:The most basic question one can ask of a model is ‘What is the effect on variable y2 of variable y1?’ Causation is ‘implementation neutral’ when all interventions on external variables that lead to a given change in y1 have the same effect on y2, so that the effect of y1 on y2 is defined unambiguously. Familiar ideas of causal analysis do not apply when causation is implementation neutral. For example, a cause variable cannot be linked to an effect variable (...) by both a direct path and a distinct indirect path. Discussion of empirical aspects of implementation neutrality leads to further unexpected results, such as that if one variable causes another the coefficient representing that causal link is always identified. (shrink)

:Statisticians have proposed formal techniques for evaluation of treatments, often in the context of models that do not explicitly specify how treatments are generated. Under such procedures they run the risk of attributing causation in settings where the implementation neutrality condition required for causal interpretation of parameter estimates is not satisfied. When treatment assignments are explicitly modelled, as economists recommend, these issues can be formally analysed, and the existence of implementation neutrality, and therefore quantifiable causation, can be determined. Examples are (...) given. (shrink)

Financial economics, and the calculations of time and uncertainty derived from it, are playing an increasingly important role in non-finance areas, such as monetary and environmental economics. In this 2001 book, Professors Le Roy and Werner supply a rigorous yet accessible graduate-level introduction to this subfield of microeconomic theory and general equilibrium theory. Since students often find the link between financial economics and equilibrium theory hard to grasp, they devote less attention to purely financial topics such as calculation of derivatives, (...) while aiming to make the connection explicit and clear in each stage of the exposition. Emphasis is placed on detailed study of two-date models, because almost all of the key ideas in financial economics can be developed in the two-date setting. In addition to rigorous analysis, substantial sections of discussion and examples are included to make the ideas readily understandable. (shrink)