247–54).
4. Actually, as will be discussed below, scientific theories are not by themselves
logically falsifiable. Rather, scientific theories along with (usually numerous)
auxiliary hypotheses may form logically falsifiable test systems (see Hausman
1988: 68–9).
5. There are a number of different types of ad hocness in Popperian philosophy;
these are discussed in detail in Hands (1988). The type of ad hocness consid-
ered here, modification solely to avoid falsification, is called ad hoc1.Popper
developed his notion of independent testability to avoid this type of ad hoc-
ness (ad hoc1ness). Another notion of ad hocness is ad hoc2ness; a theoretical
modification is non ad hoc2if some of the independently testable implica-
tions actually get corroborated. A third type of ad hocness (ad hoc3ness) was
developed more fully by Lakatos. Non-ad hoc3ness is equivalent to Lakatosian
heuristic progress.
6. The main sources for this list of criticisms are Caldwell (1984), Hausman (1985,
1988), Latsis (1976b), and Salanti (1987).
7. The Duhemian problem (Duhem 1954) arises because theories are never tested
alone, rather they are tested in conjunction with certain auxiliary hypotheses
(including those about the data). Thus if Tis the theory, the prediction of
evidence eis given by T•A=>e,whereAis the set of auxiliary hypotheses.
The conjunction T•A forms a test system and the observation of ‘not e’ implies
‘not (T•A)’ rather than simply ‘not T’; the test system is falsified, not necessarily
the theory. The Duhemian problem is a standard issue in the philosophy of
theory testing but it has only recently been recognized as an issue for economic
methodology (see Cross 1982, for instance).
8. Experimental economics is still too young to tell whether it can substantially
improve this situation. For a general discussion of the methodological implica-
tions of the literature on experimental economics see Roth (1986), and Smith
(1982, 1985).
9. Popper (1965: 42, 267, 387–8; 1959: 43–4, 93–5, 97–111; 1983: 185–6).
10. This is one source of the ‘innocuous falsification’ mentioned by Blaug (1980:
128, 259) and Coddington (1975: 542–45). The problem of such qualitative (or
generic) predictions is discussed in detail in Rosenberg (1989).
11. Popper’s most important writings on verisimilitude are contained in Popper
(1965) and (1972). Useful discussions of the topic are presented in Koertge
(1979), Radnitzky (1982), and Watkins (1984). The question of the relationship
between Popperian verisimilitude and economic methodology is examined in
more detail in Hands (1991).
12. These concepts are discussed in detail with appropriate references to Popper’s
writings in Hands (1988). Other general discussions of these Popperian concepts
include Dilworth (1986), Koertge (1979), Watkins (1978, 1984), and Worrall
(1978).
P1: JZP
9780521883504c10 CUFX202/Hausman 978 0 521 88350 4 July 11, 2007 0:54
200 D. Wade Hands
13. Many summaries of the MSRP are available in the economics literature (Blaug
(1980), Hands (1985a), Pheby (1988), and Weintraub (1985a, 1985b, 1988)
for example) but the single best presentation of the argument remains Lakatos
(1970). As with Popper’s falsificationism, only a sketch of the main argument is
provided here.
14. The definition of ‘novel fact’ has been much discussed in the Lakatosian (and
Popperian) literature. See Carrier (1988), Gardner (1982), Hands (1985b) and
Worrall (1978) on the different notions of novelty.
15. ‘A general definition of science thus must reconstruct the acknowledgedly best
gambits as “scientific:” if it fails to do so, it has to be rejected’ (Lakatos 1971:
111).
16. Particularly Blaug (1991), Maddock (1984), and Weintraub (1988), though e ven
here it depends on the exact definition of novelty one uses.
17. In addition to those mentioned in note 16, Cross (1982), de Marchi (1976) and
Latsis (1972, 1976b) should be added to this list.
18. Popper’s clearest writings on situational analysis are (1976a) and (1985); also
see Hands (1985a, 1992) and Langlois and Csontos (this volume).
19. According to Popper’s situational analysis view of social science, the action
of an individual agent is explained by describing the ‘situation’ the agent is
in (their preferences, beliefs, constraints, etc.) and the ‘rationality principle’
that all agents act appropriately (rationally) given their situation. The potential
problem arises because the rationality principle serves as the universal law in
such scientific ‘explanations’ and yet it is not clear that the rationality principle
is (even potentially) falsifiable as Popper the falsificationist would require for
the laws used in any valid scientific explanation. This is one of the reasons that
Poppern(Popper the falsificationist) was distinguished from Poppers(Popper
the philosopher of situational analysis) in Hands (1985a).
References
Blaug, M. (1976) ‘Kuhn versus Lakatos, or paradigms versus research programmes in
the history of economics’, in S. J. Latsis (ed.) Method and Appraisal in Economics,
Cambridge: Cambridge University Press, pp. 149–80.
