Tagged: HFT

Constant Change

The SEC wants to save the little guy. Again. 

A number of you alert readers sent me this story (WSJ registration required but there are similar versions) about the Gensler SEC weighing changes to how trades from retail brokerages are handled.

For those new to market structure, the SEC has a long history of adding complexity to the stock market in the name of helping retail money that ends up instead aiding computerized traders and stock exchanges.

Would that we had less noise and more substance! But we need to first understand the problem. It’s that the stock market is full of tiny trades, and not that retail traders are getting hurt.

Meanwhile, this photo is not of market structure.  Normally in early June – for the bulk of my adult life – I’d be at the NIRI Annual Conference circulating with colleagues.

Photo by Tim and Karen Quast, Jackson Hole, June 2022.

I’d have to count to know for sure, but for most of the past 27 years I’ve been there.  We took a break and this year we’re in Jackson Hole, WY, and other spots in a big loop through WY and MT seeing the west (I’m writing Tuesday night in Billings).

We were in Yellowstone much of Monday covering more than ten miles afoot, riveted by the constancy of change in nature.  Some of it is predictable, like Old Faithful and the Grand Geyser (footage here).

A lot of it isn’t.  And you can’t manage it or direct it.

The point?

Stuff constantly changes.  It’s the most inerrant feature of nature. Change is integral to human nature, which animates the stock market.

Regulators are possessed of that same nature yet want to cast the market like pewter.  Create a model and force every free-moving thing to conform.

It’s most certainly not that SEC chair Gary Gensler is smarter than millions of self-interested participants.  No, regulators want to make a mark, the same as anybody else. 

Pharaohs in Egypt hoped for immortality through pyramids.  Carnegie built libraries.  You can go to Newport and see the edifices of the rich, built to last long beyond the builders.

As ever, I have a point about the stock market. The preceding SEC administration under Jay Clayton revamped the rules, too.

I discussed their final proposed rule, Regulation National Market System II – which I called Reg Nemesis II (see what I wrote about Reg Nemesis I here) – with SEC head of Trading and Markets Brett Redfearn, who described it to the NIRI board at our request.

That rule considered many of the same things Gensler is weighing including redefining the meaning of “round lot,” currently 100 shares regardless of price, to reduce market-fragmentation that harms investors of all kinds.

After all that work, the expended taxpayer resources, the studies and lawsuits and machination, it’s set aside because the new SEC wants to build its own pyramids.   

Okay, Quast, you’ve convinced me everybody wants an ovation. Your point?

The problem is the stock market is stuffed full of tiny trades that devolve purpose from investment to chasing pennies and generating data to sell.

Which in turn is the consequence of rules. I’ve explained before that the SEC’s paramount objective is 100 shares of everything to buy and sell, all the time.

Which is impossible.

I know. I trade.  Trades fragment more at exchanges than in broker-operated markets called dark pools.  Routinely I buy in a chunk in a dark pool and then watch my trades get splintered into 5, 7, 34, 61, shares at exchanges (especially the NYSE).

The exchanges pay traders to set the bid and offer, which snap at my order like piranhas, chopping it into pieces and pricing the market with it.

In Yellowstone, nature takes its course. It’s a marvel, cinematic artistry that takes one’s breath away.  It cannot be and does not need to be directed by humans. We observe it and love it, and it changes.  Some stuff like the Grand Geyser and Old Faithful, follow a clock.  A lot of things don’t.

The same is true of human commerce. The more the few machinate interaction into exceptions and directives and objectives, the less it works.  It should in large part follow its own course, with a clear boardwalk for traipsing through the geysers.

Put another way, rather than merchandising retail trades to build pyramids, we should insist on a single set of standards, no exceptions.  And let the game be played.

There are too many complex rules, too many exceptions.  That’s the real problem.  And so the market lacks the elegance of chance, the beauty of organic and constant change.

The Big Story

Here we go again. 

Twitter fans of Fast Trading are claiming these firms help markets and especially the little guy. Now, before you check out, what’s the Big Market Story of 2021?

Retail trading. Right?  Meme stocks. The rise of the Reddit Mob.

Illustration 209856483 © Hafakot | Dreamstime.com

Editorial Note: And the rise of the #EDGEMob, the success of our trading decision-support platform, Market Structure EDGE, winner of the 2021 Benzinga Global Fintech Award for Best Day Trading Software.  EDGE helps retail traders win by seeing Supply and Demand, the very thing Fast Traders obfuscate daily.

Okay, back to our story.

Fast Trading is computerized speculation.  What most call “market-making.” In the sense that these firms buy stuff wholesale – orders from retail traders – and sell it retail (in the stock market and often back to retail traders via dark pools in bits), that’s true.

But it’s not market-making like Goldman Sachs providing research coverage on hundreds of stocks and committing to buy and sell them.

Fast Traders don’t have customers, don’t write research, don’t most times even commit to both buying and selling. They aim to own nothing at day’s end.  They profit on how prices change. Ironically, they create volatility to vacuum it away through tiny spreads.

