How A Software Glitch Essentially Killed A Major Wall Street Company

Photo: YouTube/Screenshot

This post has been updated.This week, a soft­ware glitch essen­tially killed a com­pany that had become a major force on Wall Street in the last 20 years.

Knight Cap­i­tal is one of the largest trad­ing orga­ni­za­tion on Amer­i­can stock exchanges and lost a stag­ger­ing $440 mil­lion in a bit over a half hour on Wednes­day morn­ing, push­ing the com­pany to the brink and dri­ving some of its largest cus­tomers away and the cause of it all was a set of soft­ware bugs.

High speed trading
For the last few years, there’s been a war in the stock exchange world, with each side try­ing to out­smart the other through soft­ware. In the finan­cial world, the phe­nom­e­non is known as high speed trad­ing, or high fre­quency trad­ing, where stock pur­chase and sale orders are sent and matched in mil­lisec­onds. Along the way, each wall street firm has work to find ways to exe­cute trades faster and in larger quan­ti­ties to gain an advan­tage over their competitors.

In this world, engi­neers and maths­e­mati­cians develop increas­ingly com­plex for­mu­las to shave a few cents of prof­its on a trans­ac­tion. It’s a world where soft­ware and hard­ware com­bine to deliver results that can be counted in mil­lions and bil­lions of dol­lars; a world where the finest algo­rithm is the best weapon; a world where co-locating a com­puter server a cou­ple of feet closer to the exchange machines comes at a pre­mium; a world that is not too far away from wall street phys­i­cally but eons away from the world of trad­ing that existed only a cou­ple of decades ago.

In this world, speed of trade exe­cu­tion mat­ters and in that world, few were as fast or exe­cut­ing as many trades as Knight Cap­i­tal Group. On any given day, between 15% and 20% of all trades on the NYSE and NASDAQ went through Knight Capital’s computers.

In early July, the New York Stock Exchange (NYSE) received SEC approval for a new pro­gram called the Retail Liq­uid­ity Pro­gram (RLP). The pro­gram, which was slated to kick off on August 1st, would allow retail traders to get access to bet­ter pric­ing on stock. The spec­i­fi­ca­tion for the pro­gramme had been around for less than a year but a lot of wall street firms had planned for its imple­men­ta­tion. There was some con­tro­versy when the announce­ment was made but all mar­ket mak­ers started prepar­ing for the launch, real­iz­ing that this may be another tool to fight the com­pe­ti­tion and that those who were not ready by launch time could risk los­ing out pre­cious advan­tages in the market.

Knight Cap­i­tal, as one of the largest play­ers in the mar­ket, decided that it would pro­vide that capa­bil­ity to its cus­tomers as soon as it became avail­able and that it would take advan­tage of the price dif­fer­en­tial for its own trad­ing. With only a few weeks from the announce­ment to the release, inter­nal devel­op­ment teams must have burned the mid­night oil to get this in place by the August 1st deadline.

Knight goes down
But on Wednes­day morn­ing, Knight Cap­i­tal may have made a mis­take that could cost it its exis­tence. Accord­ing to reports, the com­pany had just fin­ished upgrad­ing its soft­ware to the newer ver­sion, allow­ing it to take advan­tage of the RLP.

When the NYSE opened, trades on as many as 114 com­pa­nies started run­ning at higher than expected vol­umes. Here’s how the New York Times describes what hap­pened next:

A New York Times analy­sis of New York Stock Exchange vol­ume on Wednes­day morn­ing showed that dur­ing the first minute of trad­ing there was 12 per­cent more trad­ing in all stocks than there had been on aver­age dur­ing the pre­vi­ous seven days. By the third minute of trad­ing there was 116 per­cent more trad­ing than the pre­vi­ous week’s aver­age. The dif­fer­ence reached a peak at 9:58 a.m., when the vol­ume was six times greater. After that, trad­ing vol­ume fell off sharply, near­ing the recent aver­age at 10:15 a.m.

