Without time for translation (maybe interpret too lazy), this time goes in English…
The tension between Austrian economics and empirical tests is well known. While this kind of tension has been a long stand, this strain may have been exaggerated or misinterpreted (maybe by both sides, critics and some defenders); there does not seem to be too much distance between the empirical work of Austrians and contemporary works in epistemology as sometimes seems to be implied.
After the Popper-Kuhn-Lakatos-Feyerabend debate, it is not so clear that empirical tests are the correct approach to assess a scientific theory. Popper argued that there is no data analysis that can be independent from theory, introducing the problem of theory laden. Economic indicators, for instance, were built with a given theoretical framework, namely Keynesian inspired macroeconomics. How much confidence provides an empirical confirmation of a Keynesian model using Keynesian inspired indicators? Or, how much confidence provides an empirical rejection of an Austrian theory using Keynesian inspired indicators? It is a theory what points out what should be considered relevant data in the first place. Furthermore, because the capital theory was set aside, proper economic indicators related to the Mises-Hayek prediction are not easily available. The fact that fundamental aspects of the capital theory, a protagonist of the Mises-Hayek business cycle theory, do not find a place in our models does not make the theory wrong any more than it makes our models incomplete. [See Horwit’z comment on the relevance of capital theory]
Additionally, Kuhn argued that data interpretation, not only data selection, was dependent of the theoretical content in different paradigms, and that same terms can mean different things for different scientists.[1] Empirical tests, according to Kuhn, may work inside a paradigm, but not to resolve a dispute between theories from different paradigms because the test needs to take the paradigm for granted (namely, the paradigm is not testable). Lakatos sustained that scientists embrace a nucleus surrounded by a protection belt of assumptions. But as (Machlup, 1955) foretold, this imposes a challenge to the problem of verification in economics because a theory cannot be tested independently of empirical assumptions, which in turn can have different degrees of generality. How to empirically distinguish if there is a problem in the theory or inside the set of assumptions? Importantly, not all assumptions are observable, as (Mises, 1943) and (Lachmann, 1943) discussed, the business cycle theory assumes that expectations behave in a particular way with respect to changes in interest rates; this expectations, however, cannot be observed.[2] Feyerabend, then, concluded that is not through empirical tests, but persuasion and critical discussion the way scientists convince each other; in a debate between paradigms, each group uses their own paradigm to argue on their defense. It is theory versus theory, rather than theory versus data, what the scientist has to deal with. When one looks at «Austrian empirical work,» one may argue that, after all, Austrians are closer to post-popperian epistemology than what the critics say Austrians should do.
This, of course, does not mean that there is no role for empirical work. It means that such studies, as (Machlup, 1955) argued, should be approached as an exercise of illustration of the theory rather than as a strict empirical test.[3] One may test an assumption given the theory, but one cannot test the theory just by itself. Even under this situation interest work can be done. (Young, 2005, forthcoming), for example, measures predictions from the Mises-Hayek business cycle theory. In both cases he finds a statistical significant result. (Young, 2005) finds that labor moves along stages of production as Hayek argued should be the case.[4] The stages of production, however, is an analytical construct of the theory, it is not observable. That is the reason why Young needs to assume how to separate between different stages and their order in the production process. Namely, the structure defined by the stages of production is a set of assumptions. But these set of assumption allows illustrating Hayek’s pattern predictions of the business cycle. Also, (Young, forthcoming) finds that capital structure for the U.S. between 2002 and 2009 was affected as the Mises-Hayek business cycle predicts. Both of Young’s work, should be mentioned, focuses on specific predictions from the Mises-Hayek theory.[5]
A brief but important clarification may be in order. It is referred as a different paradigm the relationship between ‘Austrian economics’ and the ‘neoclassical synthesis’. The difference between ‘Austrian economics’ and ‘neo-classical economics’ is, for instance, different than the difference between ‘neo-classical economics’ and ‘neo-keyneisian economics.’ The later are differences within a paradigm, but do not necessary represent different paradigms as long as they see through the same terminological glasses. This is not, however, how the problem is usually approached in the literature, for instance in (Blaug, 1975; Bronfenbrenner, 1971) when trying to analyze Keynesianism as a change in paradigm. It is assumed that ‘neo-classical’ and ‘neo-keynesian’ economics represent different paradigms. This is also one of the reasons why under this approach it is hard to find scientific revolutions like the ones in astronomy or physics in the history of economics. But if ‘Austrian economics’ and the ‘neo-classical synthesis’ were to be different paradigms, then this helps to understand, among other things, the challenges in communications between these two groups; problem that is far less present between ‘neo-classical economics’ and ‘neo-keynesian economics.’ Regarding the fact that an argument may be put forward that the ‘neo-classical synthesis’ replaced ‘Austrian economics’ and, as such, this reflects an advance in economic’s science, it should be remembered that a change in paradigm does not necessary mean a step forward in the sense of getting closer to the truth, it may well mean a step backwards as well. See (Kuhn, 1962, p. 170).