Money for nothing.

So, these people on Twitter were saying Citadel and Jane Street and Two Sigma and G1X (unit of Susquehanna that buys retail flow) have better execution-quality than stock-exchange IEX.

That’s like saying sprinters have faster 100-meter times than marathoners.  Well, no kidding. They’re doing different things.  IEX is trying to increase trade-size so we can buy something meaningful. Fast Traders are after the opposite. Tiny spreads, tiny trades.

If you’re getting a headache, let me bring it all around.

“Execution Quality” is part of Reg NMS, the regulatory structure of the stock market. It’s benchmarked by thin gaps between prices, in effect. That is, a spread of a penny is no good. A spread of a tenth of a penny, awesome.

Yet Reg NMS prohibits quoting prices in increments of less than a penny, so there’s an element of irony here.  Quote sub-penny? Illegal. Trade sub-penny? The goal!

At any rate, the SEC determined that it could validate how great the market it had created worked by metering whether brokers executed trades near the best overall prices.

Well, that seems good.

Except the best price is determined by the brokers who are being measured on delivering best prices. That’s like saying, “Whoever you see in the mirror gets to judge you.”

I’m talkin’ ‘bout the man in the mirror. He’s gonna have to change his—sorry, digression. Thank you, Michael Jackson, for that awesome song.

Why use data from firms with no customers – Fast Traders serve none in the sense that Goldman Sachs or IEX do – to determine if there is market quality?

And would someone explain who benefits from a narrow spread?  Anyone? Anyone?

The parties buying and selling stuff but not wanting to own it.  That’s who.

The stock market is supposed to help investors, who want to own stuff.  Yet the rules give kudos to trading firms exploiting retail money, clouding supply and demand, and owning nothing.

That’s how retail money chased herds of buffalo off cliffs in AMC, GME and others. Market regulation crowns highwaymen champions for “narrowing the spread” while meanwhile no one knows what the hell is going on.

Well, we do.  We’re not confused at all.

But what reached a climax in 2021 besides retail trading was confusion. Public companies, do you know why your stock went up or down? Investors and traders, did the market make sense to you?

Stocks fell. Pandemic fears. Stocks zoomed. Fears were easing. Wash, rinse, repeat.

We found it entertaining, as we watched Supply and Demand and saw the market move largely in synchrony with that beat.

Low spreads don’t help public companies or traders. They help regulators justify market structure, market operators make money selling data, Fast Traders make money buying low and selling high in tenths of pennies.

All while saying the market exists for investors and public companies.

Good one, that.

Will it change in 2022?  No.

So, will you?

Traders, if you don’t know Supply and Demand, you’re kidding yourself.

And public companies, I’m not sure what else to say that I haven’t said in 17 years. We can be the people who answer an ever quieter phone, the setters of dwindling meetings as money goes quantitative, data goes quantitative.

Or we can understand Supply and Demand in the stock market.  Ask, and we’ll show you.

Two choices in 2022. Happy New Year! 

Big Strategy

Let’s have a show of hands. 

How many of you think investors woke up, several pounds heavier, the day after Thanksgiving, and opened a browser up to news out of South Africa, and said, “Shazam! Omicron!” And dumped their equities?

Second question, how many of you say that on Monday, Dec 6, investors said, “Screw it, this omicron thing is crap. Buy!” And stocks soared?

If we had a poll on our polls, I’d bet not 30% would have raised hands on either question.

So, why did the headlines say that?  And a step further, if we don’t believe humans knee-jerked the market around the past week, why suppose humans are doing it other times?

Quast, where are you going with this? What do you want us to say? 

I’d like us to come to terms as investors and public companies with the presence of automated trading strategies capable of acting

Illustration 22077880 © Skypixel | Dreamstime.com

independently.  Not as a side show, a reaction.  As valid as Ben Graham’s Intelligent Investor. Ron Baron picking stocks.

Blackrock runs over a thousand funds, the bulk of which follow mathematical models having little directly to do with earnings multiples.  Blackrock, Vanguard, State Street and Fidelity run $20 trillion of assets, most of it passive.

Yet many believe investment models follow the market, and the market is priced by rational thought. 

Why would one think stocks are priced by rational thought?  Give me data to support that view. They trade more? They own more?

Neither of those is true. My long-only investors twenty years ago were generally buy-and-hold.  Are your top 20 core Active holders in and out all the time?  Course not.

The stock market today is 100% electronic, close to 95% algorithmic, and nearly all prices are products of software. So it’s the opposite then. Buy-and-hold investors are accepting prices set by others.

A week ago Olin Corp. (OLN), the world’s largest chlorine company and owner of the Winchester arms brand, was trading near $65. Last Friday it touched $51, and now it’s back to $58. It dropped 22%, a spread of 28% from best to worst.

Anything to do with the fundamentals of Olin’s business? Active money never changed its mind, valuing OLN about $61 since early November (that’s as measurable as any other behavioral factor behind price and volume, by the way. We call it Rational Price.).

It’s volatility, Tim. Noise. 

