Dealing in Fractions of Sense

So the SEC wants to take on some of the market distortions caused by decimal (and sub-decimal) pricing.  Well, there are the things that can’t be argued about and the things that can.

Starting with what can’t be argued about: liquidity is not a free good.  In a trading market, it exists because market makers or specialists are willing to offer markets of a certain size and bid-ask spread given the usual price volatility.  Most of the time they  will make money, because there is enough informationless volume trading back and forth, that they can take a few losses when information hits the market, and informed traders temporarily make money against intermediaries until a new equilibrium is reached.

The same is true when things become more uncertain — either bid-ask spreads get wider, or sizes get smaller, or both.  I remember back in 2002 as a corporate bond manager/trader — bonds were trading in “onesies” and “twosies,” though bid-ask yield spreads hadn’t widened much.  (I.e. $1-2 million of face amount would trade at a given bid-ask context… if they got too much interest in one side or the other, the bid-ask spread would move, and fast.

So when the SEC made its move to change the tick size (minimum bid-ask difference) from fractions of dollars — eighths & sixteenths, to pennies, there was a tendency for the amount offered by intermediaries to decline.

Now if you are a small trader, you don’t care that much about this.  You were able to get your work done, and at thinner bid-ask spreads.  Life is good.

But if you are a large manager of assets, it’s not so easy.  You have a big buy or sell that you have to do, and the market gate is thin.  Give away too much of the size you want to do, and the market runs away from you, costing you money.  You might actually like markets with bigger offerings and wider spreads better.

And so you seek other trading venues out away from the NYSE and NASDAQ, and occasionally you find you can get large trades done there when there is another large trader willing to take the opposite side of your trade.

And in the process the other trading venues sometimes create fleeting trading opportunities between them and NYSE or NASDAQ, or within them.  That’s high-frequency trading [HFT].  Now, NYSE and NASDAQ profit from this because they sometimes receive payment for order flow.  They seek orders, and are willing to pay a tiny amount to get them, knowing they can make a profit on the other side of the transaction.

Given all of the above, the SEC is now reviewing a proposal to change back to fractions.  My first reaction is this favors the big over the small.  My second reaction is why not regulate/legislate and create one central order book that all orders go through for each security, and publicly display the bids and asks.

My third reaction is why not end payment for order flow.  High frequency trading would end without easy access to the deepest markets.

My penultimate reaction is, why not restructure markets so they transact once a second, or once a minute. It would not impede markets much at all in my opinion.

But my last reaction is, why not charge a teensy fee for every order placed in any venue, whether it executes or not?  It might be as small as a penny on a thousand dollar order.  Or even a penny on a ten-thousand dollar order.  Just enough that there is a disincentive to place a lot of orders where there is little intention of having them fulfilled except at advantage to the order placer.

These are better ideas than moving back to eighths and sixteenths once again — leave that alone, the existing market structure favors small traders, and that is not all bad.  Many large traders disguise themselves as small traders, and get trades done more cheaply than if they were trading in fractions.

Every trading system has its weaknesses — the challenge is to create the system that is the best for the most.  In that case, a good system will:

  • Have a central order book, or,
  • End payment for order flow, or,
  • Change markets to an auction format, or,
  • Add a fee that eliminates most non-completed high-frequency trades.

Personally, I like simplicity.  One central order book and no alternative venues, but allowing for a wide amount of order types that accommodate large orders and small orders.

So don’t go back to fractions.  They weren’t the solution to the current problem.  Better to restrict the market structure so that placing an order costs something.  Being able to place an order is a good, so there should be some cost, whether it executes or not.

What I propose here is more minimalist than other proposals, and would solve most problems from high frequency trading.  Add a small fee to each order — what could be easier?


  • Anon says:

    I must say I am highly skeptical of the idea that wider spreads would actually decrease the total trading costs of large orders. Why should the total depth more than make up for the spread? What would be the theoretical reason behind this behavior by market makers?