[1] See Horwitz, S. G. (2000). Microfoundations and Macroeconomics: An Austrian Perspective (2003 ed.) Routledge; Thomsen, E. F. (1992). Prices & Knowledge. (2002 ed.) Routledge. Also Kuhn, T. S. (1962). The Structure of Scientific Revolutions (1996 ed.) The University of Chicago Press, pp. 101-102.
[2] Other non measurable, or non observable, assumption is the stickiness of prices. See Shah, P. J. (1997). The Theory of Business Fluctuations: New Keynesians, Old Monetarists, and Austrians. Advances in Austrian Economics, 4, 33-62..
[3] (Anderson, 1949; Robbins, 1934; Rothbard, 1963) are three examples of empirical research regarding the Great Depression that could be considered an illustration of the business cycle theory. Econometric studies, of course, are not the only way to perform empirical work.
[4] Should be mentioned that not everyone agrees with Young’s approach. See Murphy, R. P., Barnett II, W., & Block, W. E. (2009). Testing Austrian Business Cycle Theory? A Rejoinder to Andrew Young. Journal of Business and Economic Perspectives, 35(2), 73-86. and Young, A. T. (2010). Illustrating the importance of Austrian business cycle theory: A reply to Murphy, Barnett, and Block; A call for quantitative study. The Review of Austrian Economics, 24(1), 19-28.
[5] Other econometric studies are Bismans, F., & Mougeot, C. (2009). Austrian Business Cycle Theory: Empirical Evidence. The Review of Austrian Economics, 22(3), 241-257.; Carilli, A. M., & Dempster, G. M. (2008). Is the Austrian business cycle theory still relevant? The Review of Austrian Economics, 21(4), 271-281; Keeler, J. P. (2001). Empirical Evidence on the Austrian Business. The Review of Austrian Economics, 14(4), 331-351. Also see the comments in Rizzo, M. J. (1978). Praxeology and Econometrics: A Critique of Positivist Economics. In L. M. Spadaro (Ed.), New Directions in Austrian Economics (pp. 40-56). Sheed Andrews and McMeel.
Punto de Vista Economico 
MUY bueno !!!!!!!!!!!! Intentando imitar a Feyerabend (y saliéndome mal, como Foreman respecto a House :-))) ) escribí algo similar en «Adiós testeo empírico adiós», http://gzanotti.blogspot.com/2010/11/adion-testeo-empirico-adios.html
Me gustaMe gusta
Otra buena literatura sobre el tema es “In Defense of Extreme Apriorism” By Murray N. Rothbard
Me gustaMe gusta
Ah! Yes… I knew the Zanotti in me had some inspiration in the real one. Reads very familiar.
As usual, outstanding post by Gabriel in his ‘Filosofia para Mi’!
Me gustaMe gusta
So this is coming to mind now, but I thought of it when I first read what you were saying about Popper-Kuhn-Lakatos-Feyerabend: you must read what Popper said about historical analysis in The Poverty of Historicism. I understand that there are legitimate arguments that what Popper and Mises were saying weren’t that distant on certain margins, but Popper did *not* believe that there is any «correct» way of interpreting any historical event in the social sciences. There are possible interpretations of history from particular perspectives, but there isn’t anything beyond that. More specifically, if I could write a sociological explanation of The Great Depression, and cite specific sociological variables as determining The Great Depression, for Popper there is no reason to favor the sociological explanation over the explanation of Austrian Economics. The same can be said of favoring the neoclassical explanation over the Austrian explanation. What I’m getting at is that Popper didn’t really believe in any empirical method of determining social causation, regardless of what he said in Conjectures and Refutations, he rejected any interpretation of history as being somehow «scientific.»