If we’re willing to characterize a 20% change in price over a week as noise, we’re saying the stock market is a steaming pile of pooh.  A real market wouldn’t do that.

But what if it’s not pooh?  Suppose it’s a strategy that performs best when demand and supply alike both fall?

Then that strategy deserves the same level of treatment in what drives shareholder value as company fundamentals. 

Do you see where I’m going?  The hubris of business news is its fruitless pursuit of human reason as the explanation for everything happening in the stock market.  And it’s the hubris of investor relations too.

Do you know Exchange Traded Funds have created and redeemed nearly $6 trillion of shares in 2021, in US equities alone (data from the Investment Company Institute)?  Nothing to do with corporate fundamentals. All about supply and demand for equities.

Bank of America said last month flows to equities globally have topped $1.1 trillion, crushing all previous records by more than 200%. Most of that money is going to model-driven funds (and 60% to US equities).

Intermediating equity flows all the time, everywhere, are high-speed trading firms like Citadel Securities, Virtu, Hudson River Trading, Two Sigma, Infinium, Optiv, GTS, Quantlab, Tower Research, Jane Street, DE Shaw, DRW Trading and a handful of others.

They follow real methods, with actual tactics and strategies.  ModernIR models show these trading schemes were 54% of trading volume the past week in S&P 500 stocks. Derivatives, a key market for Fast Traders, traced to 18% of equity volume.  About 19% was Passive models like Blackrock’s ETFs.

That leaves about 9% from Active money, your core long-only investors. 

So, what drove the stock market up and down? On a probability basis alone, it’s the 54%. 

It’s not the same everywhere, but the principle applies. For NVS Nov 18-Dec 6, 38% was Passive, just 31% Fast Trading – those machines.  For TSLA, 57% was Fast Trading, 17% from Passives.

For the record, OLN was 54% Fast Trading, 19% Passive, in step with the S&P 500.

Moral of the story?  No view of the market should ever exclude the 54%. Nor should it be seen as noise. It’s a strategy. The difference is it’s driven by Price as an end, not financial returns as an end. (If you want to know your company’s behavioral mix, ask us.)

And it’s the most successful investment strategy in the market.  That should concern you. But that’s a whole other story about the way stocks trade.

The Flaw

It was Monday, Dec 24, 2018. 

The Friday before, the S&P 500 had dropped 3%, bringing monthly losses to 13% and putting the market on bear turf.  Everybody thought, “Thank God for Christmas, we can catch a breath.”

Instead, stocks caromed down another 3% in the half-day of trading Christmas Eve.  After unwrapping presents, stocks opened the 26th with most on vacation and shot up 5%.

Remember that? 

Suppose you close your offices for the holidays and come back to find the place trashed.  It’s not supposed to happen when nobody is in.  Right?  You wonder who the hell did it.

We first got a sense of the stock market’s capacity for freakish outcomes May 6, 2010. Stocks collapsed in this “Flash Crash,” some to a penny, before resurrecting like a geyser.  We learned about “high-frequency trading,” computers buzz-sawing data, spraying prices, shredding stocks.

Illustration 180830251 / High Frequency Trading © Nadiia Prokhorova | Dreamstime.com

It’s happened a growing number of times since.  Computers thrashed through the market in March 2020 like Godzilla vs Kong, demolishing prices, splashing volume everywhere.

And it just happened again.

Everybody cleared their desks and went off to get a tryptophan high, and while we were cleaning up the dishes and looking at the college sports lineup the day after Thanksgiving, a half-day and often the lightest trading of the year, a Transformer went berserk and blew up on Wall Street, blasting almost a thousand points off the Dow.

This ought to disturb you, public companies and investors.  It’s your equity.  You leave. And some computer freakshow guns up in a hot rod and hits you with a smash-and-grab.

The earth is not flat.  And the stock market doesn’t work the way the CEO thinks.

We were in Fiji sailing before the Pandemic.  There was a low sand bar connecting what would otherwise be two islands. We anchored and walked it, observing large numbers of hermit crabs hurrying across, lengthwise.

We picked some up and put them down by the water. But no, that was distressing to them, and they flailed frantically to catch up to the rest, dragging their little hermit shells.

Later, we got it.  The tide was coming in. If you didn’t make it to the Hermit rendezvous on the right side behind some big rocks, you were swept away.

They were on a clock. They’re programmed with sea market structure. They grasped the divide between Controllables and Non-Controllables and the need to respond to both.

If hermit crabs can, we can. Traders, know when you’re vulnerable. That’s around options-expirations, holidays, month-ends, earnings.  And when Demand is falling, as it’s been doing since Nov 17, when options-expirations started (Broad Sentiment at EDGE shows you).

New month, new money, so it might reverse on a dime. Knowing, making informed choices, is the key.

Public companies, no market constituency has a greater need than you to know how the stock market works. In a sense, you’re the elevators everybody boards for the ride.

Lead riders to think that Story is safety, and you’re not only hurting them but leaving your executives and boards without defense.  The stock market is motivated primarily by things other than Story. Data is always, always, a safeguard.

