Critical Realism and Econometrics

Bache, Ida Wolden Working Paper Critical Realism and Econometrics Working Paper, No. 2003/4 Provided in Cooperation with: Norges Bank, Oslo Suggested Citation: Bache, Ida Wolden (2003) : Critical Realism and Econometrics, Working Paper, No. 2003/4, ISBN 82-7553-211-6, Norges Bank, Oslo, https://hdl.handle.net/11250/2498623 This Version is available at: https://hdl.handle.net/10419/209818 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by-nc-nd/4.0/deed.no ANO 2003/4 Oslo March 30, 2003 Working Paper Research Department Critical Realism and Econometrics by Ida Wolden Bache 14142-working-paper-4-03 16.05.03 12:49 Side 1 ISSN 0801-2504 ISBN 82-7553-211-6 Working papers from Norges Bank can be ordered by e-mail: [email protected] or from Norges Bank, Subscription service, P. O.Box. 1179 Sentrum N-0107 Oslo, Norway. Tel. +47 22 31 63 83, Fax. +47 22 41 31 05 Working papers from 1999 onwards are available as pdf-files on the bank’s web site: www.norges-bank.no, under "Publications". Norges Bank’s working papers present research projects and reports (not usually in their final form) and are intended inter alia to enable the author to benefit from the comments of colleagues and other interested parties. Views and conclusions expressed in working papers are the responsibility of the authors alone. Working papers fra Norges Bank kan bestilles over e-post: [email protected] eller ved henvendelse til: Norges Bank, Abonnementsservice Postboks 1179 Sentrum 0107 Oslo Telefon 22 31 63 83, Telefaks 22 41 31 05 Fra 1999 og senere er publikasjonene tilgjengelige som pdf-filer på www.norges-bank.no, under "Publikasjoner". Working papers inneholder forskningsarbeider og utredninger som vanligvis ikke har fått sin endelige form. Hensikten er blant annet at forfatteren kan motta kommentarer fra kolleger og andre interesserte. Synspunkter og konklusjoner i arbeidene står for forfatternes regning. 14142-working-paper-4-03 16.05.03 12:49 Side 2 Critical Realism and Econometrics∗ Ida Wolden Bache† Research Department, Norges Bank (Central Bank of Norway) March, 2003 Abstract In “Economics and Reality” (1997) Tony Lawson advocates a perspective on social reality labelled critical realism. Critical realism maintains that strict regularities between observable events are the exception rather than the rule in the social world. This is a negative argument for econometrics which is seen to rely on the identification of such regularities. By contrast, the notion of explanation sustained by critical realism does not depend on an abundance of strict event regularities. In this essay we examine whether econometrics is indeed incompatible with critical realism. This involves asking the following questions: Is Tony Lawson’s characterization of econometrics accurate? In what sense, if any, could econometrics be useful for critical realist researchers? The discussion of econometrics is confined to the so-called LSE approach to econometric modelling of macroeconomic time series. Keywords: Critical realism; Econometric methodology JEL classification: B41, C50 ∗Slightly revised version of an essay submitted to the Faculty of Social Sciences, University of Oslo in August 2002 in fulfilment of the requirements for the doctoral course “Philosophy and Methodology of the Social Sciences”. Comments from Eilev S. Jansen, Bjørn E. Naug and Fredrik Wulfsberg are gratefully acknowledged. The views expressed in this paper are my own and should not be interpreted as reflecting those of Norges Bank. †Contact address: Research Department, Central Bank of Norway, P.O. Box 1179 Sentrum, N-0107 Oslo. E-mail: ida-wolden.bac[email protected] 1 Introduction Contemporary academic economics is not in a healthy state. Over many years now problems have regularly come to light which throw considerable doubt on the capacity of many of its strands to explain, or even address, real world events or to facilitate policy evaluation. Such problems especially beset the rather dominant ‘mainstream’ or ‘orthodox’ project, centering on econometrics and formalistic ‘economic theory’ (Lawson, 1997, p. 3). The quotation is taken from the introduction to “Economics and Reality” by Tony Lawson. According to Lawson the unhealthy state of mainstream economics can be attributed to “ontological neglect”, that is, a neglect of explicit considerations of the nature of reality (Lawson, 1999, p. 274). This neglect is seen to be particularly evident in the literature on economic methodology. Most contributions to this literature, it is argued, have focused on epistemological issues