(1980) The Methodology of Economics, Cambridge: Cambridge University Press.
(1987) ‘Ripensamenti sulla rivduzione Keynesiana’, Rassegna Economica, 51:
605–34.
(1991) ‘Second thoughts on the Keynesian revolution’, History of Political Econ-
omy 23: 171–92 (English translation of Blaug 1987).
Caldwell, B. J. (1982) Beyond Positivism: Economic Methodology in the Twent ieth Century,
London:Allen&Unwin.
(1984) ‘Some problems with falsificationism in economics’, Philosophy of the
Social Sciences 14: 489–95.
(1991) ‘Clarifying Popper’, Journal of Economic Literature 29: 1–33.
Carrier, M. (1988) ‘On novel facts: a discussion of criteria for non-ad-hocness
in the methodology of scientific research programmes’, Zeitschrift fur Allgemeine
Wissenschaftstheorie 19: 205–31.
P1: JZP
9780521883504c10 CUFX202/Hausman 978 0 521 88350 4 July 11, 2007 0:54
Popper and Lakatos in Economic Methodology 201
Coddington, A. (1975) ‘The rationale of general equilibrium theory’, Economic Inquiry
13: 539–58.
Cross, R. (1982) ‘The Duhem–Quine thesis, Lakatos, and the appraisal of theories in
macroeconomics’, Economic Journal 92: 320–40.
de Marchi, N. (1976) ‘Anomaly and the development of economics: the case of the
Leontief paradox’, in S. J. Latsis (ed.) Method and Appraisal in Economics, Cambridge:
Cambridge University Press, pp. 109–27.
(ed.) (1988) The Popperian Legacy in Economics, Cambridge: Cambridge Uni-
versity Press.
Diamond, A. M., Jr. (1988) ‘The empirical progressiveness of the general equilibrium
research program’, History of Political Economy 20: 119–35.
Dilworth, C. (1986) Scientific Progress: A Study Concerning the Nature of the Rela-
tions Between Successive Scientific Theories, second edition, Dordrecht, Netherlands:
D. Reidel.
Duhem, P. (1954) The Aim and Structure of Physical Theory, (translated by P. P. Wiener),
Princeton, New Jersey: Princeton University Press.
Fisher, F. M. (1986) The Logic of Economic Discovery, Washington Square, New York:
New York University Press.
Fulton, G. (1984) ‘Research programmes in economics’, History of Political Economy 16:
187–205.
Gardner, M. R. (1982) ‘Predicting novel facts’, British Journal for the Philosophy of Science
33: 1–15.
Gellner, E. (1974) Legitimation of Belief, Cambridge: Cambridge University Press.
Glass, J. C. and Johnson, W. (1988) ‘Metaphysics, MSRP and economics’, British Journal
for the Philosophy of Science 39: 313–29.
Hands, D. W. (1985a) ‘Karl Popper and economic methodology: ANewLook’, Economics
and Philosophy 1: 83–99.
(1985b) ‘Second thoughts on Lakatos’, History of Political Economy 17: 1–16.
(1988) ‘Ad hocness in economics and the Popperian tradition’, in de Marchi, N.
(ed.) The Popperian Legacy in Economics and Beyond, Cambridge: Cambridge Univer-
sity Press, pp. 121–37.
(1991) ‘The problem of excess content: economics, novelty and a long Popperian
Tale’, in N. de Marchi and M. Blaug (eds) Appraising Economic Theories, Aldershot:
Edward Elgar 58–75.
(1992) ‘Falsification, situational analysis and scientific research programs: the
Popperian tradition in economic methodology’, in N. de Marchi (ed.) Post-Popperian
Methodology of Economics, Boston: Kluwer.
Hausman, D. M. (1985) ‘Is falsificationism unpracticed or unpracticable?’, Philosophy of
the Social Sciences 15: 313–19.
(1988) ‘An appraisal of Popperian economic methodology’, in de Marchi, N. (ed.)
The Popperian Legacy in Economics and Beyond, Cambridge: Cambridge University
Press, 65–85.
Hutchison,T. W. (1938) The Significance and B asic Postulatesof Economic Theory,London:
Macmillan (reprinted, 1960, New York: Augustus M. Kelley).
(1988) ‘The case for falsificationism’, in de Marchi, N. (ed.) The Popperian Legacy
in Economics and Beyond, Cambridge: Cambridge University Press, pp. 169–81.
P1: JZP
9780521883504c10 CUFX202/Hausman 978 0 521 88350 4 July 11, 2007 0:54
202 D. Wade Hands
Koertge, N. (1979) ‘The problems of appraising scientific theories’, in P.D. Asquith and H.
E. Kyburg, Jr. (eds) Current Research in Philosophy of Science, East Lansing, Michigan:
Philosophy of Science Association, pp. 228–251.