What about omicron? It’s an anagram for moronic. It’s 15th among 24 Greek letters. The WHO has been using them to track variants. We heard a lot about Delta. Nothing about variants Epsilon through Lambda, which exist but haven’t mattered.

The WHO skipped Nu and Xi, too culturally insensitive. Next up was omicron.

Easy scapegoat. Investors fear omicron!  Except investors didn’t do Friday, didn’t do yesterday. Friday, the only behavior up was Fast Trading. Those bloody machines. Futures expired yesterday. Volatility destroys them. That’s market structure, not a virus.

Our trouble is a stock market where machines sifting data like scientists can exploit supply/demand imbalances to profit on the pursuit of price. 

It’s the prevalent behavior in the market. It manifests starkly whenever there are unusual changes to investment, because that’s when PRICE can be exploited.  Machines don’t “think” but they’re programmed to respond.

Suppose I said, “Here are the keys to this roadster.  It’s powerful, fast, and fun. Enjoy it.”

I hand you the keys and start to walk away.  Then I stop and say, “Oh, every now and then it veers off the road for no reason.”

What would a rational person do?  Ponder that.

On the Skids

If electoral processes lack the drama to satisfy you, check the stock market.

Intraday volatility has been averaging 4%. The pandemic has so desensitized us to gyrations that what once was appalling (volatility over 2%) is now a Sunday T-shirt.

Who cares?

Public companies, your market-cap can change 4% any given day. And a lot more, as we saw this week.  And traders, how or when you buy or sell can be the difference between gains and losses.

So why are prices unstable?

For one, trade-size is tiny.  In 1995, data show orders averaged 1,600 shares. Today it’s 130 shares, a 92% drop.

The exchanges shout, “There’s more to market quality!”

Shoulder past that obfuscating rhetoric. Tiny trades foster volatility because the price changes more often.

You follow?  If the price was $50 per share for 1,600 shares 25 years ago, and today it’s $50 for 130 shares, then $50.02 for 130 shares, then $49.98 for 130 shares, then $50.10 for 130 shares – and so on – the point isn’t whether the prices are pennies apart.

The point is those chasing pennies love this market and so become vast in it. But they’re not investors.  About 54% of current volume comes from that group (really, they want hundredths of pennies now).

Anything wrong with that?

Public companies, it demolishes the link between your story and your stock. You look to the market for what investors think. Instead it’s an arbitrage gauge. I cannot imagine a more impactful fact.

Traders, you can’t trust prices – the very thing you trade. (You should trade Sentiment.)

But wait, there’s more.

How often do you use a credit or debit card?  Parts of the world are going cashless, economies shifting to invisible reliance on a “middle man,” somebody always between the buying or selling.

I’m not knocking the merits of digital exchange. I’m reading Modern Monetary Theory economist Stephanie Kelton’s book, The Deficit Myth.  We can talk about credit and currency-creation another time when we have less stuff stewing our collective insides.

We’re talking about volatility. Why stocks like ETSY and BYND were halted on wild swings this week despite trading hundreds of millions of dollars of stock daily.

Sure, there were headlines. But why massive moves instead of, say, 2%?

The stock market shares characteristics with the global payments system.  Remember the 2008 financial crisis? What worried Ben Bernanke, Tim Geithner and Hank Paulson to grayness was a possibility the plumbing behind electronic transactions might run dry.

Well, about 45% of US stock volume is borrowed. It’s a payments system. A cashless society. Parties chasing pennies don’t want to own things, and avoid that by borrowing. Covering borrowing by day’s end makes you Flat, it’s called.

And there are derivatives. Think of these as shares on a layaway plan.  Stuff people plan to buy on time but might not.

Step forward to Monday, Nov 9. Dow up 1700 points to start. It’s a massive “rotation trade,” we’re told, from stay-at-home stocks to the open-up trade.

No, it was a temporary failure of the market’s payments system. Shorting plunged, dropping about 4% in a day, a staggering move across more than $30 trillion of market-cap. Derivatives trades declined 5% as “layaways” vanished.  That’s implied money.

Bernanke, Geithner and Paulson would have quailed.

Think of it this way. Traders after pennies want prices to change rapidly, but they don’t want to own anything. They borrow stock and buy and sell on layaway.  They’re more than 50% of volume, and borrowing is 45%, derivatives about 13%.

There’s crossover – but suppose that’s 108% of volume – everything, plus more.

That’s the grease under the skids of the world’s greatest equity market.

Lower it by 10% – the drop in short volume and derivatives trades. The market can’t function properly. Metal meets metal, screeching. Tumult ensues.

These payment seizures are routine, and behind the caroming behavior of markets. It’s not rational – but it’s measurable.  And what IS rational can be sorted out, your success measures amid the screaming skids of a tenuous market structure.

Your board and exec team need to know the success measures and the facts of market function, both. They count on you, investor-relations professionals. You can’t just talk story and ESG. It’s utterly inaccurate. We can help.

