I am frustrated with my fellow Austrian economic school once again, and this time it is on the subject of aggregation. Lord Keynes has a good post demonstrating the ‘Austrian’ alternatives and showing that in accepting their alternatives, they have to accept, in one sense or another, aggregation and some validity of GDP. The interpretation I get from LK’s post is that of consistency. That is to say, if Austrians are going to claim that there are key flaws in aggregation, then their alternative to, say, GDP should not be that of aggregation. Two Austrian commentors defend the ‘Austrian’ aggregation alternatives, one claiming that while Austrians acknowledge aggregation as flawed, the aggregation measures by the Austrians are still better than that of the other schools, and the other claiming that aggregation is not flawed, they serve good purposes in applied economics.
The former commentor states:
Austrian “modifications” are trying to correct for what they perceive as some of the serious flaws (outside of the aspect of “aggregation”) in GDP estimates. Rothbard’s GO, discussed in Power and Market and America’s Great Depression, tried to correct GDP estimates for including government spending as “productive”. He used this to show that during 1929-1932 the government burden in the economy rose.
Similarly, Skousen and Huerta de Soto have tried to correct GDP for not “double counting”, that is, to count only fixed capital spending in investment but not spending in circulating capital along each stage of the production structure. The point of this revision is not to avoid Austrian criticisms of “aggregation”, but provide (in their mind) a more accurate estimate of GDP that does not overemphasize consumption spending and shows the (true, in their mind) volatility in investment spending/capital goods industries during booms and busts.
I know that people who stress GDP do not say that it is perfect (hell, the income and expenditure sides rarely equal nowadays). By perfect I didn’t mean that. Austrians know that their GDP revisions are still flawed in terms of the overall criticisms of aggregation, but better than regular GDP because they try to correct for some of its other flaws (overstating consumption/understating investment, government spending as productive, etc).
My response is quite simple: You admit that aggregation is flawed to begin with, yet you advocate the use of an ‘Austrian’ aggregate measure as an alternative. Do you not see a problem with that? To further the point, Austrians like to view economics as a deductive thing, so the whole point of aggregation has to be abandoned if they see it as flawed in the first place. That is, a premise cannot be flawed, if it is, then the argument is not valid to begin with. This is very similar to how Austrians view neoclassical economics. While they admit they dislike neoclassical economics, they still use it, so in the end, Austrians are just debating over which school has the best neoclassical theory. Same with aggregation, the debate is no longer about the validity of aggregation, it is about which school has the best aggregation method!
The latter commentor states:
I consider myself “Austrian” and I do not reject aggregates as such. I only reject using aggregates as empirical proof for a theory or for constructing economic theory. I consider aggregates extremely useful in economic analysis, however, GDP in that sense has some certain flaws such as including cost-based government spending and the exclusion of intermediary goods.
Aggregates can be useful in applied economics, i.e., applying deduced theories to reality.
This self described Austrian takes a different approach than the other commentor because he sees nothing wrong with aggregation, he/she just does not like when aggregates are used as proof for economic theory. The ideal use of aggregates are that when used in applied economics. But this is flawed also. For example, if we were to use historical aggregate measures to explain deductive theory then we leave out the reality that markets are instable, the world is that of flux, or the role of expectations. That is to say, in a world of flux (constant change) the ‘data’ is always changing, so it does not follow that historical data is somehow justifiable to explain the future. Aggregates (in the way you advocate using them) are only useful in a state of equilibrium and no further change. Such a world does not exist.