A link on Twitter sent me to the first chapter of Mary Morgan’s book, The World in the Model: How Economists Work and Think, titled ‘Modelling as a method of inquiry (pdf).’ In the chapter she sets out a brief history of the use of models in economics and argues that economists use models not to uncover economic truths but as a practical form of reasoning or enquiry.
The early economists used words to write about a few general laws – Malthus being a good example – and only a handful used anything like a model. She cites Quesnay’s Tableau Economique as the main example. By the late 19th and early 20th century the use of models had become more frequent – Edgeworth, Marshall and Fisher are cited. But not until the 1930s did modelling come to be ubiquitously used in economics, according to Morgan.
I think this understates the earlier use of modelling. Cournot and Condorcet came to my mind as I read the chapter. I also happen to be reading Roger Backhouse’s History of Economics, and his chapters on the 18th and 19th centuries cite quite a lot of relatively formal modelling such as the introduction of demand and supply curves. However, the reason for counting differently may be that Morgan equates modelling with mathematical notation, and I disagree with this narrow definition. Maps are models – think the London Tube map – and a historical account of the causes of the 2nd world war is a model too. Modelling is an effort to encapsulate causal relationships as parsimoniously as possible in the face of the complexity and messiness of the world. Economists like to use maths – more than ever, according to this paper – and probably too much so. But the maths could almost always be written out in words. Either “C=a + bY + ε” or or “consumption is proportionately related to income, above a necessary minimum and subject to random shocks”.
However, I do agree with Morgan when she writes: “Models function both as objects to enquire into and as objects to enquire with.” She describes them as a way of making informal inference – or, when applied econometrically, of formal statistical inference. Again, though, though, this dual function seems to me to characterise many disciplines. Nor do I think economists only use models, which would be madness, although some do over-use them and fail to distinguish between the model and the world. As John Kay put it, citing Alfred Korzybski, the map is not the territory. Apart from statistical methods, historical reasoning and experiments also have an important place in economics.
The book looks interesting. I have a teetering in-pile at the moment but might put it on the for-later list.