Traders, without market structure analytics, you’re trading like cavemen. Let us help.

By the way, the data do NOT show a repudiation of Tech. It’s not possible. Tech sprinkled through three sectors is 50% of market-cap. Passive money must have it.

No need for all of us to be on the skids.  Use data.  We have it.

-Tim Quast

Medium Frequency Trading

Liquidity is driving the stock market.  And I don’t mean the Federal Reserve.

I saw it firsthand in my Interactive Brokers account.  I bought 600 shares of GE at the market.  Limit orders are like turn signals in LA:  For sissies.  The average trade-size in GE is a little over 600 shares, or about $3,700.

Ergo, 600 shares should fill without issue. Well, pfffttth on that.

A hundred shares went to Island – which doesn’t exist anymore but the name lives on at Instinet, owned by Nomura. It’s an Alternative Trading System, a dark pool.

Island, for you market structure geeks (you know who you are!), invented the market as it exists today.  Look up Josh Levine and Jeff Citron (the guy behind phone company Vonage now).  And read Scott Patterson’s “Dark Pools.” 

Back to our saga. Another hundred shares filled at the NYSE’s Retail Liquidity Program, meaning I bought them from a high-frequency trader paid by the NYSE to improve the price a half-penny (price was $6.445).

If you’re suffering from insomnia, read this latest update on NYSE Arca fees. I’ll give you something to keep you awake:  a $50 American Express gift card to the first three people to tell me CORRECTLY what selling retail orders pays at NYSE Arca.

Oh, and this is why online retail brokers sell order flow and why high-speed traders buy it.  Read – or try to – the fee schedule. Think about all that complexity filtering through how algorithms route orders around the marketplace.

And think about how it might set the price for things while the good folks on CNBC are divining economic signals from what stocks do.

The exchanges are going to gush about how great this is for me.  You’re getting a better price, see!

Continuing:  Then a hundred shares executed at the very same price and time at CBOE EDGA, a stock exchange that pays traders about $0.18/100 to buy shares (to “remove,” the exchanges like to say, but adding and removing is selling and buying).  If you sell there, you pay.

Then 200 shares matched at $6.448 at a place simply labeled “DARK,” followed by another hundred in the same place, at the same price.

By breaking up my single order into five pieces (IBKR’s algo did it), the price went up.  GE’s market cap is $60 billion. It trades 130 million shares daily. Revenues are $120 billion. Should 20% of my 600-share order price it?

All these trades occurred at 9:25:52 AM MT.  I know the sequence only because the trades are ranked by succession.

This…process, or whatever you want to call it, is pricing the whole stock market.  It’s the Twister – remember that suggestive board game? – version of liquidity.

Our friends Joe Saluzzi, Sal Arnuk and team at Themis Trading are chuckling. This is daily life for them (read their book!).

But do you understand, readers?  GE’s business prospects didn’t do it.

I bought it because Market Structure Sentiment ticked up, signaling gains.  The vast bulk of trades are driven by the same motivation.  For me, it’s Medium Frequency Trading, or what probability for gains exists over a few days.

For most orders it’s far less, and for ten parties – no doubt some of them the same but there is a broker on each side of every trade – my MFT, my medium-frequency trade, was an HFT, a high-frequency trade.

Our collective acts set prices.  Trading heuristics – what do I get for this action versus that one?  Who will pay me to be ahead or behind this trade?

Exchanges and brokers say it bears little on the weight of the money in the market. Oh, really?  Try trading 600 shares of GE at the market. And why is the NYSE’s fee-schedule update 49 pages long?

Get takeout via DoorDash (a restaurant HFT, we could call them) and the price is in the cart. It’s not dependent on volume of food or whether you’re routing the meal to somebody else.

The restaurant pays for the service (or you pay delivery), but the cost is clear, and fixed. It appears to me traders made nearly three cents intermediating my fragmented order, and were paid an additional $0.51-$0.54 for setting prices.

What difference does it make? Motivation. During the placid, halcyon days of Nov 2019, stock prices still moved 2% up and down every day. In March it was over 12%. Right now it’s 3.3%. Every party to a trade gets paid. Price? Volatility.

Try reading the Nasdaq’s response to what we call Reg Nemesis II. No, seriously. Read it. Make sense?  Why is the market so fricking complicated?  Why are we told it’s good?

It lacks credulity. As do most of the daily links drawn to stock prices and headlines.

But I’ve got good news, public companies and investors. If we understand what’s happening, we can turn it to our advantage and influence it, both. I’ve got more to say on both counts soon. It’s happening.  Stay tuned.

Meanwhile, Market Sentiment bottomed last week but we’ve been moving at dog years in equities, so watch out. Things can change on a dime – especially either side of month-end window-dressing.

Rotation

There’s a story going around about an epochal rotation from momentum (growth) to value in stocks. It may be a hoax.

I’ll explain in a bit. First the facts. It began Monday when without warning the iShares Edge MSCI USA Value Factor ETF (VLUE) veered dramatically up and away from the iShares Edge MSCI USA Momentum Factor ETF (MTUM).