[…]

Once the prob­lems began, many traders said it would have made sense if the firm’s employ­ees had not caught the prob­lems for the first minute or so, given the speed at which Knight’s pro­gram was fir­ing off orders. After that, though, the prob­lems were vis­i­ble for all to see.

In under an hour, the com­pany had lost $440 mil­lion, $75 mil­lion more than all the cash it had in its reserves, or more than it had made in profit for the pre­vi­ous year. Cus­tomers, wary of suf­fer­ing from the bug, decided to start trad­ing through other firms. By Thurs­day, a num­ber of bro­ker­ages and broker-dealers (TD Amer­i­trade, Scott Trade, Fidelity Invest­ments, Van­guard Group, E*Trade and Per­sh­ing LLC, a divi­sion of BNY Mel­lon) sus­pended rout­ing trad­ing orders through the com­pany. A few came back but not all.

Mean­while, investors in the com­pany sent its price tum­bling down, along with the rest of the stock mar­kets’ as con­fi­dence in the well oiled machin­ery of stock trad­ing was shaken. The trad­ing losses forced the com­pany to open itself up to the prospect of either bank­ruptcy or a sale, with many poten­tial suit­ors circling.

For all intents and pur­poses, Knight will come out of this as a sig­nif­i­cantly smaller com­pany, if it remains a stand­alone com­pany at all. Mean­while, it has become a cau­tion­ary tale to the rest of the indus­try, leav­ing many in the finan­cial world won­der­ing if they could be next.

A dan­ger­ous pattern
The Knight Cap­i­tal fail­ure, how­ever, is not the first one related to soft­ware in the finan­cial world. A cou­ple of years ago, soft­ware was seen as the power behind a flash crash when the Dow Jones lost nearly 10% of its over­all value in a few min­utes. Last March, the dol­lar lost 5% of its value against the yen and the price of Cocoa futures lost almost 15% of their value, both inci­dents that were results of high fre­quency trading.

Mean­while, BATS, one of the sec­ondary mar­kets, tried to go pub­lic in late March but soft­ware Sna­fus force the com­pany to can­cel its IPO and even impacted the price for Apple’s stock. And even the most antic­i­pated IPO of the year, Face­book, was marred by tech­ni­cal glitches as the NASDAQ.

With 60% to 70% of all stock trades being exe­cuted by auto­mated com­put­ers, there is grow­ing con­cern that soft­ware is increas­ingly threat­en­ing the sta­bil­ity of our mar­kets. Along the way, volatil­ity has gone through the roof, which can send algo­rithms on large sale cycles that can bring mar­kets increas­ingly down.

The trend will con­tinue to grow­ing automa­tion and there is a wide­spread feel­ing around the finan­cial world that this is just the begin­ningof these types of phe­nom­e­non. In 2008, I started won­der­ing if volatil­ity was the new nor­mal; in 2012, I am increas­ingly con­vinced (and con­cerned) that algo­rithms will not only lead to increased volatil­ity but also down a path where few peo­ple may be under­stand­ing how finan­cial mar­kets are func­tion­ing, poten­tially detach­ing them from the real­i­ties of eco­nom­ics fundamentals.

Because so many peo­ple, from the cor­po­rate world to the polit­i­cal one and every­one in between, are look­ing to stock mar­kets as indi­ca­tors of the health and wealth of economies, we may find our­selves in a sit­u­a­tion where the eco­nomic indi­ca­tors are so divorced from any fun­da­men­tals to soft­ware could end up not only killing com­pa­nies (as it is doing for Knight Cap­i­tal) but pos­si­bly even take nations down with them. And that is the dark side of dark pools.

NOW WATCH: Money & Markets videos

Want to read a more in-depth view on the trends influencing Australian business and the global economy? BI / Research is designed to help executives and industry leaders understand the major challenges and opportunities for industry, technology, strategy and the economy in the future. Sign up for free at research.businessinsider.com.au.