One of my first posts on SWP (in January 06) discussed exchange capacity limits. A recent WaPo article states that this issue is increasingly a concern of market participants who worry that during times of market stress exchange message processing capacity will be overwhelmed, leading to (or exacerbating) market disruptions.
My post discussed the point that the nub of the issue was that messages (i.e., bids, asks, cancels) were unpriced resources, and that as a result customers send too many of them. This has apparently dawned on some important folks. The article quotes the NYSE’s John Thain as saying: “We don’t get paid for messages, we get paid for trades,” Thain said. “So nobody has infinite message traffic capacity.”
Well, yeah, dude, but that sort of begs the question–is it optimal (or even sane) to have unpriced messages? Similarly, it begs the question–how do you know the right amount of message traffic capacity to invest in if there is no market for it? It is ironic (or something) that businesses that are in the business of facilitating price discovery haven’t thought more seriously about discovering some prices for their most important resource.
As noted in the original post, it is not a trivial matter to devise a capacity pricing system. As Coase noted long ago, there are costs to using prices to allocate resources. But there can be costs of not using them too. And there has to be some means of allocating a scarce resource (capacity) even if one doesn’t use prices. The article suggests that the costs of inefficient utilization of, and investment in, capacity may be very large. It mentions October 87, when capacity limits nearly caused a catastrophe in US markets. The TSE’s recent embarrassment is another illustration. I therefore think it is past time for exchanges to think seriously about capacity allocation mechanisms, including pricing mechanisms.