CNBC said of Monday trading, “Data compiled by Bespoke Investment Group showed this was momentum’s worst daily performance relative to value since its inception in early 2013.”

The story added, “The worst performing stocks of 2019 outperformed on Monday while the year’s biggest advancers lagged, according to SentimenTrader. This year’s worst performers rose 3.5% on Monday while 2019’s biggest advancers slid 1.4%, the research firm said.”

A tweet from SentimenTrader called it “the biggest 1-day momentum shift since 2009.”

It appeared to continue yesterday. We think one stock caused it all.

Our view reflects a theorem we’ve posited before about the unintended consequences of a market crammed full of Exchange Traded Funds, substitutes for stocks that depend for prices on the prices of stocks they’re supposed to track.

To be fair, the data the past week are curious. We sent a note to clients Monday before the open. Excerpt:

“Maybe all the data is about to let loose. It’s just. Strange.  Fast Trading leading. ETFs more volatile than stocks. Spreads evaporating. Sentiment stuck in neutral. More sectors sold than bought….Stocks should rise. But it’s a weird stretch ahead of options-expirations Sep 18-20.  It feels like the market is traversing a causeway.”

That stuff put together could mean rotation, I suppose.

But if there was a massive asset shift from growth to value, we’d see it in behavioral change. We don’t. The only behavior increasing in September so far is Fast Trading – machines exploiting how prices change.

What if it was AT&T and Elliott Management causing it?

If you missed the news, T learned last weekend that Activist investor Elliott Management had acquired a $3.2 billion stake in the communications behemoth and saw a future valuation near $60.  On that word, T surged Monday to a 52-week high.

T is the largest component of the MSCI index the value ETF VLUE tracks, making up about 10% of its value.  ETFs, as I said above, have been more volatile than stocks.

Compare the components of MTUM and VLUE and they’re shades apart. Where T is paired with VLUE, CMCSA ties to MTUM, as does DIS.  MRK is momentum, PFE is value. CSCO momentum, INTC, IBM value. PYPL, V, MA momentum, BAC, C, value.

Look at the market. What stuff did well, which did poorly?

The outlier is T. It’s a colossus among miniatures. It trades 100,000 times daily, a billion dollars of volume, and it’s been 50% short for months, with volatility 50% less than the broad market, and Passive Investment over 20% greater in T than the broad market.

T blasted above $38 Monday on a spectacular lightning bolt of…Fast Trading. The same behavior leading the whole market.  Not investment. No asset-shift.

What if machines, which cannot comprehend what they read like humans can, despite advances in machine-learning, artificial intelligence (no learning or intelligence is possible without human inputs – we’re in this business and we know), improperly “learned” a shift from growth to value solely from T – and spread it like a virus?

Humans may be caught up in the machine frenzy, concluding you gotta be in value now, not realizing there’s almost no difference between growth and value in the subject stocks.

Compare the top ten “holdings” of each ETF. Easy to find. Holdings, by the way, may not reflect what these ETFs own at a given time. Prospectuses offer wide leeway.

But let’s give them the benefit of the doubt. What’s the difference between MRK and PFE? V, MA, and PYPL and C, BAC and, what, GM and DIS?

Stock pickers know the difference, sure.  Machines don’t. Sponsors of ETFs wanting good collateral don’t.  Except, of course, that cheap collateral is better than expensive collateral, because it’s more likely to produce a return.

Such as: All the worst-performing stocks jumped. All the best-performing stocks didn’t.

What if this epochal rotation is nothing more than news of Elliott’s stake in T pushing a domino forward, which dropped onto some algorithm, that tugged a string, which plucked a harp note that caused fast-trading algorithms to buy value and sell momentum?

This is a risk with ETFs. You can’t trust signs of rotation.

We have the data to keep you from being fooled by machine-learning.

Short-Term Borrowing

Half the volume in the stock market is short – borrowed. Why?

It’s the more remarkable because stocks since late December have delivered an epic momentum rebound. A 15% gain is a good year. Half the sectors in the market were up 15% in just the last 25 trading days.

Yet amid the stampede from the depths of the December correction, short volume, the amount of daily trading on borrowed shares, rose rather than fell, and remains 48%.  That means if daily dollar-volume is $250 billion, $120 billion is borrowed stock.

What difference does it make? We’ve written before that the stock market now has characteristics of a credit market.  That is, if lending is responsible for half the volume, the market depends on short-term loans rather than long-term investment.

And share-borrowing, credit, will give the market a false appearance of liquidity.

Think about the sudden and massive December declines that included the worst-ever points-loss for the Dow Jones Industrial Average.  Was that a liquidity problem? Does a V-shaped recovery signal a liquidity problem?

Before the Dodd-Frank financial legislation, large banks might carry a supply of shares to meet the needs of customers, especially stocks covered by equity research.

With rigid value-at-risk regulations now, banks don’t hold inventory.  The supply chain for the stock market has shifted to proprietary fast traders, which don’t carry inventory either. They borrow it.

