Reg NMS

There is renewed discussion of the possibility of banning payment for order flow in the United States, and the proposed ban appears to be addressing some issues in the wrong way.  Those issues are oddities of "Reg NMS" --- that's "National Market System" --- that convert a dozen or so stock exchanges into a single, abstract "stock market".  And my proposal, ultimately, is that we ban exchanges from charging commissions to liquidity takers.

I want to note that this is ultimately an accounting requirement --- and Reg NMS is why it matters.  If I place an order to an exchange to buy a stock for $64.03, and the lowest price at which anyone is willing to sell is $64.05, the order sits there until someone comes along to sell at $64.03.  If and when they do, the exchange typically charges each of us a commission; I'm really paying $64.031, and they're clearing $64.028.  If the seller were not allowed to pay a commission, you could report the price as $64.031 instead, and charge me a .3 cent commission instead of .2 cents; the same cash changes hands in the same ways, we're just reporting the trade differently.

The problem is that, for many orders subject to certain concessions to the laws of physics (namely the speed at which information can be transmitted), an order to trade "at market" has to go to the exchange with the "best" price.  Because of this, some exchanges actually have a negative commission for liquidity providers --- they might charge me $64.029, and only pay the seller $64.026.  By telling a seller (or, more to the point, the seller's broker) who will only clear $64.026 by selling on my exchange that they can sell there for "$64.03", they can attract a sell order that could perhaps have cleared more (net of commissions) somewhere else.  By requiring that the commission be paid by the trader whose order is resting on the book, you're simply aligning the reported bids and asks with the prices that would actually be obtained by a trader hitting them.

Most of "payment for order flow" similarly constitutes a sleight-of-hand involving careful but economically meaningless distinctions between "prices" and "commissions".

bounded cognition in evolutionary game theory

Suppose you have a set of agents that, for behavioral and strategic reasons, all "cooperate" with each other, but would recognize if one of the other agents started to "defect"; cooperating would be an evolutionarily stable strategy in this context.  If the society gets larger you might expect there to be a point (Dunbar's number, for example) where the agents can't keep track of all of the other agents anymore; suppose, in fact, that we have 1400 agents, each of which is designed to keep track of 140 agents.  As long as all but a couple of agents continue to "cooperate", you're still fine; if the number creeps above 20 or so, then the information required to keep track of who has been cooperating and who has been defecting gets to 140 bits, and so one might suppose that would overwhelm the agents, and there would be a tipping point around there where cooperation would break down.