Lawson On Closed Systems In Economics

For Lawson is the type of explanation common to mainstream economics is what he terms the formalistic-deductive framework. He zeroes in on economics laws which state that ‘whenever event x then event y’, an event regularity, deterministically defined. Such statements can develop into laws, or predictions of outcomes, or allow us to deduce outcomes. So, increasing the minimum wage (x) will cause unemployment in the poorest elements of society (y). Naturally, x and y are not restricted to two events or two variables, and can be any number of x’s causing y’s, what matters here is the structure x then y and the relationship of causality between then, it is irrelevant whether this is an empirical study, or begins with a set of principles simply imagined. This methodology, translated into the language of mathematics forms the basis of the mathematical modelling which is the whole mainstream project – other approaches are prima facie dismissed.

Such a method, like all methods, has certain ontological commitments, commitments regarding the nature, structure and raw stuff of reality. Yet, ‘the ontological presuppositions of the methods of mathematical modelling used by economists are rarely questioned or even acknowledged’. It is never questioned whether these ontological presuppositions fit with the social reality under investigation. It is Lawson’s claim that these presuppositions “appear not to arise very often in the social realm” and hence are wholly inappropriate for understanding economic reality. This deductive model presupposes that “event regularities prevail in the social realm” – they do not. Lawson is fond of quoting Marx: ‘In the analysis of economic forms…neither microscopes nor chemical reagents are of use’ – “the nature of the subject matter in question is such that the noted tools are not appropriate to its investigation”. In order to maintain this form of deductive x then y reasoning that is necessary to create mathematical models, economists are required to make two ontological commitments, which create the closed system necessary for prediction, two commitments which the post-autistics and other have noted are wholly inappropriate given even the obvious features of social reality. Lawson names these intrinsic and extrinsic closure. Here we find the root of the two axioms of economics many have found so problematic: atomistic conceptions of agents and bodies (intrinsic closure) and almost complete isolation and lack of intersubjectivity of these agents (extrinsic closure).

Intrinsic Closure

First, intrinsic closure. Say we have the situation, a standard deductive-formalistic one, where stimulus q is applied to x in order for y to happen. I am an economist (perhaps one working in econometrics and this is largely statistical), and I do an experiment in the field, giving a bar of chocolate to one of my friends, Rob. Rob loves the bar of chocolate. If I model this I can assume that giving x (a human person), q (a bar of chocolate) results in y – if x and q then y. You say I gave David (x) a gift of chocolate (q), using my model I would assume he enjoyed it (y) – again if x and q then y. What must we assume for our deductive model to be correct? Firstly we must assume that every x is the same, I must assume that all x’s, that is human beings, all are constant, that they are always the same – I must assume intrinsic consistency, that one can be substituted for the another. Yet in social reality this is not at all the case. David is allergic to chocolate and is hospitalised for six months.

Secondly I must also assume that all x’s respond to the stimulus q in the same manner, that is, y – in Lawson’s words, that only one outcome is possible, only one exit – the situation must be reducible. Yet in the situation I am this is not the case. I gave Rob another bar of chocolate (q) usually loves chocolate so would normally react with pleasure (y) but he is feeling ill today, so he doesn’t want some chocolate, human beings are not intrinsically reducible, they do not respond in the same way to the same stimulus across time or space – “the manner in which a given person may answer a specific question, conduct himself or herself in the market place, or perform some other specified act, is likely to depend upon whether he or she is alert or sleepy, optimistic or pessimistic, starving or well fed…”. Yet, to create a closed system which allows one to predict matters, or create laws, economics must assume the system is intrinsically closed. They must therefore assume the agents in the system are like atoms that can be substituted for one another and that respond in the same way given a stimulus, that the agents involved (be they chocolate bars – I could buy a different brand, of course – or people) are intrinsically consistent. This also allows them to aggregate the results, I send out a hundred chocolate bars and I get a hundred happy people would be my prediction – all the effects can easily be added together because I am assuming every thing is the same. Yet in the real world some people would not like chocolate, or be ill that day – they would not be the same and not respond passively like atoms.

Extrinsic Closure

Perhaps more significant is extrinsic closure. Economic activities must be therefore considered in isolation from any external influences, where the economic transfers and the actors involved is considered in something akin to a vacuum – methodological individualism, non-embedded in the real world. Given our exciting chocolate example, ( if x and q then y) I am unaware that Rob has been told by Andrew that he is fat and that he shouldn’t eat a bar of chocolate. This throws a spanner in the works, instead of if x and q (and I am able to predict y), I have x and q and a (Andrew). To keep my economic law working I need to assume that x and q will never encounter some unknown factor coming in madly from outside. The system must therefore be extrinsically closed against such things for it to stand.

In short, in order for my chocolate model to work I have had to make non-trivial socio-ontological axiomatic claims. It would have made no difference if I hadn’t done the experiment and simply (as is very common in economics) imagined my model from the get go. It is a feature of the deductive style that is incompatible with reality:

it is these tacit requirements of the deductivist mode of explanation, rather than
substantive [reality observing] considerations, which lead to economic models
predicated in terms of atomistic individuals acting in relative isolation

Without using my chocolate technique, Lawson concludes this is the problem with econometric and theoretical economics. In my chocolate model, the event regularity I modelled I took a regularity from the real world, made a model and then used it to predict the outcome – which I got entirely incorrect – because my predictive model assumed closure, because is a necessary feature of deductivism when applied to social
situations. Fullbrook again: “Deductivism ‘presupposes a ubiquity of social closures, while in the event, social closures of an interesting kind appear to be extremely rare’”. Whatever model I might define from the data dissipates on contact with reality, because reality is not a closed system: “Identifying stable econometric relationships requires that the system under investigation be closed off from extraneous effects (the extrinsic closure condition) and from changes in the internal states of those under analysis (the intrinsic closure condition)”. I must therefore assume that individuals can be cordoned off in such a way that behaviours are not interdependent and that they never change, that they are atoms. Yet in reality neither things are possible.

Contrary to popular belief it is then much mainstream economics itself that is assuming away the laws of (sociological) gravity.