and, moreover, have been influenced by the philosophies of Karl Popper. Lawson calls for a reorientation of economic methodology towards a greater emphasis on ontology. The perspective on social reality which he himself advocates is labelled critical realism and is strongly influenced by the transcendental realism of Roy Bhaskar (see e.g. Bhaskar, 1978). Lawson’s main criticism against econometrics is that it turns upon the identification of strict regularities between observable events. According to critical realism such regularities are the exception rather than the rule in the social world. This is a negative argument for econometrics, but not for critical realist research as the notion of explanation sustained by critical realism does not depend on an abundance of strict event regularities. In this essay we examine whether econometrics is indeed incompatible with critical realism. This involves asking the following questions: Is Tony Lawson’s characterization of econometrics accurate? In what sense, if any, could econometrics be useful for critical realist researchers? The discussion of econometrics will be confined to one particular approach to econometric modelling of macroeconomic time series, the so-called LSE methodology. However, we make no presumption that the LSE methodology offers the most relevant approach to empirical macroeconomics, or that other branches of econometrics, such as microeconometrics and financial econometrics, are less important. The essay is organized as follows: Section 2 gives a brief account of critical realism as presented in Lawson (1994, 1997, 1999). This will be contrasted with the alternative approach of empiricism. Section 3 provides an overview of the main tenets of the LSE approach to econometric modelling. Section 4 then addresses Lawson’s critique of econometrics. Section 5 concludes. 2 An outline of critical realism Critical realism (CR) is an anti-empiricist philosophy of science. Hence, it is useful to contrast CR with empiricism. Following Smith (1998, ch. 3), we take empiricism to cover the falsificationism associated with Karl Popper, as well as the various brands of positivism such as the logical positivism of the Vienna Circle and the standard positivism associated with Carl Hempel. Despite their differences these approaches share a set of foundational assumptions that can be contrasted with those of CR. 2 2.1 Ontology An important characteristic of CR is a strong emphasis on ontology, that is, the study of the nature of reality, the study of what really exists. CR distinguishes between three levels of reality; the empirical, the actual and the deep or non-actual (Lawson, 1994, p. 263). The empirical level refers to our experience, perception and impression of the actual. The actual level consists of events and states of affairs. The non-actual refers to the unobservable structures, mechanisms, powers and tendencies governing actual events. These structures and mechanisms exist independently of our experience of them. Importantly, the three levels of reality are unsynchronized or “out of phase” with one another. A given event may be perceived differently by different people. Moreover, a single event is typically determined by multiple, possibly counteracting, mechanisms. Thus, the governing mechanisms cannot be “read straight off” the event. As an example of the lack of synchrony between events and mechanisms, Lawson uses a leaf falling from a tree. The falling leaf is always subject to gravity, but the law of gravity cannot be inferred directly from observing the leaf as the leaf is also subject to aerodynamic and thermal forces (Lawson, 1997, p. 22). An alternative perspective is that reality consists merely of the objects of experience. This ontological position is referred to as empirical realism. Lawson (1997, p. 19) argues that empirical realism is implicit in the empiricist claim that all knowledge is grounded in experience. Empirical realism denies the existence of an unobservable deep or non-actual level of reality. Some empiricists conflate the empirical and the actual levels of reality and adopt what is referred to as a flat ontology (Smith, 1998, p. 298). By contrast, CR proposes a multileveled or stratified ontology.1 2.2 Empirical regularities and causal explanations A common foundation of empiricist approaches is that an empirical regularity is both necessary and sufficient for establishing a causal law. Empirical regularities, or constant conjunction of events, consist of two or more events occurring together in similar conditions. For an empiricist the existence