    It is also my understanding that the academic literature says that HFT decreases price impact of large orders. (for example

    I also think all of the solutions you mention have significant draw-backs… Going the opposite way, what do you think of Cowen’s idea of decreasing the minimum spread, forcing high frequency market making bots to compete more on price than on speed?

  • Greg says:

    Another tobin tax proposal? Are you serious or just French?

    Calling it a “fee” doesn’t change anything. It is a stupid idea David.

    We all know the politicians won’t leave any fee/tax small. They will scream about evil speculators and make your dumb idea into a huge fee.

    The airplane ticket fee that was allegedly going to be used to fund air traffic control (and security costs pre 9/11) was diverted into the “general fund” and pissed away on pork barrel nonsense before the ink was dry. Air traffic control runs on CRT tubes and WW2 surplus radar, while Congress has its own private gym at taxpayer expense.

    We also know this stupid tax/fee idea of yours won’t be paid by Wall Street anyway. Like all the other so-called taxes on businesses, it gets passed along to the consumer.

    And the most insulting bit of your proposal David is that it is simply the moronic tobin tax by a different name.

    You usually have some great ideas, even when I disagree with you. But trying to propose a tobin tax by fraudulently labeling it a transaction fee? Come on David. You are better than that

  • ljump12 says:

    I think you make some good points here, but I also think some are misguided. I’ll try to review a few of the points interjecting my thoughts.

    First, you are correct in that a change to bigger tick sizes will undoubtedly hurt the retail investor trying to buy or sell 100 shares. It will help the larger institutions move bigger blocks, and in some sense that may trickle back into the hands of the investing public if their 401k/pension funds are the ones moving the blocks. I’ve often thought about this because people often villanize HFT for “sniffing out” large orders, and trading with them. But in a true market, isn’t this what should happen? If a huge buyer steps in, and demand now severely outpaces supply, should the price not go up? I believe the price should go up, until the market deems that it has gone up too much. At that point sellers will step in and bring the market back to a reasonable level. Just because it’s a pension fund, or 401k does not mean they should have carte blanche to buy or sell as many shares as they want without moving a market.

    Perhaps the problem is that institutions are being judged on how well their getting orders done compared to the inside. Take some small cap stock that used to be quoted a quarter wide, now it’s quoted 2-3 cents wide. Before the trader that wanted to move 5k shares could get it done at the inside, now he’s moving the stock 10-15 cents!! But wait, even though he’s gotten a better price than he would have in a quarter wide scenario, the stock now looks like its moved — and he’s done a piss poor job.

    Your analysis of the payment for order flow is wrong. Exchanges like Nasdaq and NYSE do not pay for order flow, however market makers like Citadel, Knight, and Citigroup do. In general I think payment for order flow isn’t the best practice, as it’s questionable if the payment ever really makes it back into the retail investor’s hands. I think having the market makers compete to give the best execution quality is a better option, as it’s the savings are immediately passed back to the retail investor in their execution price.

    The penultimate reaction is an interesting one. It’s certainly something I’ve though about myself, and I agree that holding an “auction” once a second that matches up buyers and sellers would not impede the markets too much. While this option may eliminate the “speed race”, it would need to be well thought out. It would be huge change to the market, and I’m afraid that it might just introduce problems of its own.

    The final reaction is one that i’ve heard tossed around a lot. In my opinion it’s the worst option. By charging to display quotes, you’re immediately going to widen spreads. If i’m a market maker in an ETF, how am I going to make a tight market? Every time the underlying securities move, I have to cancel and replace my quotes to reflect that underlying change. This gets exponentially worse in ETF’s that currently carry tight liquid spreads, but don’t trade often. It’s a dangerous path to go down in my opinion.

    Hope my .02 on some of these won’t go to complete waste. I think that we have a lot of good in our current markets, especially for the retail investor. A lot of this hoopla over HFT in my opinion is way overblown, and really does not affect the mom and pop investor at all.

1 Trackback