95

Economists disagree. They do not agree on whether they disagree, or on why. Some think that they disagree mainly over values, but some studies show that they also disagree over facts. Economists seem to be excessively confident in their estimation of parameters, a classic expert bias. Using taxation as an example, this article points out that the classic distinction between facts and values is too simplistic. Concerning distributive issues, there is obviously no such thing as value-free measurement: What is measured is often the difference from a norm, neither entirely normative, nor merely descriptive. Moreover, parameters have meaning only within a given model. Disagreement may be over the model itself, or the choice of appropriate simplifying assumptions. Choosing a model also mixes normative and positive issues.