Working on the Knight moves

By now, most everyone is familiar with the story of Knight Capital Group.  Basically, an algo went bonkers, a bunch of stocks got roiled at the open, and Knight lost $440 million.  Over the weekend, Knight found a white knight to infuse it with capital in a highly dilutive convertible preferred sale.

Today,, in conjunction with Nanex, posted a great piece on exactly what happened on that fateful morning.  As you read the Traderswire post, do yourself a favor and click through to the Nanex link.  As usual, Nanex provides pictures of the mess that are worth a thousand words.  Or $440 million.



Of Books and Guillotines

Being a fairly boring guy with too much time on my hands, I tend to read a lot.  Much of what I read these days pertains to our capital markets — from postmortems of the 2008 crisis, to analyses of the 2010 flash crash, to wonky books on the nature of high frequency trading.  I also read some Stephen King now and then, which often leads to less horrific outcomes than those we often meet in the financial arena!

In the book “Confidence Men, ” one is left to decide whether government ineptitude or Washington/New York collusion placed the big banks in the profitable position they are today — after nearly cratering the global economy.  The part about Ken Lewis extorting $20 billion from Hank Paulson is terrific.

There were no heads on pikes.  Just rapid returns to profitability.

In “Dark Pools, ” (a book oddly lacking much content about actual dark pools but long on the surface level analysis of high frequency trading) a moral argument is put forth.  Should the fastest, most ruthless, amoral, wealth-mongering individuals be legally compensated with hundreds of millions of dollars while breaking down the very structure of our capital markets? (Note:  the value judgement is the author’s, not necessarily mine.)

The question is interesting in the abstract, but rhetorical in a world in which I watch stocks close on their VWAP (Volume Weighted Average Price) every day as if they were a jet being eased onto the runway by a skilled pilot.

We practice our trade in a world where the maker/taker fee system, set up by the exchanges, compensates the bad actors for the very behavior of acting badly.  But if my algo can swallow yours, (while creating trade volume, of course), good for me.  Price discovery be damned.

I’m not about to opine on the issue of societal good versus individual financial gain in this blog post.  It’s a discussion I’d welcome, if any readers would like to discuss it in the comments section.

It just strikes me that more, and more extreme, writings are cropping up that favor draconian resolutions to today’s (perceived or real) government/corporate exploitations.  The 1% is under assault; both for their level of wealth and the methods they used to attain it.

“You didn’t build that” might be better stated as “You shouldn’t have built that.”  You shouldn’t have built the “too big to fail” bank (with all due respect to Sandy ‘Born Again Financier’ Weill.)  You shouldn’t have built the algorithms that now constitute over 70% of the stock market’s volume while old-school investors recede into the shadows.  You shouldn’t have built a policy where bondholders were exempt from normal, market-based haircuts while the rest of the country was in financial flames.

But, I digress.  I often do.

After just having finished “Confidence Men, ” I started perusing the Wall Street Journal.   In the editorial section, I came across a very interesting question…  Would capital punishment for financial crimes be effective in quelling what we’ve been experiencing in the financial markets?  It’s a funny question in its absurdity, but the history of times when the question was not so absurd shines some light.

Rather than paraphrasing, here’s the link in its entirety.  Issac Newton and all.

Clowns to the Left of Me, Jokers to the Right

We frequently blog about the structural strains put on the market by algorithmic and high frequency trading.  We’ve also blogged (more than a few times) about the distortions in price discovery caused by central bank intervention, zero interest rate policy, etc. (most recently pointing out that 80% of the annual equity premium since 1994 was attributable to trading in the 24 hours before a Fed announcement).

This morning, however, saw a convergence of the two we seldom get to “enjoy.”

First, today’s episode of “Algos Gone Wild” was set at the NYSE, where 148 symbols began trading wildly and with ludicrous volumes at the open.  While there seems to be some linkage to Knight Capital, the specifics are not yet clear.  What we do know is that these stocks started swinging inexplicably by magnitudes of 10% or more.

A quote from a director of floor trading:  “Stocks are moving all over the place, they are weird, and they are trading like millions of shares 100 shares at a time.”

An example, provided by CNBC, would be Molycorp.  This stock traded 5.7 million shares in the first 45 minutes of trading.  Typically, Molycorp trades about 2.65 million shares per day.

If there is an upside to this story, the NYSE quickly halting the affected stocks may very well have averted another flash crash.  Small consolation in the scheme of things.

While the algo drama was unfolding, Louis Bacon put out a letter regarding his multi-billion dollar Moore Global Investments fund.  Bacon is voluntarily returning $2 billion to his investors, for the following reasons:

  • “The markets have been riskier and less liquid.”
  • “Markets are increasingly distorted by central banks’ attempts to squeeze drops of growth from an over-indebted private sector across much of the developed world.”
  • The US markets are hindered by “a caustic political environment and an anti-business administration.”
  • “The [Eurozone] banking authorities have been a special case in ineptitude, [waiting to raise bank capital] until it is largely infeasible.”
  • “Disaster Economics, where assets are valued based on their ability to withstand a lurking disaster as opposed to what they may yield or earn, is now the prism through which investors are pricing markets.”