of such regularities justifies the claim to universal laws of the form “whenever event xthen event y” (Smith, 1998). The logical positivists arrived at scientific laws by inductive reasoning, that is, by moving from the observation of particular instances to the formulation of general laws. Recognizing that scientists often proceed by first positing general laws and theories and then engaging in empirical research, the standard positivist and falsificationist approaches put greater emphasis on deductive reasoning. The hypothetico-deductive (H-D) model of explanation requires that the statement about what we are trying to explain (the explanandum) must be logically deducible from a statement expressing relevant initial conditions and at least one universal law (the explanans). As explaining an event according to the H-D model is the same as identifying universal law(s) under which the event can be subsumed, this model is also called the “covering 1As noted by Smith (1998, p. 299), idealism can also be said to commit to empirical realism. In contrast to most empiricists, however, idealists distinguish clearly between events and perceptions, that is, between the actual and the empirical domains. 3 law” model of explanation.2According to the H-D model there is a logical symmetry between prediction and explanation, a notion which is referred to as the symmetry thesis (Blaug, 1992, p. 5). CR rejects the covering law model of scientific explanation. According to CR science does not depend upon identifying constant conjunctions of events. As formulated by Lawson: The aim is not to cover a phenomenon under a generalization (...) but to identify a factor responsible for it, that helped produce, or at least facilitated, it. The goal is to posit a mechanism (typically at a different level to the phenomenon being explained) which, if it existed and acted in the postulated manner, could account for the phenomenon singled out for explanation (Lawson, 1997, p. 212). Thus, science and explanation should be concerned with uncovering the structures and mechanisms that govern events. In any given situation several mechanisms may be operative and these mechanisms may amplify each other or cancel each other out. If a particular structure is at work, an event may or may not be produced. Hence, an event regularity is neither necessary nor sufficient for a causal statement. Moreover, there is no logical symmetry between explanation and prediction.3 Given that the objective of science is to identify the “real” mechanisms underlying events, how should scientific investigation proceed according to CR? According to Lawson the appropriate “mode of inference” is retroduction (or abduction). Retroduction involves (...) the movement from knowledge of some phenomenon existing at any one level of reality, to a knowledge of mechanisms, at a deeper level or strata of reality, which contributed to the generation of the original phenomenon of interest (Lawson, 1997, p. 26). However, Lawson is vague about the actual process of retroduction: Not much can be said about this process of retroduction independent of context other than it is likely to operate under a logic of analogy or metaphor and to draw heavily on the investigator’s perspective, beliefs and experience (Lawson, 1997, p. 212). In other words, there are no simple recipes for how to come up with hypotheses about causal mechanisms. Nor are there any simple rules for evaluating such hypotheses. Runde (1998) sets out broad conditions for testing causal explanations of events. He explicitly acknowledges, however, that these conditions are insufficient to identify the superior explanation in all situations. 2.3 Naturalism and its limits If science requires the identification of empirical regularities, then a condition for science to be possible is that the requisite regularities are there to be found. CR argues that even in the natural world strict event regularities typically occur only in experimental situations. Lawson (1994, p. 261) refers to systems in which constant event conjunctions occur as closed 2The H-D model of explanation is attributed to Karl Popper, Carl Hempel, and Peter Oppenheim. As noted by Blaug (1992, p. 5), it was one of the model’s critics who originally labelled it the “covering law model”. 3According to Collier (1994, p. 58), evolutionary biology is an example of a scientific discipline with high explanatory power but low ability to predict specific outcomes or events. 