Mechanics of performativity, Pt. 1
February 18, 2007
Recently, Daniel Beunza brought to my attention the on-going discussion among organisational theorists about the notion of performativity and the way this concept is applied to study of financial markets (and other organisational environments).
Having taken part in the research project where the cases of the options and futures markets were analysed (one of the products of which is the “performativity” paper), I feel that some of the arguments raised in the discussion around economic performativity need to be addressed, so that the general message of the concept is made clearer and future discussions would are more informed.
One of the arguments I repeatedly encountered is expressed nicely here:
[P]erformativity lacks meaningful boundaries for what can be asserted and why it is taken to be true and real, rather it simply in post hoc fashion labels whatever happens to emerge in reality as: performativity.
In other words, the post is claiming that performativity is very good at predicting the past. Indeed, it is undeniable that it is very difficult, given the presents methodological state-of-the-art, to predict which theory-environment constellations would produce performative phenomena (It is the place to mention that some fascinating methodological developments may bring a change here). Nevertheless, this does not take away from the strength of the performativity concept. Being able to look back on a historical phase and analyse the true paths of effective influence that expert bodies of knowledge, (e.g. economics, accounting) had on organisation evolution is not to be sniffed at.
Another strand of criticism is what I call the ‘stuck clock’ argument. The old saying goes: ‘even a stuck clock tells the right time twice a day’. The same was said about performativity of the Black-Scholes model in financial markets: the model was not accurate to begin with, but as the market matured and the conditions in it resembled the assumptions of the model, the prediction it made were gradually more accurate. Hence, the model did not affect the market, but simply ‘was there’ when the market became more model-like.
This argument calls for an elaborate, empirical answer, and I will try to give one in the following posts. However, at stage it is important to note that there was a relatively short time during which the 1973 version of Black-Scholes predicted options pricing accurately, and that is roughly between 1978 and 1987; before and after this period there were (and still are) significant discrepancies between market prices and the model’s predictions.
If the Black-Scholes model is one of the most successful theories in the history of economics, why is it that it worked well only for nine years? More specifically, maybe this particular historical path of the model (going in and out of accuracy) is trying to tell us something more general about how theories in the social sciences actually operate. That is, that we should take into account not only the internal intellectual rigour of the theory but also its effects on the field it aims to describe.