Studies of this new population of do-it-yourself traders invariably show that increased access to trading tools and market data creates the illusion of market competency and encourages poor decision-making. Even before net access, do-it-yourself investors tended to make poorer investment decisions than those who used financial advisors or, best of all, invested in entirely unmanaged “index” funds. The main reason for DIY investors’ poor results? Amateurs trade too much. Meanwhile, online trading brokerages—whose profit comes almost solely from commissions on trading—have a stake in getting their users to make as many trades as possible.
As the data shows, the same brokerage houses profit from the same trading activity in the same way they always have, while retail investors’ actual percent of profits and trading accuracy goes down. That all this is occurring during the Internet revolution hasn’t helped these new investors make informed decisions either. On the contrary, when people have access to more data through which to predict the future, their confidence in their predictions increases much more than the accuracy of their predictions.27 So the more data an online site can provide its traders, the more secure they will feel trading. Likewise, the more control a trader feels over his activities, the more trades he will make and the more conviction he will have about them—and the poorer he will do.28
Fully aware of the psychological influence of information access and the illusion of control, online trading brokerages develop advertising campaigns that exploit both of these vulnerabilities. eSignal proclaims, “You’ll make more, because you know more.”29 An advertisement for Ameritrade explains that online investing “is about control.”30 The ever-growing assortment of online trading tools for individual investors may create the sense that the game has opened up but offers the individual investor little more than the vicarious thrill of participation. There are lots of buttons, menus, data points, and choices, but most of them make no real difference to the outcome. The trading itself is real, as are the commissions and losses, but the access to equal footing on the playing field is a digital simulation.
DO ALGORITHMS DREAM OF DIGITAL DERIVATIVES?
Professional traders face similar challenges in the digitized environment. Like those in other disrupted industries, many finance workers have been replaced by networks and computers.
Human stockbrokers, in addition to providing access to markets, used to be responsible for giving clients the best information on a stock or sector, as well as advice and numbers on allocations and future earnings. Trades then went through the specialists—designated market makers—who owned a pool of a particular stock, which they used to fill orders when there was no ready counterparty. They were required to serve as buyers of last resort, preventing a stock from crashing unnecessarily due to a temporary lack of liquidity. Yes, they made money doing this, exploiting their privileged position on the trading floor to buy low and sell high. But their activity reduced volatility and kept markets more consistently liquid and orderly.
Thanks to the net, customers can now access markets—or at least virtual trading desks—directly. The digitized marketplace doesn’t require brokers or specialists to function. Trades are executed via “straight through” processes that connect buyers and sellers from around the world. Real-time quotes from electronic trading desks make pricing more transparent; the decentralized nature of the networked exchanges anonymizes deal making and reduces favoritism; and digital record keeping increases accountability and archives any nefarious practices for future audit and prosecution. With multiple and competing exchanges instead of just one specialist, traders also get tighter spreads between the ask and bid, which means less money is leaked out of each trade.
But the elimination of human specialists also eliminates the buyers of last resort and the dampening of volatility they provided. Into the power vacuum and hungry for this increased volatility come the high-frequency traders (HFTs) and algorithms—computer programs that look to profit exclusively through the exploitation of temporary, even microsecond-long, imbalances in trading.
Sure, in many cases, having an algorithm that scans for a lack of volume of a particular security on an exchange somewhere is a good thing. The high-frequency trading algorithm may charge a few pennies extra for having found the stock someone wanted, but the specialist charged a few pennies for his services, too. The problem is that, unlike the specialists who were obligated to reduce volatility in the shares they serviced, HFTs like volatility. For instance, one typical HFT strategy is to provide liquidity in a particular stock until the market shows signs of instability. The algorithm then suddenly withdraws all its bids and offers, leading to an immediate dearth of demand and a precipitous price drop. The smart algorithm, knowing it can make this happen, has already bet against the stock with derivative options. When the other algorithms realize what’s happening, they freeze up, too, leading to a “flash crash.” The stock goes down, but for no real-world reason. It’s just collateral damage from the game itself.31
Another common algorithm strategy is to flood the quote and order systems with fake trades—orders of intent but not full executions—to convince human traders (or other algorithms) that the market is moving in a particular direction. More than 90 percent of all quotes are fake gestures of this sort, generated by computers.32
Algorithms run on ultrafast computers connected as physically close to the stock exchange computers as possible. This gives them a processing and latency advantage over their peers and any remaining human traders. A well-located algorithm can take action between the moment a more distant human or computer makes a trade and the moment that trade is fully executed.33 This is essentially “front-running” other people’s purchases—buying the shares they intend to buy and then selling them to those same traders at a profit. Although it may cost the trader only a few pennies extra, those pennies add up when algorithms perform these routines millions of times a day, extracting real value from the market.