4 systems. Experimental activity can thus be viewed as an attempt to close a system. Scientists conducting an experiment try to insulate a particular mechanism by focusing on a limited number of variables and excluding possible disturbing influences. In the social world, critical realists argue, it is not generally possible to create experimental closures. Rather, the social world is fundamentally open. As formulated by Roy Bhaskar: The chief epistemological limit on naturalism is not raised by the necessarily unperceivable character of social scientific inquiry, but rather that they only ever manifest themselves in open systems; that is, in systems where invariant empirical regularities do not obtain. For social systems are not spontaneously, and cannot be experimentally closed (Bhaskar, 1989, p. 45). The openness of the social world is closely related to the possibility of real choice. If human beings have real choice it means they could always have acted differently. A precondition for choice is that the world is open in that all events could have been different (Lawson, 1994, p. 269). Thus, open system analyses acknowledge that individuals are complex and internally structured and so may respond differently in similar situations at different occasions. Accepting that the social world is inherently open and that strict event regularities are unlikely, how might we detect the mechanisms of the social world? Lawson (1997, p. 204) argues that even if we reject the possibility of strict event regularities occurring in the social world, we are not thereby forced to take the view that events occur randomly in a totally unsystematic fashion. Instead, regularities come in varying degrees of strictness. Occasionally certain mechanisms may dominate and produce so-called demi-regularities or stylized facts.4 Such demi-regularities are useful for detecting the structures and mechanisms governing events as they allow at least partial access to the deep level of reality. Notice that, like empiricism, CR commits to naturalism.5Both in the natural and social sciences explanation involves a reference to the mechanisms and structures of the non-actual level of reality, and in both cases the appropriate “mode of inference” is retroduction. However, it is recognized that structures and mechanisms in the social world differ from the structures and mechanisms operating in the natural world. For instance, social structures are likely to be only relatively enduring and to vary more over time and across space than the objects of natural science. 4A demi-regularity, according to Lawson (1997, p. 204), is a “(...) partial event regularity which prima facie indicates the occasional, but less than universal, actualization of a mechanism or tendency, over a definite region of time–space.” 5Smith (1998, p. 347) defines naturalism as: “(...) any approach to knowledge construction which assumes that the methods and assumptions of the natural sciences can and should be applied to the objects of analysis in the social sciences.” Although critical realists and empiricists are both committed to naturalism, they differ in their views on what are the appropriate methods and assumptions to use. 5 2.4 Lawson’s critique of econometrics According to Tony Lawson econometrics involves searching for covering laws or universal event regularities in a probabilistic form. More precisely, he argues, econometricians are implicitly committed to a thesis called regularity stochasticism:6 (...) for any (measurable) economic event y a stable and recoverable relationship between a set of conditions x1, x2,..., xnand the average or expected value of y (conditional upon x1, x2,..., xn), or some such, is postulated (Lawson, 1997, p. 76). It is clear that once we accept the critical realist view of the social world as inherently open, we expect a stable and recoverable relationship between economic events to be the exception rather than the rule. Lawson (1997) identifies three strict conditions that must be satisfied for a stable empirical relationship to obtain, namely the extrinsic closure condition, the intrinsic closure condition and, if the relationship involves aggregate or macro-level variables, an aggregational condition for (aggregate) closure. We consider these in turn. The extrinsic closure condition is satisfied when the phenomena under study can be isolated from disturbing external influences. That is, any variable omitted from the econometric model should be uncorrelated with the variables already included. The individuals of analysis may have more than one internal state (for example, a person could be sad or happy) and may behave differently depending on the internal state. Hence the