Chicago Fed Does What The SEC Has Failed To Do


While the SEC and CFTC tinker around the edges and lack the resolve to really take on the issue of high frequency trading, the Federal Reserve Bank of Chicago has once again come out with some creative and innovative ideas to tackle the problem of HFT.  If you recall, back in October of 2012, Carol Clark of the Chicago Fed published “How to keep markets safe in the era of high-speed trading” .  While that paper specifically dealt with risk issues, the Chicago Fed’s new paper gets to the root of the problem with HFT and addresses the concerns of institutional investors. The new paper, which was authored by John McPartland, is titled “Recommendations for Equitable Allocation of Trades in High Frequency Trading Environments”


The author quite keenly does not try to ban any activities but rather has six suggestions to deter unwanted behavior.  This approach is very smart because it gets around the issue of having to identify and regulate HFT. He states that:


“The thesis of this paper is that, rather than attempting to ban these techniques (which could likely be difficult to enforce in practice), one could change the character and economics of the trading environment so as to disincent these and similar undesirable trading techniques. Rather than propose solutions that preclude specific HFT strategies, we propose to simply change the economics of the trade match and trade allocation processes, to strike a more equitable balance between the high frequency trading community and the investment management community.”


Here is a summary of the authors six suggestions:

– Where appropriate, utilize a new trade allocation formula that is intermediate between the Pro Rata trade allocation formula and the Price/Time or FIFO trade allocation formula.

-Create a new, optional, term limit order type that, as part of the trade allocation process, would reward traders for the time that their orders are committed to be resting in the Order Book.

-Completely dark orders or the hidden portion of resting orders that are not fully displayed (lit) in the Order Book should go to the very end of the queue (within limit price) with respect to trade allocation.

-Prior to trade allocation, multiple small orders from the same legal entity entered contemporaneously for the sole purpose of exploiting the rounding conventions of a trading venue should firstly be aggregated as a single order and should carry the lowest allocation priority time stamp of all of the orders so aggregated.

-Rather than running a continuous trade match, trading venues should divide their trading sessions into periods of one half second. At a completely random time within each half second period, the trade match and trade allocation algorithms should be run once.

-Visibility into the Order Book should be no more granular than aggregate size at a limit price. Market participants should neither be able to view the size of individual orders nor any other identifiers of any orders of others. This more granular information is not information that any market participant needs to make a fully informed economic decision as to the instantaneous value of the instrument being traded.

We are particularly fond of the last suggestion.  We have long argued that the data feeds that exchanges provide to their clients contain too much granular information.  A few years back we caught a few exchanges giving out private information on hidden order flow to their data feed subscribers.  While we have suggested that data feeds should only contain information that you could see on the tape, the author takes this one step further:

“No market participants should be able to see any other identifying data in the Order Book that would reveal the identity or origin of the other market participants that have entered orders. No market participant should be able to see the time stamps of any orders in the Order Book other than their own. No market participants should be able to see the individual lot sizes of orders entered, other than their own. Such granular data is not information that any market participant legitimately needs to make an informed economic trading decision.”

The author further states:

By obtaining this more granular pre-trade information, high frequency traders could (1) more efficiently reverse engineer the trading algorithms of their competitors and (2) more effectively discriminate among the counterparties whose resting orders are resident in the Order Book. It is difficult to see how either of these activities serves the public good.”

We hope the SEC and CFTC read this Chicago Fed paper carefully.  We intend to review some of their other suggestions in further detail in upcoming notes and hope the SEC and CFTC also decide to carefully review the papers suggestions.  While the industry has long thrown up its hands and proclaimed that you can’t put the genie back in the bottle, this paper has shown us that maybe its time to forget the old genie and look for a new bottle.