Me gustaMe gusta
Bueno, hoy terminé de leer «El Método de la Economía Política» (Zanotti), luego leí «In Defense of Extreme Apriorism» (Rothbard), luego «Mises: ¿Rothbard o Machlup?» (Zanotti) y oh casualidad, apareció este post ayer. Que, obviamente, me llevó a leer «Adiós Testeo Empírico Adiós». ¡¡Me copé mal!!
Ahora bien, lo primero que quiero preguntar es: ¿por qué decís que la discusión de neoclásicos y keynesianos está dentro del mismo paradigma?
Por otro lado: «Feyerabend, then, concluded that is not through empirical tests, but persuasion and critical discussion the way scientists convince each other; in a debate between paradigms, each group uses their own paradigm to argue on their defense. It is theory versus theory, rather than theory versus data»
Ok, de acá podemos concluir, como dice Gabriel en su post que «Todo es teoría» y que sea buena o mala lo podemos ver después del diálogo con otras teorías.
Ahora bien (y por ahí esta es pregunta para Gabriel que sostiene esta postura) ¿Por qué, si todo es teoría, es mejor para el diálogo con otros paradigmas la interpretación Machlup de Mises que la interpretación Rothbard? Pienso, a la larga son dos teorías… ¿qué tiene de más dialoguista una que otra?
Luego, cuando Rothbard dice: «The assumption that firms aim at maximizing their money profits is simply a convenience of analysis; it permits the elaboration of a framework of catallactics (economics of the market) which could not otherwise be developed»
¿No es esto un reconocimiento de que no todo es deductivo y a priori?
Y, por último, ¿el rol que le quedaría a la experimentación sería el de «ilustrar» de manera de poder «conversar mejor» con otras ciencias?
Bueno, miles de dudas. Me gustó mucho el tema, espero que puedan aclararme un poco el panorama.
Abrazo!
Me gustaMe gusta
«….Y, por último, ¿el rol que le quedaría a la experimentación sería el de “ilustrar” de manera de poder “conversar mejor” con otras ciencias?»
YES
Eso contesta por qué la interpretación Machlup es importante.
Me gustaMe gusta
Buen, tarde pro seguro. Muy relacionado con este tema me surgió otra duda leyendo «Praxeology: The Methodology of Austrian Economics» de Rothbard. Allí él dice:
«To the economic historian, economic law is neither confirmed nor tested by historical facts; instead, the law, where relevant, is applied to help explain the facts. The facts thereby llustrate
the workings of the law.»
Ahora bien, si no entendí mal, la conclusión de Rothbard es parecida a la de Feyerabend y a la de Gabriel: Los hechos de la realidad, los datos, no pueden refutar ni verificar una teoría, a la larga es teoría contra teoría. Entonces, ¿por qué la mirada de Rothbard es distinta?
¿Por qué es menos dialoguista si le da el mismo rol que Machlup a la experiencia (el de ilustrar cómo la teoría trabaja)?
Saludos!
Me gustaMe gusta
Dentro de la síntesis neoclasica hay distintas posturas, del mismo modo que lo hay dentro de Austriacos. Por ejemplo anarcocapitalistas y liberales clasicos, o free banking versus encaje 100%, pero hay algo que los hace a todos «Austriacos.» Lo mismo sucede con la economía «neoclasica» y «neokeynesiana.»
Hay ciertos conceptos, que podríamos decir, pertenecen al core, que son común a los distintos puntos de vista dentro de la síntesis neoclásica, por ejemplo racionalidad, qué se entiende por conocimiento/información, etc. Hay un núcleo y terminología en común.
Creo que el artículo de Rothbard es difícil. Por un lado lo critica a Machlup, pero por el otro habla de supuestos agregados a la teoría. Pero eso es lo que Machlup hace, creo, con mayor complejidad.
Quizás este post (y comentario) también te sea de interés: https://puntodevistaeconomico.wordpress.com/2011/02/14/%c2%bfque-tan-apriori-es-la-imposibilidad-del-calculo-economico-en-mises/
Me gustaMe gusta
Recién miré un poco el debate de Ken Rogoff y Paul Krugman y me vino a la mente esto que decís. Claro, es el mismo paradigma porque le da cierto rol al gobierno en «optimizar» el mercado, pero discuten a muerte si el gasto público tienen que ser para cualquier cosa o para proyectos más específicos, etc. ¿Sería este un caso de discusión en el mismo paradigma?
