My instant reaction is that applied economists regard data as observable when it can be measured or collected in a relatively theory-independent manner. For example, the CPI consists of recording actual prices and calculating an index number. While there are theoretical issues involved of a statistical kind in the creation of the index, a specific economic theory is not involved. That is not to say that economic theory is not involved in the interpretation: for example, in the question of whether or not the CPI is a good measure of "true" inflation. It is just to say that people with opposing theoretical views can still look at the data as a common object on which they can theorize. Constrast that with NAIRU, which can only be recorded when one is willing to make a commitment to a particular economic theory -- and different economic theories give clearly different measures of NAIRU. It is interesting that economists sometimes forget the provenance of their data. The U.S. Congressional Budget Office publishes NAIRU data based on a particular theory, which is used by some economists as an input to a analysis involving an inconsistent theory. Observability on this standard surely admits of degrees, depending on the level of theoretical commitment or, perhaps, forms families of relatively shared theoretical commitments. Kevin Hoover At 04:21 PM 11/15/2004 +0000, you wrote: >----------------- HES POSTING ----------------- >Dear list members, > >I would be very interested in your opinion in the following matter. One >frequently reads in econometric papers that certain economic variables >(e.g. inflation, GDP, unemployment) are termed "observable" while others >(e.g. the NAIRU, potential output, the causal effect of some x on some y) >are called "unobservable". I have two questions about this: > >1) What do you think is the explanation for this difference/distinction? > >Please be as specific as possible (e.g. if you think it is a matter of >direct vs indirect measurability, say exactly what that means; similarly >if you think it's a matter of measurable vs estimable etc. etc.). > >2) What are the loci classici for discussions of the concept of observable >IN ECONOMICS? > >Again, please be as specific as possible: I am NOT looking at testability, >verfifiability, falsifiability etc. but observability. One place is >certainly Oskar Morgenstern's On the Accuracy of Economic Observations. >Uskali Maki sometimes talks about observables but I think he uses a >different concept (according to which variables such as inflation would >UNobservable). > >Many thanks! > >Julian Reiss > > > >----------------- FOOTER TO HES POSTING ----------------- >[log in to unmask] >http://eh.net/mailman/listinfo/hes ****************************************************** "It has been my experience that folks who have no vices have very few virtues." Abraham Lincoln ****************************************************** ****************************************************** Kevin D. Hoover Professor of Economics and Chair Department of Economics University of California Davis, California 95616-8578 Tel. (530) 752-2129 Fax (530) 752-9382 E-mail: [log in to unmask] Web page: http://www.econ.ucdavis.edu/faculty/kdhoover/index.html ******************************************************