To be clear, the algorithms are providing no service. Any liquidity they might create is more than compensated for by the liquidity they take away when they’re seeking to generate volatility or panic selling. They disadvantage not only human brokers but also the individual investors that a digital stock market was supposed to empower. Algorithmic trading doesn’t happen on a laptop connected to the net by Wi-Fi. It requires the kind of hardware, connectivity, and real-estate location that only the wealthiest, most established firms can afford. It may be disruptive to trading, but it only enhances the advantages of the traditional players—or at least the firms they worked for before they were replaced by machines.
New exchanges are emerging to counteract some of these trends. Ironically, they evade algorithms by slowing down the speed at which they communicate their trades. As we’ve seen, algorithms exploit traders by intercepting their communications as they make their way from one trading desk to another. If a trader is selling a thousand shares of a stock, the order typically goes to a few different virtual trading desks. Because these desks are located in different places, the order may get to one desk a half second before it gets to the second one, or even a whole second before it reaches the third. An ultrafast computer can see what’s happening on the first trading desk, then race ahead to the second and third desks and front-run the trade before the original order reaches them. To counteract this, the new anti-algorithmic exchanges calculate how long it will take orders to reach the furthest exchanges, then send orders out in a time-delayed fashion. They send the order to the furthest exchange first and to the closest exchange last. This way the order reaches all the exchanges at the same moment. Even if the algorithm witnesses the first order, it can’t front-run the others, because they’ve all been executed.34
And for right now, anyway, that’s what constitutes a winning strategy on Wall Street. It’s a game being played between algorithms exploiting the trading protocols. It has nothing to do with providing capital to growing companies, and everything to
do with extracting value from the investment economy by undermining the very premise of open markets. It is gaming the system.
Stock is intended to be an instrument through which entrepreneurs can raise capital for new businesses or expansions. In exchange for cash, the investor gets a piece, or share, of the company. Once that initial transaction has been completed, however, the rest of that share’s journey is inconsequential. Any increase in share value goes to the trader, not the company. The only way the company gets more capital is by selling or issuing more shares.
Still, the value of a share of stock is related to the fortunes of the company. If a company grows, shareholders own pieces of a bigger pie; if the company makes sufficient profits, these are paid up to shareholders as dividends. That’s why investors traditionally make decisions based on research about a company, its management, and the business conditions surrounding it.
In contrast to investors, who are looking to grow money over time by assessing the true value of companies, traders seek to profit from the changing prices of stocks and bonds. The underlying worth of a company doesn’t really matter. The trader is looking at ebbs and flows, trend lines and moving averages, bubbles and crashes. For the trader, the massive amounts of data and processing capabilities unleashed by digital technology are important only insofar as they offer new ways of strategizing moves in the game.
Digital publishers from Bloomberg to Yahoo Finance are more than happy to satisfy the traders’ insatiable appetite for charting and data visualization. Acting a bit like algorithms themselves, traders employ stochastics and momentum oscillators to bet on volatility itself. Of course, their trades are effortlessly preempted by the real algorithmic traders—the algorithms themselves. Those algorithms, in turn, battle other algorithms, all competing to strategize on top of each other’s strategies, in successive layers of derivative transactions, each one more abstracted from the last.
Those who play in this space, from individual technical traders to the operators of algorithmic programs, feel they have gotten into the very core of the game—the rule writing itself. In actuality, they may be dominating the landscape of trading, but they are operating on a plane far removed from the companies and investors whose shares they are leveraging. The tremendous volume of activity they generate only disconnects the marketplace further—and extracts more value—from the commerce they were originally created to fuel.
The sheer volume of derivative finance dwarfs that of the real markets. In 2013, the value of the derivative market was estimated at approaching $710 trillion35—that’s almost ten times the size of the nonderivative global economy36 and forty-seven times the size of the U.S. stock market.37 Less conservative estimates put the derivatives market at $1.2 quadrillion, or more than twenty times the world economy.38 It’s only fitting that in 2013, a derivatives exchange called Intercontinental Exchange ended up purchasing the NYSE itself.39 The stock market—already an abstraction of commerce—was swallowed by its own abstraction.
As markets are increasingly driven by all this virtual gamesmanship, they become more volatile and harder to read in terms of any fundamentals. Minor swings of companies and sectors lead to bubbles and crashes of unprecedented speed and magnitude. The sheer turbulence of all this digital trading creates something of a weather system all its own.
Indeed, the more that algorithms dominate the marketplace, the more the market begins to take on the properties of a dynamic system. It’s no longer a marketplace driven directly by supply and demand, business conditions, or commodity prices. Rather, prices, flows, and volatility are determined by the trading going on among all the algorithms. Each algorithm is a feedback loop, taking an action, observing the resulting conditions, and taking another action after that. Again, and again, and again. It’s an iterative process, in which the algorithms adjust themselves and their activity on every loop, responding less to the news on the ground than to one another.