need for a condition for intrinsic closure. The intrinsic closure condition can be understood as implying that a cause always produces the same effect (Downward et al., 2002, p. 482). It will be satisfied if every possible intrinsic state of the unit of analysis is specified, and only one outcome is possible for each state given the specific conditions.7 There can be no general presumption that the individual units in a population are identical. Instead, populations are typically heterogeneous. This implies that the extrinsic and intrinsic closure conditions are not sufficient when the postulated relationship involves aggregate or macro-level variables. Aggregate closure requires either that restrictions are imposed on the individual units so that stability obtains in the aggregate despite the heterogeneity, or that explicit account of the heterogeneity is taken in the model specification.8 Thus econometrics, Lawson (1997) argues, involves making closure assumptions. He points out that a common response to the breakdown of econometric models is to suggest that closure has not yet been achieved. Faced with instability in estimated relationships, econometricians suggest that a variable has been erroneously omitted from the model specification (extrinsic closure has not been achieved), or else that the model is not formulated at a sufficiently disaggregate level (the intrinsic and/or the aggregational closure conditions are not satisfied). From a critical realist perspective, however, the presumption that closure is achievable is a fallacy, thus attempts to create closures are likely to fail. This follows from a recognition that 6Mainstream theoretical economics, it is argued, is committed to a special case of regularity stochasticism, namely regularity determinism (Lawson, 1997, p. 98). 7To illustrate; in neoclassical consumer theory intrinsic closure is achieved by assuming that individuals are utility maximizers (they have only one intrinsic state), and a unique utility maximum exists (there is only one possible outcome for each internal state). 8Neoclassical consumer theory imposes restrictions on individuals’ utility functions to ensure that the representative consumer model holds, that is, that aggregate behaviour looks as if it was generated by a single “representative” consumer with income equal to aggregate income. A necessary and sufficient condition for the representative consumer model to hold is that individual utility functions have the so-called Gorman form. See e.g. Varian (1992, p. 153–154). 6 4.1 The aims and possibilities of econometrics According to Hendry the objectives of empirical modelling involve “discovering sustainable empirical relationships between observed economic magnitudes”. This apparently fits nicely with Lawson’s characterization of econometrics as engaged in a search for covering laws. I would argue, however, that the sustainable relationships Hendry refer to are less strict and less universal in scope than the “exceptionless” regularities required for a covering law according to Lawson. After reading several early contributions to econometrics, Kevin Hoover concludes that Yes, they [the early econometricians] sought robust regularities, but they expected neither the precision nor the freedom from context and precision that is implied in Lawson’s covering law characterization (Hoover, 1998, p. 7).19 This characterization applies equally well to the LSE econometricians. Although they search for invariant relationships, it is clear that they do not expect to find relationships that are invariant to any kind of shock that may hit the economy. It is recognized that, occasionally, economies are subject to abrupt changes that may well cause statistically well-specified and theory-related econometric models to “break down” in the sense of producing systematic forecast failure. Furthermore, it is recognized that some events could be inherently unpredictable. The LSE approach clearly acknowledges the complexity of the economy and the problems involved in measuring economic magnitudes. This is evidenced in the following quotation: Modern economies are complicated, dynamic, non-linear, simultaneous, highdimensional, and evolving entities; social systems alter over time, laws change; and technological innovations occur. The resulting time-series data samples are heterogeneous, non-stationary, time-dependent and inter-dependent. Economic magnitudes are inaccurately measured, subject to revision and important variables are not observable: worse still, available samples are relatively short and highly aggregated (Hendry & Krolzig, 2001, p. 101). Moreover, the LSE methodology recognizes the difficulties involved in isolating sustainable relationships between economic variables. An empirical model is seen to be derived from the underlying DGP through a series of reductions. Interestingly, the conditions for the reductions to be valid, that is, to involve no loss of information, relate to the closure conditions identified by Lawson. Valid marginalization with respect to variables that are excluded from the analysis corresponds to satisfaction of the extrinsic closure condition. The requirement of parameter constancy relates to the intrinsic closure condition as well as the condition for aggregate closure. The LSE approach emphasizes that the validity of the implied reductions should be evaluated through extensive testing of the empirical model. We can interpret this as saying that closure, to some extent at least, can be justified by statistical criteria. Tests for omitted variables and tests for parameter instability could help establish whether closure, at least locally, has been achieved. Tests of super exogeneity can be used to evaluate whether the Lucas critique applies 19Similarly, on the subject of covering law notions in economic theory, Hausman (1999, p. 287) argues that “The dominant view of economic theory since the early nineteenth century is that it captures only some of the relevant factors and that consequently its empirical implications will be inexact and sometimes badly off the mark. (...) As economic methodologists have long understood, a ceteris paribus law implies no strict regularity among the properties it explicitly mentions”. 13 to the model at hand. Interestingly, there seems to be little evidence for the empirical relevance of the Lucas critique (Ericsson et al., 1998, p. 376). Accepting the view that econometrics is about discovering “robust” regularities, we must make probable that the requisite regularities are indeed to be found. According to Lawson so-called demi-regularities may occasionally turn up even in the social realm, reflecting that certain mechanisms come to dominate over limited regions of time and space. These demiregularities are insufficient to the requirements of econometrics as defined by Lawson, but are perhaps more like the robust regularities required for a less “ambitious” econometric project. Moreover, there are different views on the degree of openness of the social world within the critical realist camp itself. Drawing on insights from institutional and behavioural economics, Downward et al. (2002) argue for a modified realist ontological position. Their argument is that although the social world is potentially, or logically, a purely open system, in practice it is characterized by “quasi-closures” which may manifest themselves in stable patterns of events. Situations of quasi-closure come about because social agents seek “ontological security” in an open and complex social world. Faced with the challenge of making decisions in complex situations, agents are inclined to develop habits, conventions and routines. Such habit and routine behaviour will be reflected in social institutions which may prove to be stable over periods of time. 4.2 Could econometrics be useful for critical realists? According to CR the aim of science is to uncover the mechanisms and structures of the deep level of reality, and these need not manifest themselves in regularities at the empirical level. This implies that, even if we accept that econometrics relies only on “robust” regularities and that such regularities do in fact obtain in the social world, we have yet to demonstrate that econometrics could be useful for critical realist researchers. The relevant question is whether econometrics could help produce explanations in terms of mechanisms. As noted by Downward et al. (2002), the implications of critical realism for empirical research methods (which, of course, are not exhausted by econometrics) are not entirely clear. Lawson dismisses econometrics as incompatible with critical realism, but nevertheless insists that his explanatory project contains a “significant empirical component” (Lawson, 1997, p. 221). Activities such as the computation, graphing and tabulation of summary statistics about the economy are all regarded as informative. As argued by Hoover (1998, p. 16), much of econometrics is just sophisticated versions of these “legitimate” activities. In fact, one of the main purposes of empirical econometric models according to Hendry (1995a) is to provide data summaries. The extent to which econometrics