, , ,

During the VI International Conference: The Austrian School in the 21st Century, Adrian Ravier asked me if my presentation had any conflicts with Garrison’s model of the ABCT (Austrian Business Cycle Theory). My presentation was an overview of the work I’ve been doing with Peter Lewin for the last 2 years or so. In our work, we apply finance to capital theory and business cycles, in particular to the ABCT.

Since I don’t think I answered in the most clear way, and since this might be a topic of general interest, I will try to explain my answer here. Of course, these are my thoughts and I don’t intend to put my words on Lewin.

Of course I see differences between the work I’ve doing in applying finance to the ABCT and Garrison’s model. But I don’t see them as contradicting each other. The reason is that these two approaches have different objectives.

According to Garrison (2001, xii) himself, the model he presents has been developed for pedagogical reasons. In other words, it is a graphic model designed to teach the general insights of the ABCT to undergraduate students next to other business cycle theories. This is a very important contribution. The model was so successful that it started to be used at the graduate level and even to drive empirical academic research. No surprise such application comes with issues (for what is worth, I heard Garrison himself express this same concern at a conference). To extend the model to a more complex scenario becomes too troublesome. Empirical research is either inconclusive or contradictory and only sometimes supports the ABCT thesis.

In our papers we mentioned, more than once, the limits of this model when applied to empirical research. In particular the use of stages of production is troublesome because these are a conceptual tool more than an objective characteristic of reality. Empirical research has been mostly focus on the behavior of stages of production as depicted in Garrison’s model.

Another issue is what in the conference I referred to as too much aggregation. The ABCT literature emphasize that the problem is mal-investment. But the investment variable is an aggregate one (as is consumption). This simplification is done to emphasize the story going on in the Hayekian triangle with the stages of production. And because during a boom the economy goes outside the PPF the model invites the interpretation that there is over-investment instead of mal-investment. Another issue with the PPF is that it shows the familiar curved shape. This shape makes sense when looking at the production of goods and there is decreasing marginal returns. But the PPF in Garrison’s model is between USD consumed and USD invested. One USD not consumed is one USD invested, and so the PPF should be a straight line (this comes from: Y = C + I where C = Y – I which has a slope of minus one).

I don’t think these issues are that important given the objective of the model. The problems I mention are more due to what I see as a misapplication of the model than faults of the model itself.

Our focus, however, is different. When I see the work on finance and business cycle we’ve been doing I don’t see an undergrad pedagogical tool. What I see is an attempt to (1) provide a more solid foundation of capital theory and the ABCT, which in turn is arguably the most challenged issue by non-Austrians. (2) Show how the ABCT story is captured in a framework that reflects how investment decisions are actually made (not present in Garrison’s model itself). (3) Use (1) and (2) to communicate/talk with “non-Austrian” economists, that is, have bridges between these two paradigms as they were present in the first half of the twentieth century. I would say that very little, if anything, we frame in finance would be considered highly contestable by most economists. We boil down the ABCT story to a simple well known effect called duration (both, Macaulay and modified).

In summary, I don’t see a “clash” between our work and Garrison’s model because each approach has a different objective and therefore different assumptions and complexities. I do see problems, however, with how the model has been applied recently, which is different than seeing problems with the model. I myself with Ravier have worked within the framework of Garrison’s model. Also, along with Alexandre Padilla, we developed a mathematical version of Garrison’s model.

I hope this brings some clarity to the work Lewin and I have been doing.