Such systems go out of control because the feedback of their own activity has become louder than the original signal. It’s like when a performer puts a microphone too close to an amplified speaker. It picks up its own feedback, sends it to the speaker, picks it up again, and sends it through again, ad infinitum. The resulting screech is equivalent to the sudden market spike or flash crash created by algorithms iterating their own feedback.
Traditional market players scratch their heads at these outlier events because they can’t be explained in terms of trading activity between humans. What made that bubble burst? Was it market sentiment, a piece of news, or something being overbought? None of the usual suspects indicated trouble. That’s why it has become popular to label these gaps in rationality “black swans”—as if they are utterly unpredictable anomalies.
In fact, they are entirely predictable. We might not know exactly when these extreme events are coming, but we know they will, because that’s the way nonlinear systems express themselves. We are not witnessing momentary crises in the capitalization of business; we are watching a high-stakes video game among the nonhuman players of the wealthiest investment houses. At best, we humans are carried along for the ride.
These boom-and-bust cycles—whether in the microscopic moment of a single hijacked trade or in the macromoment of a major market crash—are the extractive function of the stock market at work. As in a self-similar fractal, the same process occurs on all levels simultaneously—bubbles within bubbles. Every time we humans bid on a stock, we trigger a computerized auction among algorithmic sellers that creates a miniature bubble, just for us—forcing us to buy at the “top” of that micromoment’s trading. We humans also lose in the long run, hanging tight to our portfolios as algorithms sell on market highs, dissolve a few decades of our investing, and then buy again at the bottom, after taxpayers have “bailed out” the financial institutions pretending to have been decapitalized in the process.
These chaotic systems may exhibit emergent behaviors and even predictable patterns, but they have almost nothing to do with the underlying landscape from which they spring forth. This is why traditional business analysts, economists, and central bankers alike are at such a loss to comprehend markets in traditional terms. As former World Bank Senior Economist Herman Daly puts it: “Just as in physics, so in economics: the classical theories do not work well in regions close to limits.”40
The process of capitalization has been accelerated into something other than itself. Instead of integrating the marketplace, digital technology generates derivative systems that extract value for their operators through sheer churn. It’s synthetic growth.
Only a true digital native could understand this as a way of doing business.
INVESTMENT GAMIFIED: THE STARTUP
When investing gets so separated from real economic activity, finding funding for a company—without falling into the growth trap—is hard. Entrepreneurs must play the same abstracted game as investors but from the other side of the board.
One of the smartest technologists I know, a young woman from the West Coast I’ll call Ruby, decided to launch a company on a whim. She was not interested in making money or even promoting a new technology; she wanted to test her theories about how the ebbs and flows of the startup market worked and whether she could win at the game by getting herself acquired.
So Ruby did exhaustive research on emerging interests and keywords in the technology and business press, as well as conference topics and TED subjects. What were venture capitalists getting interested in? Moreover, what sorts of technical skills would be valuable to those industries? For instance, if she concluded that big data was in ascendance, then she would not only launch a startup related to big data but also make sure she created competencies that big data firms required, such as data visualization or factor analysis. This way, even if her company’s primary offering failed, it would still be valuable as an acquisition—for either its skills or its talent, which would be in high demand if her bet on the growing se
ctor proved correct.
She ultimately chose geolocation services as the growing field. She assembled teams to build a few apps that depended on geolocation—less because the apps themselves were so terrific (though she wouldn’t complain if one became a hit) than because of the capabilities those apps could offer to potential acquirers. Working on them also forced her team to develop marketable competencies as well as a handful of patentable solutions in a growing field with many problems to solve. The company was purchased, for a whole lot, by a much larger technology player looking to incorporate geolocation into its software and platforms. The employees, founder, and investors who believed in her are now all wealthy people.
Ruby is not cynical; she is a hacker by nature, and merely gamed a system that she knows is already a game. She reverse engineered a startup based on market conditions, industry trends, and nascent investor fads. I asked her if she could do it again—with me as a partner or investor this time. She shook her head. “I’m glad I did it, but it was kind of boring,” she replied. “Besides, we’re at the wrong moment in the cycle right now. Maybe next year. If you need funding for anything in the meantime, though, just let me know.”
Her success may be unique in that she did it as a fun experiment, but Ruby’s approach of retrofitting a company to the startup market is all too common. The smartest hackers understand that their skill at hacking technology may be less important than their skill at hacking the digital marketplace. To them, it’s all just code—and even if it’s not, it’s more like code every day. The economy is less a place to create value than a system to game. Hell, everyone in finance and banking is already gaming the system, extracting money from what used to be the simple capitalization of business ventures. Why not create business ventures that game the gamers at their own game?
Throwing Rocks at the Google Bus: How Growth Became the Enemy of Prosperity Page 21
