Tagged: data

Trading Fast and Slow

 

Happy 2nd Half of 2020!  I bet we’re all glad we’re halfway there. Karen and I would’ve been in Greece now sailing the Ionian islands, luxury catamaran, sunset off starboard, Hurricane in hand.

Instead my checklist leaving the house has expanded from wallet, phone, keys, to wallet, phone, keys, mask.  No vacation for you. Just a Pandemic and executive orders.

A CEO of a public company said, “Why does my stock trade only 60 shares at a time, and how do I fix it?”

I was happy the team at ModernIR had highlighted shares-per-trade (one of several liquidity metrics we track).  Every public-company CEO should understand it.

After all, your success, investor-relations folks selling the story to The Street (and investors, whether you can buy or sell shares before the price changes), depends on availability of stock. Not how great your story is.

Because there’s a mistaken idea loose in the stock market. We understand supply is limited in every market from homes for sale to shishito peppers at the grocery store. Yet we’re led to believe stocks are infinitely supplied.

This CEO I mentioned asked, “So how do I fix it?”

This isn’t AMZN we’re talking about, which trades less than 30 shares at a time, but that’s over $70,000 per trade.  AMZN is among the most liquid stocks trading today. The amount you can buy before the price changes is almost eight times the average in the whole market.

AAPL is liquid too, not because it trades almost $16 billion of stock daily but because you can ostensibly buy $35,000 at a pop, and it trades more than 400,000 times.

Back to our CEO, above. The company trades about 3,000 times per day, roughly $3,000 per trade.  Liquidity isn’t how much volume you’ve got. It’s how much of your stock trades before the price changes.

That matters to both the IR people as storytellers and the investors trying to buy it.

I’ve shared several vignettes as I experiment with our new decision-support platform democratizing market structure for investors, called Market Structure EDGE. I couldn’t buy more than 10 shares of AAPL efficiently. My marketable order for JPM split in two (96 shares, 4 shares), and high-speed traders took the same half-penny off each.

The CEO we’re talking about meant, “What can we do differently to make it easier for investors to buy our shares?”

The beginning point, at which he’d arrived, which is great news, is realizing the constraints on liquidity.

But. If you have the choice to buy something $3,000 at a time – oh, by the way, it’s also more than 4% volatile every day – or buying it $70,000 at a time with half the volatility, which “risk-adjusted return” will you choose?

And there’s the problem for our valiant CEO of a midlevel public company in the US stock market today. Fundamentals don’t determine liquidity.

To wit, TSLA is more liquid than AAPL, over $42,000 per trade.

Forget fundamentals. TSLA offers lower risk.

Ford and GM trade around $4,000 per transaction, a tenth as liquid as TSLA.  Heck, NKLA the maker of hydrogen trucks with no products and public via reverse merger is twice as liquid as our blue-chip carmakers of the 20th century.

Prospects, story, don’t determine these conditions. These vast disparities in liquidity invisible to the market unless somebody like us points them out are driven by DATA.

The reason liquidity is paltry in most stocks is because a small group of market participants with deep pockets can buy better data from exchanges than what’s seen publicly.  I’ll explain as we wrap the edition of the MSM in a moment.

The good news is the SEC wants to change it.  In a proposal to revamp what are called the data plans, the SEC is aiming to shake up the status quo by among other things, putting an issuer and a couple investors on the committee governing them.

I’ve been trying for 15 years to achieve something like this, and so has NIRI, the IR professional association (they longer than me!). I should retire! Mission accomplished.

Heretofore, the exchanges and Finra, called “Self-Regulatory Organizations (SROs),” have been able to create their own rules. In a perfect world full of character, we’d all be self-regulatory. No laws, just truth.

Alas, no.  The exchanges provide slow regulatory data to the public and sell fast data for way more money to traders who can afford it.

By comparing the slow data to the fast data, traders can jump in at whatever point and split up orders and a take a penny from both parties.

This happens to popular brokerage Robinhood’s order flow by the way.  Those retail customers get slow data, and the traders buying the trades use fast data.

That’s not the problem. The problem is that by buying Robinhood’s order flow at slow-data prices, and selling it at fast-data prices, T Rowe Price’s trades get cut out, front-run, jumped past.

Your big investors can’t buy your stock efficiently (passive money doesn’t care as it’s just tracking a benchmark).

That’s why our CEO’s company trades 60 shares at a time.

For stocks like AMZN the difference between fast and slow is small because they’re the cool kids. And size grows. For the rest, it’s the crows in the cornfield.

You can talk to investors till you turn blue. It won’t solve this problem. The only thing that will is when investors join with public companies and get behind eliminating Fast Data and Slow Data and making it just Data.

We’ve got a shot, thanks to this SEC, and the head of the division of Trading and Markets, Brett Redfearn.  We should all – public companies and investors – get behind it. If you want to know how, send me a note.

Clashing Titans

While Karen and I consumed Arctic Char in Iceland, stock exchanges sued the SEC.

Talk about a big fish story.