We define liquidity as the number of shares that can be traded before the price changes.  Prior to electronic markets, trade-sizes were ten times larger than today.  The mean trade-size the last five days was 181 shares, or about $13,500 against an average market price of $74.61.

But a few liquid stocks skew the average.  AAPL’s liquidity is over $23,000, its average trade-size. WMT is the average, about $13,000. GIS is half that, about $6,800.

AAPL is also 57% short – over half its liquidity is borrowed.  And AAPL is used as collateral by 270 Exchange-Traded Funds (ETFs). Related?

(Side note: Why would AAPL be used more than other stocks in an index if ETFs are tracking an index? Because ETFs only use a sample, often the biggest stocks that are liquid and easy to borrow.)

These three elements – fast traders, high borrowing levels, ETFs – are intertwined and they create risks of inflation and deflation in stocks that bear no correlation to fundamentals.

The market, as we’ve said before, always reflects its primary purposes. If the parties supplying the market with shares are borrowing them, they have an economic interest that will compete with the objectives of those buying shares as an equity investment.

Second, borrowing is a back-office brokerage function. With massive short-term securities lending, the back office becomes as important as the cash equities desk. And it’s a loan business, a credit market (a point made by the insightful academics comprising the Bogan family).

And ETFs? If you want to know how they work, read our white paper. ETFs are not pooled investments. They are collateralized stock substitutes. Derivatives.

Collateral is something you find in a credit market. ETF collateralization, the wholesale market where ETF shares are created and redeemed, is a staggering $400 billion per month in US equities, says the Investment Company Institute.

It’s cheap and easy for brokers to borrow the shares of a basket of stocks and supply them as collateral to the Blackrocks of the world (does Blackrock then loan them out, perpetuating the cycle?) for the right to create and sell ETF shares (or provide them to a hedge-fund customer wanting to short the whole Technology sector).

And how about the reverse? Brokers can borrow ETF shares and return them to Blackrock to receive collateral – stocks and/or cash that Blackrock puts in the redemption basket to offer in-kind for ETF shares.

These are the mechanics of the stock market.  It works well if there’s little volatility – much like the short-term commercial paper market that froze catastrophically during the financial crisis.

We are not predicting doom. We are highlighting structural risks investors and public companies should understand. The stock market depends for prices and liquidity on short-term borrowing. In periods of volatility, that dependency will amplify moves.

In extreme cases, it’s possible the stock market could seize up not through investor panic but because short-term borrowing may freeze.

How might we see that risk? Behavioral volatility. When the movement of money becomes frantic behind prices and volume where only a few firms like ours can see it, market volatility tends to follow (as Sept 2018 behaviors presaged October declines).

Currently, behavioral volatility is muted ahead of the Fed meeting concluding today, loads of earnings, and jobs data Friday. It can change on a dime.

Big Movers

You can’t expect the stock market to reflect earnings. I’ll explain.

By week’s end, 20% of the S&P 500 will have reported, and earnings are up 17% over the same period last year so far (normalized to about 7% sans federal corporate tax reform legislation).

Yardeni Research, Inc. reports that price-to-earnings ratios in various categories of the market are not misaligned with history.  The S&P 500 trades just over 16 times forward expected earnings, about where it did in 2015, and in 2007 before the financial crisis, and well below levels before Sep 11, 2001.

Sure, by some measures valuations are extreme. Viewed via normative metrics, however, the market is as it’s been. From 1982-2000, PE ratios were generally rising.  From there to 2012, they were generally falling. Yet between we had multiple major market corrections.

Which returns us to my incendiary opening assertion that earnings today don’t drive stocks. What does? The money setting prices. Let me explain.

Buy-and-hold money tends to buy, and hold. Most conventional “long” equity funds must be fully invested, which means to buy something they must sell something else.  Buying and selling introduces tax, trading-commission, and volatility costs, which can cause stock-picking investors to underperform broad indexes.

The Investment Company Institute reported that 2016 turnover rates among equity funds averaged 34%, or about a third of positions annually. Passive index and exchange-traded funds tout low turnover. State Street, sponsor for the world’s largest ETF, SPY, claimed 2017 turnover was 3%.

We’ll come to the fallacy of low turnover in ETFs.

First, Big Reason #1 for the movement of stocks is arbitrage. Follow the money. Using our proprietary statistical measures of behavior in stock trades, nearly 46% of market volume (20-day ave.) in the Russell 1000 (which is over 90% of market cap) came from high-speed traders.

They are not investors. These machines trade tick data in baskets, aiming most times to own nothing at day’s end. The objective is to profit on intraday price-moves.  For instance, 52% of Facebook’s daily trading volume is high-speed machines. Less then 9% is Active investment by stock-pickers.

Viewed another way, there’s a 46% chance that the price of stocks reflects machines trading the tick. Since less than 12% of Russell 1000 volume was fundamental, there is but a one-in-eight chance that earnings set prices. High-speed trading is arbitrage – profiting on price-differences.

Don’t fundamentals price the market long-term? Again, that would be true if the majority of the money setting prices in the market was motivated by fundamentals. That hasn’t been true this century.

