Models and mechanisms

The most important question in economic methodology concerns the relationship between economic models and reality. How can economic models based on unrealistic and even false assumptions provide an explanation of real-world economic phenomena? A good starting point is Cartwright's (2002) idea that models are believed to say something about the world because they are intended to describe causal processes or mechanisms operating in the world. In economics, such mechanisms are typically sequences of actions and changes in the constraints facing agents (changes in prices and incomes). They link an initial stimulus to some outcome. Thus the multiplier links a rise in government expenditure (ΔG) to a rise in income greater than ΔG, the mechanism involving a series of steps in which rises in income cause consumers to spend more and rises in spending generate rises in income. Creating a theoretical model involves finding a set of simplifying assumptions such that it is possible to demonstrate, to generally accepted standards of rigor, that the outcome will follow from the initial conditions. The model shows that the mechanism is coherent – that it could operate, if only under some ideal set of conditions.

Cartwright is correct to observe that deriving theoretical results normally requires ‘a very finely honed model’ (2002, 146).