Machines Will Do Shady Things in the Markets Too (Bloomberg View)

Machines Will Do Shady Things in the Markets Too

https://bloom.bg/2hvQDLw

This post originally appeared in Money Stuff.

A useful way to think about modern electronic market structure is that in the olden days humans traded stocks and options, making markets based on gut instinct, and then those humans were replaced by computers that used algorithms that largely replicated the humans’ gut instincts but more efficiently. But also, in the olden days, those humans did various shady things, and over time the computers have started to replicate the humans’ shady-thing-doing abilities, because, you know, the shady-thing-doing tradition runs deep.

And so the way human markets work is that Pension Fund X will buy a lot of Microsoft Corp. stock through Dealer Y, and Hedge Fund Z will call up Dealer Y and say “hey who is buying all this Microsoft stock,” and the dealer won’t say “oh it’s Pension Fund X” — that would be a hideous violation of client confidentiality — but she might give the hedge fund some “market color.” The exact flavor of market color will depend on her relationships with Pension Fund X and Hedge Fund Z, and on the norms of the particular market they trade in, but she might say something like “we’re seeing long-term real-money flows into Microsoft” or whatever. And Hedge Fund Z will conclude that there’s more Microsoft buying to come, and so will buy Microsoft itself to profit by selling into those additional long-term real-money flows. And Pension Fund X will have to pay a little more to complete its Microsoft buying, and will feel aggrieved that its trading strategy leaked out into the market and that it was “front-run” by Hedge Fund Z.

This story is a little fanciful because no one trades stock over the phone anymore, but you can computerize the whole thing. Here’s a story about the “Intellicator Analytic Tool,” a tool that Nasdaq Inc. has proposed to give option-market color to its customers:

Every minute, it would spit out numbers corresponding to a part of the options market and show whether investors in that market segment were bullish or bearish.

While it wouldn’t reveal the identities of investors, the Intellicator could reveal the “customer type” of buyers or sellers in thinly traded markets. For instance, it could show whether a trade was initiated by a small investor or big money manager, by identifying certain traders as “professional customers” who place larger volumes of orders each day.

This is useful color, for the recipient of the color, which means that it will make the subject of the color feel aggrieved: 

But if the Intellicator reveals a small order was initiated by a big investor, others could infer that the investor is about to buy or sell many options, potentially affecting the price of the underlying stock. An algo trader could quickly buy or sell that stock, resulting in a worse price for the investor on the options.

It all just feels so much more naked when the computers do it. When the humans do it, it is all a vague mosaic of intuition and inference and “color.” When the computers do it it is all there in black and white, specifically quantified, plugged into trading models that can make decisions based directly on these signals. It is also less dependent on relationships: Hedge Fund Z, in my fanciful story, needed to know to call up Dealer Y, and she needed to take its call. But Nasdaq will sell its Intellicator, with no clubbability restriction on who gets what market color. “If this data is made publicly available, customer trades could be adversely impacted if bad actors attempt to utilize this data to manipulate the market,” said Sifma in a comment letter. Electronic trading has in many ways democratized Wall Street, but it can also democratize the shadiness.

If you'd like to get Money Stuff in handy email form, right in your inbox, please subscribe at this link. Thanks! 

This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.

    To contact the author of this story:
    Matt Levine at mlevine51@bloomberg.net

    To contact the editor responsible for this story:
    James Greiff at jgreiff@bloomberg.net

    General

    via Bloomberg View https://bloom.bg/2ma9ax6

    November 12, 2017 at 06:02AM