could be useful in critical realist research is related to the role played by demi-regularities in this research. Lawson admits a role for demi-regularities at two stages in a realist research project. First, demi-regularities have a role in the context of discovery. Demi-regularities help direct the research process and contribute to the generation of hypotheses about causal mechanisms. Second, demi-regularities could have a role in the assessment of causal explanations. Econometric models potentially have the capacity to reveal non-spurious and non-obvious (demi-) regularities. In a sense this will be particularly true of an approach to econometrics which takes the relationship between the unknown data generating process and the empirical 14 model seriously, such as the LSE approach. We argue that the capacity to reveal unobvious and robust regularities makes econometric models useful for critical realist researchers. This view gets some support from Downward et al. (2002) who, in line with their view that the social world is characterized by situations of “quasi-closure”, admit a greater role for econometrics in realist research than does Lawson. Thus, we agree with Hoover (1998) that econometrics is compatible with critical realism. Econometric models potentially contribute to causal explanations by summarizing data and by revealing interesting demi-regularities. However, as emphasized by Sayer (2000, p. 22), statistical explanations are not explanations in terms of mechanisms, and so, from a critical realist perspective, the explanatory power of the econometric models themselves could be limited. In order to uncover the sought-after structures and mechanisms of the non-actual level of reality, critical realists would use econometrics as a supplement to other research methods. 5 Concluding remarks In “Economics and Reality” Tony Lawson advocates an open-systems ontology for the social sciences. Econometrics, it is argued, supposes that closures are widespread in the social world. Defining econometrics as a search for exceptionless regularities, econometrics is incompatible with critical realism. Examining the writings of David Hendry and his coauthors, however, it seems that econometrics is better characterized as a search for robust, but not necessarily strict, event regularities. The LSE approach also emphasizes the importance of testing the validity of the reductions or closure assumptions implicit in the empirical model. By revealing interesting and non-obvious partial regularities, econometric models could contribute to causal explanations in the critical realist sense. Hence, econometrics is useful for critical realists, at least as a supplement to other research techniques. What is more, by highlighting the conditions under which stable empirical regularities obtain in the social world, critical realism could contribute to a better understanding of the potential and the limitations of econometrics. 15 References Bhaskar, R. (1978). A Realist Theory of Science. Hassocks: Harvester Press, 2nd edition. Bhaskar, R. (1989). The Possibility of Naturalism: A Philosophical Critique of the Contemporary Human Sciences. New York: Harvester Wheatsheaf, 2nd edition. Blaug, M. (1992). The Methodology of Economics: Or How Economists Explain. New York: Cambridge University Press, 2nd edition. Clements, M. P. & Hendry, D. F. (1998). Forecasting Economic Time Series. Cambridge: Cambridge University Press. Clements, M. P. & Hendry, D. F. (1999). Forecasting Non-stationary Economic Time Series. Cambridge, Massachusetts: The MIT Press. Collier, A. (1994). Critical Realism. An Introduction to Roy Bhaskar’s Philosophy. London: Verso. Downward, P., Finch, J. H., & Ramsay, J. (2002). Critical realism, empirical methods and inference: A critical discussion. Cambridge Journal of Economics, 26, 481–500. Eitrheim, Ø. & Nymoen, R. (1988). LSE-skolen i økonometri [The LSE School in Econometrics]. Arbeidsnotat 1988/11, Norges Bank. Ericsson, N. R., Campos, J., & Tran, H.-A. (1990). Pc-give and david hendry’s econometric methodology. Revista de Econometria, 10, 7–117. Ericsson, N. R., Hendry, D. F., & Mizon, G. E. (1998). Exogeneity, cointegration, and economic policy analysis. Journal of Business and Economic Statistics, 16(4), 370–387. Favero, C. & Hendry, D. F. (1992). Testing the Lucas critique: A review. Econometric Reviews, 11, 265–306. Favero, C. A. (2001). Applied Macroeconometrics. Oxford: Oxford University Press. Frisch, R. (1933). Editor’s note. Econometrica, 1, 1–4. Gilbert, C. L. (1990). Professor Hendry’s econometric methodology. In C. W. J. Granger (Ed.), Modelling Economic Series (pp. 279–303). Oxford: Oxford University Press. Hausman, D. M. (1999). Ontology and methodology in economics. Economics and Philosophy, 15, 283–288. Hendry, D. F. (1983). Econometric modelling: The ’consumption function’ in retrospect. Scottish Journal of Political Economy, 30, 193–220. Hendry, D. F. (1995a). Dynamic Econometrics. Oxford: Oxford University Press. Hendry, D. F. (1995b). Econometrics and business cycle empirics. The Economic Journal, 105, 1622– 1636. Hendry, D. F. & Krolzig, H.