Friday, August 3, 2012

Dark Days at Knight Capital

Despite all efforts to corral Wall Street to avert a crisis, avoid market collapse, and instill confidence in the system, guess what happens. Yet another major misstep in the marketplace by one of its big participants.  And not just the rare market mistake that occurs once every year or two.  Missteps, hiccups, and strange collapses seem to be occurring these days just about every other week.

This week, it's Knight Capital, the equities market-making firm that announced losses of over $400 million after it launched new software in its trading systems.  Software errors and technology glitches led the firm's black boxes to spew large orders of errant trades. By the time the firm's humans (not machines) could discover what was happening, it was too late. The losses had piled up on trades the firm had no idea it had booked and would have never wanted to make in the first place.  The losses wiped out about half of its book-value equity, and now it struggles to survive intact. Until this week, Knight existed quietly in a niche role in stock trading (institutional equity brokerage, trading and market-making) and had been successful and well regarded in equity markets.

The trading mistakes resulted after Knight implemented new code to capitalize on a new form of retail-related trading with the New York Stock Exchange. They follow a series of embarrassments and other blatant mistakes all around the financial system the past two years.  Knight's loss reminds us of the 2010 "flash crash," when computerized trading involving futures and equities led to a sudden, shocking, unexplained nose-dive in stock markets.  Regulators afterward implemented safe-guard measures to reduce the probability of another flash crash. Or they thought they did. This week, regulators observed the unusual activity at Knight. The overall market corrected itself, but it was too late to save Knight from itself.

Beyond flash crashes and trading losses from bad computer code or faulty systems, there have been steady occurrences of bad events--almost enough to scare retail investors away from markets for the rest of the decade.  In just the past several months, MF Global, the large futures brokerage, collapsed after taking on large trading positions in Europe markets and losing hundreds of millions in customer funds. The upstart trading system BATS, specializing in stock match-making and high-frequency trading, planned its own IPO, but canceled it because of technology problems in its own infrastructure.

JPMorgan Chase this spring announced over $5 billion in trading losses from trading credit-default swaps in what was supposed to have been a safe hedge on its balance sheet.  The Nasdaq Exchange and Morgan Stanley are being blamed for the problems Facebook had on opening day in its IPO. Facebook and others say Nasdaq's systems mishandled orders and trading  in the first moments of trading in Facebook stock.  Nasdaq has acknowledged some errors, but claims those errors have little to do with underwriters' pricing the IPO too highly and with the precipitous decline in Facebook share prices since Day 1 of the IPO.

Last month another futures brokerage Peregrine Financial, less well known in futures markets, collapsed, and customers and regulators are panicking to locate their own funds.  And now Knight Capital.

Knight Capital is being described as a high-frequency, algorithmic-trading firm, although its evolution is more conventional.  The firm was launched in the 1990s to act as a market-maker of Nasdaq-traded stocks.  Retail brokerage firms from every corner of the U.S. send their brokerage orders to Knight or similar firms that always stand ready to make markets (buy or sell), based on order flow from investors.

As equities markets advanced, became more deregulated, dispersed and decentralized and trading volumes grew and trading became more computerized, Knight advanced, too.  It retained its market-making roots, but transformed into a sophisticated institutional trading organization with a prowess for high-frequency trading and black boxes humming all over its offices.

It traded on behalf of clients, counter-parties, institutions and itself. Clients routed their trades to Knight because Knight promised them fast execution and best prices (when they bought or sold stock).  If mom or pop or a Vanguard mutual fund bought or sold stock, chances are the trade was routed toward Knight Capital for execution, or at least Knight had a chance to see it and make a rationale bid or offer. Or let's say, Knight's computers.

All major institutional players engage in some form of high-frequency trading today, if only to survive and have a chance to squeeze tiny profits from the system--whether they are Knight, Goldman Sachs, hedge funds like Citadel or one of the many "algorithmic traders" with names like Getco, Jump Trading, Sun Trading, Peak 6, Quantlab or Susquehanna.  Some firms like Getco trade for their own accounts, jump into and jump out of markets in nanoseconds to make minuscule profits based on market tendencies and discrepancies.  Market-making firms like Knight promise clients they offer technology advantages to get them into and out of markets with fast execution, best prices, and handsome profits.

JPMorgan is so large and so well-capitalized (and some argue, so important to the global system) that its billions in losses in the second quarter caused an unsightly black eye, but nothing more than a financial sprained ankle. Knight's losses wiped out half the firm's capital base, and it now struggles for survival.  It may not exist a week from now.

What will likely happen?

The firm's CEO Thomas Joyce says the firm has sufficient "excess capital" based on SEC requirements. That means the firm meets SEC minimums for capital and the SEC can't force the firm to shut down immediately. It doesn't mean the SEC can't swarm the firm with regulators and investigators to see what happened, decide whether there should be penalties, or urge the firm to wind down or sell itself immediately.

At Knight, losses over $400 million imply the firm needs new funding immediately--new capital, new long- and short-term funding. Funding is necessary for ongoing, everyday operations--to support trading and market-making positions, to fund deposits at exchanges and clearing organizations, to pay down worried short-term lenders, or to pledge more collateral for other lenders. The longer it takes to replace the $400 million, the less likely Knight will survive in any form.

Already brokerage firms, trading counter-parties, institutional traders and hedge funds have stopped funneling trades its way, reducing revenue flow. They stand on the side lines to see how this story will unfurl. Perhaps they will resume trading once they see an outcome comfortable to them. Few outcomes, however, will be comfortable for Knight shareholders and employees.

The likely outcome?  With lenders, creditors, clients and counter-parties retreating and not willing to engage with Knight until after figuring out what happened--and with regulators and the public crying "mismanagement" or lashing out at technology-based trading, Knight Capital will likely have to sell itself in entirety or in bulk pieces.

Once it is confirmed that the firm's loss was due to technology or programming errors and not fraud or an attempt to do something illegitimate, larger financial institutions funds will see value in its parts or whole.  There is value in its order flow from hundreds of broker/dealers and hedge funds with long-term relationships with the firm (if they all choose to return when the madness dims). There is value in its existing memberships, tie-ins, and plug-ins to equity and futures exchanges. And there may be value in the existing black boxes that have worked well in normal markets (if we assume that the problem this week was all due solely to the firm's lack of patience in testing new trading code).

Bankruptcy is always a route to get to the best possible result for shareholders, who will no doubt suffer. CEO Joyce's acclaimed career in equity markets may end and his successful efforts to turn Knight into an important market participant could be forgotten.

Have we come to expect mammoth market mistakes to be the new normal, despite the good intentions of new regulation and oversight? What else could possibly occur this month, when we usually think Wall Street slows down for vacation?

And just think, all the turmoil and chaos this week because of a 15-minute, computer mistake in Jersey City.

Tracy Williams

For more, see also:

CFN:  JPMorgan and Trading Losses, 2012
CFN:  MF Global and Its Demise, 2011
CFN:  Facebook and Its Rough IPO, 2012

No comments:

Post a Comment