Log in

16 July 2014 @ 10:53 pm
Trying to figure out what the SEC is trying to accomplish with a tick-size experiment.  
I'm trying to figure out the motivation behind the SEC's experiment:


(Via a reference in Kid Dynamite's blog post)

In short, they're picking a few thinly-traded stocks, and requiring that prices be a multiple of 5 cents, instead of the now-traditional 1-cent (usually). The first-order effect of this is pretty clear: It transfers money from investors to market-makers. In the past the SEC has worked in the opposite direction, both with decimalization and going after dealers for "colluding to keep bid-ask spreads wise". In particular, NASDAQ had a phase where lots of bid and ask prices were even; the SEC viewed this as an attempt to mistreat investors.

It's possible that this is an attempt to favor traditional market-makers over the high-frequency traders: With prices changing by 5 cents at a time, there's less opportunity for a high-frequency trader to pick up a trade. (Or viewed the other way, for a high-frequency trader to give an investor a better deal by a penny or two.) They may be thinking about the smaller bid and ask sizes that naturally go with a smaller spread, but if that's their objective, it seems like there would be better ways of accomplishing that.

For example, instead of the current Regulation NMS (National Market System -- it details how the many exchanges interact, and those rules do much to enable high-frequency traders), the SEC could experiment with slower trades: Have exchanges match trades every second. As an individual investor, I'm rarely in that much of a hurry, and I'd rather get a deal that's better by a dollar than a deal that's 300 milliseconds faster.
Tags: ,
Steuardsteuard on July 17th, 2014 11:57 am (UTC)
It seems like it's got to be an anti-HFT thing. But I completely agree: moving to "auctions resolved once per second" as a model would seem to be a much more effective tool. (At a minimum, it would eliminate the value of spending loads of cash for co-located data centers.)