Kuhn, T. S. (1970) The Structure of Scientific Revolutions, second edition, Chicago: Uni-
versity of Chicago Press. (First edition 1962, Chicago: University of Chicago Press.)
Lakatos, I. (1970) ‘Falsification and the methodology of scientific research programmes’,
in I. Lakatos and A. Musgrave(eds) Crit icism and the Growth of Knowledge , Cambridge:
Cambridge University Press, pp. 91–196.
(1971) ‘History of science and its rational reconstructions’, in R. C. Buck and R. S.
Cohen (eds) Boston Studies in the Philosophy of Science, Vol. 8, Dordrecht, Netherlands:
D. Reidel, pp. 91–136.
Latsis, S. J. (1972) ‘Situational determinism in economics’, British Journal for the Philos-
ophy of Science 23: 207–45.
(ed.) (1976a) Method and Appraisal in Economics, Cambridge: Cambridge Uni-
versity Press.
(1976b) ‘A research programme in economics’, in S. J. Latsis (ed.) Method of
Appraisal in Economics, Cambridge: Cambridge University Press, pp. 1–41.
McCloskey, D. (1985) The Rhetoric of Economics, Madison Wisconsin: University of
Wisconsin Press.
Maddock, R. (1984) ‘Rational expectations macrotheory: A Lakatosian case study in
program adjustment’, History of Political Economy 16: 291–309.
Pheby, J. (1988) Methodology and Economics: A Critical Introduction, London: Macmillan.
Popper, K. R. (1959) The Logic of Scientific Discovery, London: Hutchinson.
(1965) Conjectures and Refutations, second edition, New York: Harper & Row.
(1970) ‘Normal science and its Dangers’, in I. Lakatos and A. Musgrave (eds) Crit-
icism and the Growth of Knowledge, Cambridge: Cambridge University Press, pp. 51–8.
(1972) Objective Knowledge: An Evolutionary Approach, Oxford: Oxford Univer-
sity Press.
(1976a) ‘The logic of the social sciences’, in T. W. Adorno et al. (eds) The
Positivist Dispute in German Sociology (translated by G. Adey and D. Frisby), New
York: Harper & Row, pp. 87–104.
(1976b) Unended Quest, La Salle, Illinois: Open Court.
(1983) Realism and the Aim of Science (edited by W. W. Bartley III) Totowa, New
Jersey: Rowman & Littlefield.
(1985) ‘The rationality principle’, in D. Miller (ed.) Popper Selections, Princeton:
Princeton Universtiy Press, pp. 357–65.
Radnitzky, G. (1982) ‘Knowing and guessing: if all knowledge is conjectural, can we then
speak of cognitive progress? on persistent misreading of Popper’s work’, Zeitschrift f¨ur
Allgemeine Wissenschafstheorie 8: 110–21.
Rosenberg, A. (1989) ‘Are generic predictions enough?’, Erkenntnis 30: 43–68.
Roth, A. E. (1986) ‘Laboratory experimentation in economics’, Economics and Philosophy
2: 245–73.
Salanti, A. (1987) ‘Falsificationism and fallibilism as epistemic foundations of economics:
a critical view’, Kyklos 40: 368–92.
Smith, V. L. (1982) ‘Microeconomic systems as an experimental science’, American Eco-
nomic Review 72: 923–55.
(1985) ‘Experimental economics: reply’, American Economic Review 75: 265–72.
P1: JZP
9780521883504c10 CUFX202/Hausman 978 0 521 88350 4 July 11, 2007 0:54
Popper and Lakatos in Economic Methodology 203
Watkins, J. (1978) ‘The Popperian approach to scientific knowledge’, in G. Radnitzky
and G. Anderson (eds) Progress and Rationality in Science, Dordrecht, Netherlands:
D. Reidel, pp. 23–43.
(1984) Science and Skepticism, Princeton, New Jersey: Princeton University Press.
Weintraub, E. R. (1985a) ‘Appraising general equilibrium analysis’, Economics and Phi-
losophy 1: 23–37.
(1985b) General Equilibrium Analysis: Studies in Appraisal, Cambridge: Cam-
bridge University Press.
(1988) ‘The NeoWalrasian program is empirically progressive’, in de Marchi, N.
(ed.) The Popperian Legacy in Economics and Beyond, Cambridge: Cambridge Univer-
sity Press.
Worrall, J. (1978) ‘The ways in which the methodology of scientific research programmes
improves on Popper’s methodology’, in G. Radnitzky and G. Anderson (eds) Progress
and Rationality in Science, Dordrecht, Netherlands: D. Reidel, pp. 45–70.
P1: JZP
9780521883504c10 CUFX202/Hausman 978 0 521 88350 4 July 11, 2007 0:54
204
