Behavioural economics, antitrust and regulation

Posted on

What follows are some initial thoughts on how behavioural economics might affect antitrust and regulation.  They are not particularly advanced, but some points of the points raised I have not seen elsewhere (even when I remember to search for “behavioral economics” and not just “behavioural economics”).  Wikipedia has a good general introduction to behavioural economics, and the European Commission has also published a short overview for the press.  (Disclosure: the European Commission is my employer, even though I am on secondment to NYU at the moment.  I was not involved in the preparation of that document.  All opinions are my own, etc…)

There is understandable reluctance by economists to transpose conclusions about the behaviour of individuals to the behaviour of firms: arguably much of the analysis carried out by firms before they decide on a strategy is precisely to widen the bounds of their rationality, avoiding the conceptual limits of individuals.

This sidesteps the question, though, of what the implications of behavioural economics should be when an antitrust analysis depends on consumer behaviour. The Microsoft bundling cases in the US and the EU depended in part on how consumers would react to the bundling into Windows of Internet Explorer and Media Player.  Given that behavioural economics suggests that consumers are influenced by “defaults” to a greater extent than traditional (rational market) economics would predict, then the negative effects of the bundling on consumer welfare would be greater than that which traditional economics would suggest.

Other biases also might be relevant: availability bias – where individuals are disproportionately influenced by highly visible or memorable factors – might suggest that incumbency carries greater advantages than would generally be expected.  Consumers might prefer a highly visible dominant company over a less visible new entrant, even if that new entrant offered a better product.  Endowment effects and loss aversion might also suggest benefits to being an incumbent.

Behavioural economics might also be relevant to the design of remedies, where  those remedies rely on consumer behaviour.  As consumer behaviour is affected by how issues are presented (framing), then it is important to look at how new choices are presented to consumers, and not just whether they are.

I also wonder about the implications of this for mobile phone markets: behavioural economics suggests that most mobile phone price plans are sufficiently opaque that it is in reality impossible for an individual to know which plan is best for their usage – which is why Thaler and Sunstein propose a RECAP (Record, Evaluate and Compare Alternative Prices) system to make the necessary information more transparent. They suggest this as a measure to benefit consumers.  But if mobile phone price plans mean that there is little or no price competition (because individuals do not have the information they need to make an informed decision), what are the implications of that for antitrust analysis or telecoms regulation?

Also, notwithstanding the caution of economists in extending this to firm behaviour, there do seem to be some potential areas of interest.  For example, most studies suggest that the majority of M&A activity result in a loss of shareholder value.  Why then do firms persist?  Leaving aside venality (chief executives of larger firms get paid more) or incompetence, what about the over-confidence bias identified in individuals by behavioural economics?  If people tend to over-estimate the likelihood of good things happening, and under-estimate the likelihood of bad things happening, then might that explain why so many firms enter into mergers that lose money?  It might also suggest that when faced with claims of efficiencies, antitrust authorities should be sceptical.

Perhaps the most general lesson of behavioural economics, which is based on actual behaviour rather than hypothesised models, is that antitrust and regulation should be subject to far more empirical analysis than is currently the case.