-M. (2001). Automatic Econometric Model Selection Using PcGets 1.0. London: Timberlake Consultants Ltd. Hendry, D. F. & Richard, J.-F. (1982). On the formulation of empirical models in dynamic econometrics. Journal of Econometrics, 20, 3–33. Hoover, K. D. (1998). Econometrics and reality. Unpublished manuscript, University of California, Davis. Lawson, T. (1994). A realist theory for economics. In R. E. Backhouse (Ed.), New Perspectives on Economic Methodology (pp. 257–285). London: Routledge. Lawson, T. (1997). Economics and Reality. London: Routledge. Lawson, T. (1999). What has realism got to do with it? Economics and Philosophy, 15, 269–282. 16 Lucas, R. E. (1976). Econometric policy evaluation: A critique. In K. Brunner & A. H. Meltzer (Eds.), The Philips Curve and Labor Markets, Vol 1 of Carnegie-Rochester Conference Series on Public Policy (pp. 19–46). Amsterdam: North-Holland Publishing Company. Pesaran, M. H. & Smith, R. (1995). The role of theory in econometrics. Journal of Econometrics, 67, 61–79. Runde, J. (1998). Assessing causal economic explanations. Oxford Economic Papers, 50, 151–172. Sayer, A. (2000). Realism and Social Science. London: SAGE Publications Ltd. Smith, M. J. (1998). Social Science in Question. London: SAGE Publications Ltd. Varian, H. R. (1992). Microeconomic Analysis. New York: W. W. Norton and Company, Inc., 3rd edition. 17 18 WORKING PAPERS (ANO) FROM NORGES BANK 20022003 Working Papers were previously issued as Arbeidsnotater from Norges Bank, see Norges Bank’s website http://www.norges-bank.no 2002/1 Bache, Ida Wolden Empirical Modelling of Norwegian Import Prices Research Department 2002, 44p 2002/2 Bårdsen, Gunnar og Ragnar Nymoen Rente og inflasjon Forskningsavdelingen 2002, 24s 2002/3 Rakkestad, Ketil Johan Estimering av indikatorer for volatilitet Avdeling for verdipapirer og internasjonal finans 2002, 33s 2002/4 Akram, Qaisar Farooq PPP in the medium run despite oil shocks: The case of Norway Research Department 2002, 34p 2002/5 Bårdsen, Gunnar, Eilev S. Jansen og Ragnar Nymoen Testing the New Keynesian Phillips curve Research Department 2002, 38p 2002/6 Lindquist, Kjersti-Gro The Effect of New Technology in Payment Services on Banks’Intermediation Research Department 2002, 28p 2002/7 Sparrman, Victoria Kan pengepolitikken påvirke koordineringsgraden i lønnsdannelsen? En empirisk analyse. Forskningsavdelingen 2002, 44s 2002/8 Holden, Steinar The costs of price stability - downward nominal wage rigidity in Europe Research Department 2002, 43p 2002/9 Leitemo, Kai and Ingunn Lønning Simple Monetary Policymaking without the Output Gap Research Department 2002, 29p 2002/10 Leitemo, Kai Inflation Targeting Rules: History-Dependent or Forward-Looking? Research Department 2002, 12p 2002/11 Claussen, Carl Andreas Persistent inefficient redistribution International Department 2002, 19p 19 2002/12 Næs, Randi and Johannes A. Skjeltorp Equity Trading by Institutional Investors: Evidence on Order Submission Strategies Research Department 2002, 51p 2002/13 Syrdal, Stig Arild A Study of Implied Risk-Neutral Density Functions in the Norwegian Option Market Securities Markets and International Finance Department 2002, 104p 2002/14 Holden, Steinar and John C. Driscoll A Note on Inflation Persistence Research Department 2002, 12p 2002/15 Driscoll, John C. and Steinar Holden Coordination, Fair Treatment and Inflation Persistence Research Department 2002, 40p 2003/1 Erlandsen, Solveig Age structure effects and consumption in Norway, 1968(3) – 1998(4) Research Department 2003, 27p 2003/2 Bakke, Bjørn og Asbjørn Enge Risiko i det norske betalingssystemet Avdeling for finansiell infrastruktur og betalingssystemer 2003, 15s 2003/3 Matsen, Egil and Ragnar Torvik Optimal Dutch Disease Research Department 2003, 26p 2003/4 Bache, Ida Wolden Critical Realism and Econometrics Research Department 2003, 18p Ida Wolden Bache: Critical Realism and Econometrics Working Paper 2003/4 KEYWORDS: Critical realism Econometric methodology - 14142 14142-working-paper-4-03 16.05.03 12:49 Side 4