The NYSE, Nasdaq and CBOE, representing about 57% of trading volume over a combined twelve platforms, asked a federal court to halt the SEC’s plan to test changes to trading fees and credits via a Transaction Fee Pilot program set to begin late in 2019.

It’s the more curious because investors transacting in markets generally support the planned study. Reading through hundreds of comment letters (including our own, offered in support), we tallied around $24 trillion of assets backing the test.

What’s got the exchanges in such a tizzy about a probationary effort – not an actual rule-change, mind you – that they’ve lawyered up?

The answer lies in the four key tenets of Regulation National Market System, a sort of current Magna Carta for stock-trading in the USA.  Every investor-relations professional and investor should know baseline facts about it.

In 1975 when the USA was mired in screaming inflation, a plunging dollar loosed from its gold moorings in 1971, an oil crisis, and failure in Vietnam, the US Congress decided to throw a fence around stock markets as a vital strategic interest.

So they passed the National Market System amendments to the Exchange Act of 1934, now part of the leviathan United States Code, 15 USC, Section 78c.

Oh boy.

The wheels of regulation mire in swampy muck, so it was thirty years later when the SEC finally got around to fulfilling the vision (no longer needed, in my view) Congress saw in 1975.  Enter Reg NMS.

The regulation has four pillars.  First, exchanges must ensure that stocks trade only at a single national best price (the Order Protection Rule). Second, the SEC – and this is the essence of the Fee Pilot – capped (the SEC said “harmonized”) what exchanges could charge so nobody would be priced out of access (The Access Rule) to stock quotes. It also required a spread between the bid to buy and offer to sell. No locked (same bid, offer) or crossed (higher bid than offer) markets.

Third, the law prohibited stock-quotes (unless under $1 in value) in increments of less than one penny (in effect, guaranteeing speculative traders a profit, because while quotes are in pennies, trades are often in far smaller increments at variations of midpoints) so every stock would have a bid and offer separated by at least a penny.

Finally, the law changed how revenue from data (Market Data Rules) would be allocated.

Reg NMS was implemented in 2007.

I think a legal case could have been made that both decimalization and Reg NMS six years later were unconstitutional. Nowhere does our governing charter give Congress the power to set prices and commandeer property not for public use.

Yet it did so by forcing exchanges to share what before had been proprietary intellectual property – data – and setting what they could charge for services. So how ironic is it that now exchanges are suing to keep this structure?

Not ironic at all when you realize that big exchange groups (the newest entrant IEX is not suing, supports the fee pilot, and didn’t build business on data and technology services) have exploded in size, profits, revenues and influence under Reg NMS.

Humans are an intelligent species. We adapt. The exchanges discovered that they could pay traders to set the bid and offer, and sell the data generated in that process – and then sell all kinds of services for using that data effectively.  Genius!

The pilot plan threatens this construct.

You’ll read that the concern is brokers routing trades for payments instead of where it’s best. That’s to me a red herring.

We oppose prices set by participants wanting to own nothing. They distort fair value, supply, demand, and borrowing. If you use our analytics, you know it’s 44% of trading volume directly, and over 70% indirectly (we calculate that 94% of SPY volume is arbitrage).

If half the volume is intermediation, the market is a mess.

The stock market is supposed to match investors and public companies. Reg NMS derailed the market’s central purpose. That’s my opinion. Predicated on data. I run a technology firm that for the entirety of the Reg NMS regime has measured the collapse of rational thought and capital-formation consequent to this regulation.

The stock market today is a great place to trade stuff. Exchange Traded Funds have prospered because they’re by law dependent on arbitrage, profiting on different prices for the same things. The regulatory structure requires different prices for the same things.

Let’s summarize what’s occurring. The exchanges have invested billions of dollars to make money under Reg NMS. And they’ve succeeded.

The SEC now realizes that Reg NMS hurts the root purpose of equity capital markets and it wants to test ways to roll back rules promoting short-term trading.

The exchanges are opposed because short-term trading is the very cornerstone of profitable data and technology services.

I don’t fault them.  But come on, guys. The SEC has finally seen how the market has devolved into a laser light show of speculation and fleeting intermediary profiteering that has pushed meaningful capital-formation into private equity.

That in turn defrauds mom and pop investors of the Intel Effect (I don’t have to explain it), and my profession, investor relations, of a thriving job market.

Could we run this test, please, and see what happens, exchanges?

Exchanging Data

Do we need another stock exchange?

I’ve been asked this question repeatedly since Bank of America Merrill Lynch, Charles Schwab, Citadel Securities, E*TRADE, Fidelity Investments, Morgan Stanley, TD Ameritrade, UBS, and Virtu Financial agreed Jan 7 to collaborate on seeking approval for a 14th official US stock market.

The answer? It depends on who “we” is, or are.

Adam Sussman of block-trading firm Liquidnet wrote that it’s an effort to lower trading costs which, thanks to high prices from exchanges for data feeds, have gone the opposite direction of trading commissions.

As to further fragmentation – more venues, less aggregation of buyers and sellers – Sussman says amusingly (the whole piece is funny) that “fragmentation is like having kids – after you have three of them, you just go numb to the pain.”

