Simplicity in Econometric Modelling : Some Methodological Considerations

In this paper, two main issues in econometric modelling are being addressed, one is more philosophical, and the other is more practical oriented. First, simplicity as a modelling concept is connected to important strands in scientific philosophy, namely the ?traditional scientific methodology?, the ?rhetorical? approach and Hermeneutics. The ?traditional scientific methodology? emphasises that ?objective? observation in empirical science is possible, while the other two approaches are much more critical in that respect. The ?rhetorical? approach focuses on the idea of writing as a way to persuade the (scientific) reader, while the hermeneutical approach puts more emphasis on ?understanding? (the philosophical concept of ?Verstehen?) in the context of theories and not explanation. It is shown that simplicity has an important role to play in all of these approaches. The second issue relates to simplicity in the context of practical empirical econometrics. Here it is argued that data restrictions make it inevitable that even so-called ?general-to-specific? modelling strategies, as most prominently suggested by David Hendry, are also confined to a simple framework. This does not imply, though, that such approaches should be discarded. As argued above, simplicity does appear to have a number of methodological advantages, and thus this ?limitation? need not be a problem and may even ultimately be regarded as beneficial for empirical research