Me gustaMe gusta
Puede depender de cómo uno quiere distinguir entre paradigmas. Personalmente los veo como dentro del mismo. Pero puede ser discutible. Hay algo de arte en identificar un paradigma, lamentablemente no es tan claro y explicito. Por lo tanto siempre queda algo a la interpretación de cada uno.
Me gustaMe gusta
Dear Nicolás, it’s so weird answering in English but I’ll continue your trend : )
Very interesting post! I’d like to say a few things about it.
What’s the use of emprical testing in science? I’d say it’s most important use is for critical control of the theories. Is that possible when we know that observation is theory-ladden? Absolutely! There’s no logical secquence from the idea that observation is theory-ladden to the so-called «fact» that empirical tests can’t be used as controls for theories (I see a non-sequitur fallacy here). Furthermore, science is full of examples where theories -before alternatives theories even existed- started to have very serious empirical challenges. For instance, Newton’s theory was «suffering» from severe questioning before Einstein’s relativity.
This, in a nutshell, is the role that Popper’s gives to empirical testing: another way of exercizing critical control of our theories. Is this perfect? Obviously not -it’s as fallible as we all are. Does it help? Yes, it does. And I believe that the methodological consecquences of adopting fallibilism and falsification are much more fruitfull than another competing ideas (I find extremely complicated, for instance, to say that philosophy, theology, metaphysics and science are all the same -point that I have discussed in a number of occassions with our friend Gabriel).
You said that «After the Popper-Kuhn-Lakatos-Feyerabend debate, it is not so clear that empirical tests are the correct approach to assess a scientific theory». There are indeed things that can be debated. However, if something was pretty clear after the famous debate is that most of Kuhn’s notions were obscure, unclear and its core -the notorious «paradigms»- were a myth impossible to find in any science. By the way, it’s interesting to note that Kuhn himself, while writing as a historian of science, never used his own methodology of scientific paradigms (I think Aaron Saal wrote an excelent piece comparing Kuhn with an squizofrenic).
Finally, I totally agree with you that social sciences have a number of complications for empirical that physical sciences don’t have. Since I’m no positivist, I’m very open to the idea that -maybe- social sciences aren’t empirical sciences and can never be empirical sciences (and that wouldn’t make them «meaningless» or unimportant). Having said this, I also would like to say that it’s hard to argue with the idea that a scientific theory has to state in advance some states of the world which contradict it, some situations that aren’t compatible with the theory (if you don’t do that, in the best case you have a tautology and in the worst you’re not as honest as you should be). Is this scenario of stating in advance impossible cases unreachable for social sciences? I don’t think so…
Best wishes!
Me gustaMe gusta
Hello Federico, always nice to have your comments here.
I agree with most of what you said. Indeed I agree that empirical studies do have a role in science. Even with theory-laden and other problems. I didn’t want to imply that is useless and there is no role at all, but certainly is not that strong test that, at leas in economics, is usually implied to be the case in the practice of what would be Kuhn’s normal economics. I also argued that empirical studies do have a role within a paradigm, but they may have a harder challenge to solve a dispute between paradigms.
I have the impression than when pushed hard, the issue between one point of view and other boils down to a discussion of how plausible we think the approach, assumptions, etcetera is, and in that limit I have doubts of how much empirics can contribute. How to empirically solve the Keynesian and Austrian debate on the causes of the Great Depression? At least in cases like economics, it is quite usual to have theories that do not only assume that certain steps hold (assumed in the sense that they are derived from the theory propositions), but also that some relevant aspects are not observable behave in a particular way or are present. Expectations and stickiness may be an example; one can assume rational expectations, for instance, but one cannot observe them.
Does this sound more or less what you had in mind?
Best!
NC
Me gustaMe gusta
Nicolás! You make very good points and -to be honest- for many of them I have no response. In fact, I’d like to take this oportunity to ask you a question: Do you think that, in principle, assumptions could be tested through observable consecquences logically derived from them? Would this be possible? In physical sciences many metaphysical assumptions that exist there are tested following a similar procedure.
Such a procedure, by the way, would be even compatible with Friedman’s instrumentalism since it fouses mostly on the predictive aspect of a theory.
Cheers!