Pretty strong words from the 238th richest American and uber-successful hedge fund manager since 1987.

To get some sense of how prevalent the Louis Bacon-type behavior has become in the hedge fund industry, Google “hedge funds calling it quits.”  Bring a cup of coffee, because you’ll be spending a lot of time reading your computer screen.

These are indeed weird times.  Navigating them requires extreme care in both capital allocation decisions and sizing of positions.

To paraphrase the 1972 song by Stealers Wheel:  “Fed to the left of me, algos to the right.  Here I am stuck in the middle.”


If it were hockey we’d call it a Facewash. In IPO’s it’s called a Zuckerpunch.

The Facebook launch on Friday was much like that of a North Korean missile – much hyperbole, much cheerleading, displays of excited citizens crowding the public square, a launch…  then a thud.

At this point, I have no interest in in opining on Facebook’s pricing, revenue persistence, market penetration, or all that fundamental stuff.  Rather, I thought it may be interesting to point out some of the oddities that occurred during the day of the Taepodong-2… I mean, Facebook launch.

Let’s start with the opening.  Slated to launch at 11:00, the first trade was delayed until 11:20.  While Nasdaq is playing the role of Korean Rocket Scientist today, Telis Demos from Trading Technology wrote a nice piece on the cause of the delay.  Here are a few quotes from the article (although I encourage you to read it in its entirety by clicking the link):

“And it was one quote cancellation sneaking into a five-millisecond window that caused about 20 painful minutes, watched live by the world on CNBC television, of the delay to the opening Facebook’s public offering on Nasdaq’s US market.”

“In brief, the problem was that the system took two extra milliseconds to calculate the opening price. Because of a decision before to allow continuous order placement during IPOs, cancellations kept “fitting in between the raindrops”, in the words of Bob Greifeld, Nasdaq’s chief executive, in the five milliseconds it was taking to determine a price.”

“As a result, Nasdaq had to manually override the process, which took up those 20 minutes. But the manual process meant that individual order confirmations were not sent out until almost 1.50pm, hours after the 11.30am opening print. Confusion reigned, and many blamed it for dampening demand for Facebook stock.”

I’m no rocket scientist myself, but this sure looks like the fingerprints of high frequency trading.  Sneaking a big order cancellation in a five millisecond window would have been impossible for a human being – even the late, great Benevolent Dictator with all his superhuman skills.

The next oddity is brought to us by Nanex Research.  Again, I suggest you visit their site for some great graphic illustration, but I’ll summarize a few of the conclusions here:

  • Between 11:54 a.m. and 1:50 p.m. Nasdaq quotes stopped coming through the conventional channels.  To quote Nanex, “Those who are co-located and get the direct feeds, namely HFT’s, didn’t experience this problem, as trades continued to come from Nasdaq.”
  • At around 1:50 p.m., trades began mysteriously executing 120 milliseconds before the bids printed.  Nanex calls this “fantaseconds, ” or evidence that HFT’s may now be able to 1) trade faster than the speed of light, or 2) orders are being routed in an unconventional manner that may be outside of the regulations.
  • Later that same second (yes, as in 1/60 of one minute), trades were executed 900 milliseconds before the quotes were printed.
  • At the same time, there were stretches where Facebook trades accounted for 100% of all Nasdaq stock trades.

The final extreme characteristic of this launch was Morgan Stanley’s role in supporting the missile’s $38 price.  From Reuters:

  • “Morgan Stanley may have spent billions of dollars to support the stock price by buying shares in the market. Some market participants said that the underwriters had to absorb mountains of stock to defend the $38 level and keep the market from dipping below it.”
  • Morgan Stanley had access to 63 million shares in the over-allotment option, giving them plenty of ammo to hold the line at $38.
  • Reuters also points out:  “As an indication of the cost, had Morgan Stanley bought all of the shares traded around $38 in the final 20 minutes of the day, it would have spent nearly $2 billion.”  That’s a lot of cabbage, if the stink over JP Morgan’s trading loss is any indication.

I think the autopsy of this IPO in the coming weeks will prove to be both interesting and educational.  Has the HFT become self-aware and turned on its creator (Nasdaq)?  What new procedures will be in place to keep a debacle like this from happening on future IPO’s?  How much of a hit to capital will Morgan Stanley ultimately feel?

It’s rare that market sausage-making happens on a stage as large as the Facebook IPO.  Maybe that will help be a catalyst for better understanding our brave new trading world.

Of correlations and ETF’s

For those who followed our monthly letters and conference call, you’ll recall the theme for both was “correlation.”  In recent months, we’ve noticed increasing correlations between and among the constituents of various markets.  In the products we manage, all of which are meant to be non-correlated, correlations have doubled in the past year or so.  While we have taken action to drive down correlations within our portfolios, the question still remained… “What is causing this dangerous correlation trend?”

While I typically consider CNBC to be fast food financial news (ie. the USA Today of investing), their stock correspondent, Herb Greenberg, posted a timely article considering what we’ve been observing here at Altair.  Rather than summarizing it, I’ll just attach the link.  Suffice it to say, we agree with Herb.  The elimination of leveraged ETF’s and the robots that trade them would bring some sanity back into the marketplace.