How about fund flows?  Assembling data from EPFR, Lipper and others and accounting for big outflows in February, about $40 billion has come into US stocks this year.

Using Investment Company Institute data and estimates for Mar and Apr this year, ETFs have by comparison created and redeemed some $1.5 TRILLION of shares. Fund flows are less than 3% of that figure.

These “in-kind” exchanges between ETF creators and big brokers that form the machinery of the ETF market are excluded from portfolio turnover. If they were counted, turnover rates in ETFs would dwarf those for conventional funds. And the objective behind creations and redemptions is not investment.

ETF creators make money by charging brokers fees for these transactions (which are tax-free to them) and investing the collateral. Brokers then trade ETFs and components and indexes to profit on the creations (new ETF shares sold to investors) and redemptions (returning ETF shares to ETF creators in exchange for collateral to sell and short).

Neither of these parties is trying to produce an investment return per se. They are profiting on how prices change – which is arbitrage (and if ETF creations are greater than redemptions, they permit more money to chase the same goods, lifting markets).

Summarizing: The biggest sources of movement of money and prices are machines trading the tick, and ETF creators and brokers shuttling tax-free collateral and shares back and forth by the hundreds of billions. If pundits describe the market in fundamental terms, they are not doing the math or following the money.

And when the market surges or plunges, it’s statistically probable that imbalances in these two behaviors are responsible.

Paid Access

A day ski pass to Vail will now set you back $160-$190. It’s rich but I’m glad the SEC isn’t studying skiing access fees. It is however about to consider trading access fees and you should know, public companies and investors. These are the gears of the market.

We all probably suppose stock exchanges make money by owning turf and controlling access. Right? Pete Seibert, Earl Eaton and their Denver investors had a similar ski vision when in 1962 they bought a hunk of Colorado mountain down from the pass through which Charles Vail had run Highway 6. Control turf, charge for access.

In stock trading it started that way too. The Buttonwood Agreement by 24 brokers in 1792 that became the NYSE was carving out turf. Brokers agreed to give each other first look at customer orders and to charge a minimum commission.

This became the stock-exchange model. To trade at one, you had to have access, like a ski pass. Floor firms were called two-dollar brokers, the minimum commission. If you wanted to offer customers more services – say, beer at Fraunces Tavern with a stock trade – you could charge more. But not less.  No undercutting on price.

In the ski business, Nederland, Loveland, Wolf Creek and other ski slopes along the continental divide will undercut, letting you in for half Vail’s cost – but Vail wraps world-class value-adds around its access fees, like Mountain Standard and The Sebastian.

Suppose all the ski resorts could charge only a maximum rate for passes and were forced to send their customers to any mountain having a better price.  It would be inconvenient for travelers arriving in Vail via I-70 to learn that, no, the best ski price is now at Purgatory in Durango, five hours by car.

And it would be like today’s stock market (save for speed). The three big exchange groups, plus the newest entrant IEX, and tiny Chicago Stock Exchange, comprising currently 12 separate market centers, can charge a maximum price of $0.30/100 shares for access to trade. And still they all undercut on price.

That’s because rules require trades to match between the marketwide Best Bid or Offer (BBO) – the best price. As Vail would do in our imaginary scenario, exchanges must continually send their customers to another exchange with the best price.

How to set the best price? You can only cut price so much.  More people will still go to Vail because it’s close to Denver on the Interstate, than to Purgatory, halfway between Montrose, CO and Farmington, NM off highway 550.

Now suppose Purgatory paid to chopper you in from Vail. It might not move you out of The Sebastian, but you’d again have the stock exchanges today. While access fees are capped (and undercut), exchanges can pay traders to bring orders to them.

That’s called a rebate. Exchanges pay brokers incentives to set prices because if they can’t attract the BBO part of the time, they don’t match trades, don’t capture market share, can’t generate valuable data to sell to brokers (Only IEX is eschewing rebates).

The problem for investors and companies is that trades motivated by rebates are like shill bids at art auctions (which by the way are prohibited). They set the best price for everybody else yet the shill bidder doesn’t want to own the painting – or the shares. That’s high-frequency trading. It’s 40% of market volume on average and can be 60%.

Bloomberg reported yesterday the SEC is planning to study access fees through a pilot trading program next year. We’re encouraged that it may include a group of securities with no rebates. But the initial framework begun in 2016 under Mary Jo White aimed to lower access fees, and the study right now contains those plans.

Why? Exchanges are already lowering them. How about setting a floor on access fees so exchanges can make a decent return matching trades and don’t have to engage in surreptitious incentive programs to compete? I got the idea from the Buttonwood Agreement and 200 years of history.

Say all exchanges charge baseline access fees. If this exchange or that wants to wrap more value around fees – better data or more technology or beer – they can charge more.

Whatever happens, we hope (and I asked Chairman Clayton by email) the SEC makes an issuer committee part of the process. Without your shares, public companies, there’s no market. We should have a say. That’s why we have to know how it works!