Me gustaMe gusta
I’m not completely sure. I guess some of them may be tested, but such test may require an «interpretative theory», and this may keep the problem in place.
How do we know how facts were interpreted by the individuals? Surely, in many cases a very plausible assumption can be put forward, but that’s still an assumption. The problem is when this plausibility is not so common outside a given ‘paradigm.’ Rational expectations versus subjective expectations could be an example. For the neoclassical economics, it is almost unquestionable that rational expectations should be assumed, this is not the case for an Austrian economics because he posits that the complex phenomena that the market is cannot be summarized in the result of rational expectations. This assumption may be plausible in some simple scenarios, but probably not, for instance, when dealing with business cycle problems.
I don’t know of an empirical solution for a different position on this. One can always change some of the implicit ceteris paribus variables that are not explicit.
Me gustaMe gusta
My freat friend Federico said: «…Is that possible when we know that observation is theory-ladden? Absolutely! There’s no logical secquence from the idea that observation is theory-ladden to the so-called “fact” that empirical tests can’t be used as controls for theories (I see a non-sequitur fallacy here).». I am not so sure. Again, what´s the meaning of «empirical testing»? Kuhn has already said that each paradigm has its puzzled solving, so, you can practice empirical testing inside each pararigm, assuming the meaning of singular statements from the theory that it´s assumed in the paradigm. Out-side the paradigm, the empical testing has no any meaning. So, «I am sorry Guido but»…………….. :-)))) !!!! All that matters is dialogues between theories. And THIS is a popperian point.
Me gustaMe gusta
Dear Gabriel!!! Thank you for your very interesting reply!
Firstly, Kuhn said so many things : ) And, as you know, he had to weaken the notion of «paradigm» to the extent of depraving it from any significant meaning. So, it’s complicated to talk about the existence of an «inside» and an «outside» of the paradigm. Moreover, if one takes Kuhn’s description of paradigms such as presented on «The structure…» no such thing ever existed in the realm of social sciences! (or any other science).
Anyway, more important than this is the fact that I can’t imagine an empirical science whose theories are compatible with every possible state of the world thinkable. If I’m right, then, (critical) empirical testing is a must for scientific activity.
Best wishes!
Me gustaMe gusta
Ok, so our difference is the interpretation of «paradigm» in Kuhn. That´s a good point, he has never been clear about it. Just a little thing. Take in account that Koyré´s lectures were essential to him, and Koyré had been attending the Gotinga´s circle with Husserl………………..
Me gustaMe gusta
With regard to the neo-classical vs neo-Keynesian frameworks as being the same paradigm, one should have to wonder first what one understands by neo-Keynesian. But if we refer to what really Keynes believed, those paradigms are absolutely different. The neo-classicals consider that markets have an inherent tendency to clear (match supply and demand) only hampered by price stickiness, which can be price restrictions just like government laws or the usual market stickiness for some prices. So for them, the general framework is one of full utilization of resources, and they consider excess capacity and unemployment a particular aspect being brought by this friction of the system. Keynes, on the other hand, considered that markets do not have an inherent tendency to clear, because they always rest in a position of underutilization of resources and unemployment; there he coined the term «underemployment equilibrium». He didn’t mean a position of static equilibrium (i.e., where nothing changes), but a position in which changes doesn’t change the system into one with more capacity utilization. So in his paradigm, the «general theory» (as he called his book) is one that considers this underutilization of resources as the norm, being a full utilization of resources just a particular case of this more general one.
Me gustaMe gusta
«To what *really* Keynes believed, those paradigms are absolutely different.» That may be a reading, but the point is that it may be debatable what really Keynes believed, that interpretative. What was Mises *really* position on fractional reserve banking? It is incredible how the same paragraph gets opposing interpretations.
I see now difference in the hard core between neo-classical and neo-keynesian economics. That was the two examples on the post more than Keynes himself. However, the Austrian core is different.
The difference between neo-classsical and neo-keynesian economics is on whether certain assumptions hold, but there’s no discussion on what those assumptions mean or that they are relevant. Expectations may be an example.
Are there other possible delimitation of a paradigm, maybe, if someone wants to follow a different approach. At least, for the purposes of the comparison in the post, is the common nucleus what I use as proxy to distinguish between paradigms. It is this hard core what gives different meaning for same words (and even try to answer different questions).
Me gustaMe gusta