Hugo A. Keuzenkamp

Rejection without falsification: On the history of testing the homogeneity condition in the theory of consumer demand

Abstract The analysis of consumer demand is used as an example to illustrate the interplay between theoretical and empirical research in economics. A few simple assumptions and axioms lead to testable implications such as homogeneity of degree zero in prices and incomes of Marshallian demand functions. What are the implications of the resulting tests? We argue that the tests should not be interpreted as intentional efforts to falsify the micro-economic theory of consumer behaviour. The case study is a history of rejections without falsification. Instead, we interpret this episode as a sequence of sophisticated specification searches.

A Precursor to Muth: Tinbergen's 1932 Model of Rational Expectations

This paper describes a model, published in 1932, in which explicitly rational expectations are used to study a dynamic problem of intertemporal inventory allocation. The model is very similar to Muth (1961), except that the stochastic process in the latter is serially correlated, whereas in Tinbergens model the disturbances are serially uncorrelated. Hence, Tinbergen was able to solve a rational expectations model thirty years before Muths much more well known contribution. Copyright 1991 by Royal Economic Society.

The enigma of simplicity

Many scientists believe that simplicity is a crucial element in their quest for knowledge. The order, which is found in chaos, it is thought, facilitates understanding, prediction and intervention. This seems particularly proper in statistical science. Simplicity seems to be related to inductive power. Some even hold simplicity as a sign of beauty and assign it intrinsic value (see Derkse, 1992). Simplicity has many faces. Still, exactly how simplicity is used in scientific reasoning remains enigmatic. Until recently, simplicity remained part of the realm of metaphysics. Recent arguments in the philosophy of science, statistics and econometrics generated wide agreement that the role of simplicity in scientific reasoning is in need of clarification. A multidisciplinary conference was organized and held at Tilburg University in January 1997 to help clarify the definition and role of simplicity in scientific reasoning. The fruits of the conference are presented in this volume of papers. The idea that simplicity matters in science is as old as science itself. Consider Occam’s razor, ‘entia non sunt multiplicanda praeter necessitatem’: entities are not to be multiplied beyond necessity. A problem with Occam’s razor is that nearly everybody seems to accept it, but few are able to define its exact meaning and to make it operational in a non-arbitrary way. Hence, the theme for this volume might well be summarized by the general question of how to clarify the enigma of simplicity in scientific inference. This leads to several specific questions, namely:

The complexity of simplicity

This paper is concerned with the practical problems associated with understanding and defining the concept of simplicity. Different attempts that have been made to define simplicity and, in particular, definitions based on counting parameters, are discussed and analyzed. The limitations of these attempts, especially as applied to economics, are illustrated by means of several econometric examples, including single-equation models, systems of equations, alternative functional forms, and probability distributions.