Michael Friedman, formerly of proprietary trading shop and technology vendor Trillium Management, said at TABB Forum (registration required) that these trading firms representing perhaps more than half of all volume resent how the exchanges keep raising prices for market data that brokers themselves create.

Before the exchanges IPO’d – all but IEX are now owned by public shareholders – they were member-owned, and members didn’t pay for data. Coincidentally the new market is called MEMX, or Members Exchange, anachronistically hailing a different era.

Friedman artfully unfolds market structure, explaining how a bid to buy shares at $9.08 at the NYSE cannot execute if the Nasdaq has a bid to buy at $9.10 because buyers willing to pay more are given legal priority and the trade must route out to the Nasdaq.

What if these firms were to route all the best trades – ones wanting to be the highest bid to buy or offer to sell – to themselves?  They could conceivably ravage market-share among big exchange groups until costs fell to a new equilibrium.

I think there are two other big reasons for this new cooperative.

One is easy to understand. Brokers are required to prove to customers that they provide “best execution,” or trading services that are at least as good as the average.  Paradoxically, that standard is predicated on averages for customer trades in the market – which concentrate heavily into the largest firms, including several MEMX backers.

If the order flow is consistently better than the average, it’s conceivable these firms could use their own data for free to meet best-execution requirements, a tectonic fist-bump amidst market rules.

So how would they boost odds that their data are better?  Look at who’s involved. They are mostly retail brokerage firms, or firms buying retail flow.

At Fidelity, about 97% of the firm’s retail orders are “nondirected,” lacking instructions about where the trades should occur. And well over 50% of those orders are sent to Virtu and Citadel.

Schwab says 99.6% of its trades are nondirected and 70% of them go to Virtu, Citadel and UBS.

And guess what?  Retail orders are permitted under rules to, in the jargon of market structure, “price-improve” trades.  The NYSE says its Retail Liquidity Program “can be used by retail firms directly as well as by the brokers who service retail order flow providers.”

Interactive Brokers, a firm for sophisticated retail traders and hedge funds, says retail orders with a limit, or set price, can be hidden from display at exchanges in increments of a thousandth of a dollar better than the displayed one, and the orders will float with a changing bid to buy or offer to sell.

That is, if the best bid to buy everyone sees is $9.08, a hidden limit order can be set at $9.081 and bounce like a bobber, staying always a fraction of a penny better than visible prices.

Under market rules, stocks cannot quote in increments below a penny. But they sure can trade in smaller increments, and they do all the time.

By aggregating retail order flow that market rules give a special dispensation to be better than other orders the members of MEMX believe they can not only match more orders but create the best market data.

How is it possible? Regulators wanted to be sure the little guy wouldn’t get screwed, so they give retail trades preference. They never dreamed innovative high-speed traders would buy it, or take advantage of rules permitting these trades to have narrower spreads.

It may work.

The problem is that the advantage MEMX hopes to leverage is a regulatory one that gives special access to one kind of activity.  (Editorial note: As we’ve written repeatedly, it’s just as Exchange Traded Funds have proliferated not by being better but through unique regulatory advantages giving them a private, wholesale block market with no transparency).

What’s it mean to investors and public companies? Investors, you could be picked off because MEMX could have compounded capacity to price-improve non-displayed orders. Public companies, something other than capital-formation is driving markets, which is not in your best interest.

We’d prefer a fair, level playing field serving investors and issuers, not rules permitting exceptions traders can game.

Data Darkness

There’s apparently a reality TV show called “Dating in the Dark.”

It must lack the cachet of Survivor or The Bachelor because you don’t hear much about it. The gist is that a number of people of opposite sexes wander around in utter blackness falling in love. You wonder how that’s superior to the displayed market – so to speak.

But in the equity market, dating in the dark is a big deal. I’m talking about how stock orders find each other. Take Coca-Cola (KO), which reported yesterday. From July 8-12, according to Fidessa’s Fragulator, 25.6% of trades occurred on KO’s listing exchange, the NYSE. But 29.4% were on the FINRA NYSE tape, a reporting facility for trades between brokers rather than on exchanges.

The remaining 44% of KO’s trading mostly met in displayed markets at the Nasdaq, BATS and Direct Edge, and the NYSE’s derivatives-centric platform called NYSE Arca, formerly the ECN Archipelago.

Why does this matter to you, IR professionals? It’s important to understand what’s happening. This is the market you manage – the equity market for your shares.

So, FINRA – the Financial Industry Regulatory Authority – is trying to address concerns that a large amount of stock-dating in the dark is bad for markets. That volume of KO’s on the FINRA NYSE tape? It’s “dark pool” trading, where buyers and sellers meet secretly and anonymously through brokers acting like millionaire matchmakers.

Last week FINRA sent a proposal to its members that would create new reporting rules for dark pools. If adopted, alternative trading systems, or facilities where the principal function is matching trades but the regulatory structure is one for broker-dealers rather than the regime exchanges operate under, would report their trades to FINRA on a delayed basis using a unique market-participant identifier. That way, FINRA would know what trades and volume occurred in each facility to better identify market-manipulation. (more…)