Trading Sites—Destroyed, Revealed, Restored

Abstract

How did companies at the Trade Center respond to the destruction brought about by the attack on September 11th? In this paper we look through the concrete and glass facade of the twin towers into the socio-technical networks of people, machines, and ideas that constituted the trading rooms. We follow the traders of an investment bank adjacent to the Trade Center in their escape away from Ground Zero to a makeshift trading room in New Jersey. We accompany them in their efforts at restoring trading operations, which revealed a socio-technical network of relations, connections, bandwidth politics, and time-critical data normally hidden from view. We support these findings with interview materials from a focus group with heads of technology of major Trade Center companies. Successful recovery, we found, was a combination of planning and spontaneity, of redundancy and self-organization, typical of firms with non-bureaucratic and non-hierarchical forms. With these findings in hand, we visit the debate on the redevelopment of Lower Manhattan, and propose for the district at large the same recipe that worked for the firms: rather than pursue top-down detailed urban planning based on the world of finance that we know today, we propose instead to emphasize lateral ties and promote organizational diversity as a basis for innovation.

Introduction

So accustomed have we grown to the image of the World Trade Center as a facade, two tall rectangles cut against the skyline of Manhattan, that we are usually unaware that it is only the outsides of the buildings that are displayed to us. In the countless media representations, rarely do we see inside the towers. And although we see photographs of the victims of the attack and learn about their personal lives, rarely do we hear about the work that was done behind that curtain wall of concrete and tinted glass.

Between the restaurant at its pinnacle and the retail shops in its basement mall, the Trade Center was above all a place of finance—not retail banking, of course, but a type of financial activity that involved trading. If, before September 11th, you were to choose a floor at random for exploration, you would more likely than not end up in a trading room, those vast open spaces where traders, salesmen and analysts buy and sell stocks and bonds. Cantor Fitzgerald, for example, the bond-trading company that suffered most in the attack, was a trading room. Morgan Stanley, the largest tenant of the towers with 23 floors, had many trading rooms.

In this paper we examine a trading room that was damaged in the September 11th attack. The trading room, part of an international investment bank, was located in the World Financial Center, directly adjacent to the World Trade Center. Two years before September 11th, we began ethnographic field research in the trading room as part of a project to study how the social organization of trading was changing in response to new information technologies. What, we wondered, is the role of place under conditions of global connectivity? Our findings, as the reader shall see, are rich in paradox: the more that timely information is available simultaneously to all market actors, the more advantage shifts from economies of information to processes of interpretation. The trading room, so abundant in information, is a place of interpretation. Our findings about the social organization of trading in the era of quantitative finance, reported below, have implications for current debates about the redevelopment of Lower Manhattan.

Since September 11th, we continued our ethnographic research as we followed the traders in their relocation to New Jersey. Our findings there confirmed our preliminary insights that traders place a high value on physical proximity to facilitate the kinds of association that are so important for their work. It also revealed that trading rests not only on social organization but also within a complex set of technical relations. The attack on September 11th damaged the trading room and completely disrupted the technologies that are so fundamental to modern trading. This disruption lays bare the socio-technical character of these relations. The breakdown of technology is society made visible.

In addition to our direct observations in the relocated trading room, we also report below on our conversations with senior managers of other financial firms in the WTC complex. How were firms able to respond so rapidly and effectively when their operations had crumbled, quite literally and so devastatingly, all around them? In passages from these conversations reported below, our interlocutors give voice to the fear and loss they experienced as they were working in those terrible days, and they tell of how strong personal ties were keys in the recovery process. As we shall see, organizational responsiveness rested less on contingency plans and hierarchical command structures than on heterarchical structures of self organization and lateral coordination. In short, the kinds of distributed intelligence that are so important in the day-to-day operation of the trading rooms formed the basis of organizational response to crisis. Our paper, thus, opens with a brief analysis of the social organization of a trading room.

Modern finance, place, and technology

The equities room of International Securities, like its counterparts at the Trade Center, offers a sharp contrast from the conventional environment of corporate America.1Gladwell (2000) discusses parallel efforts that exploit the ways in which architecture and organizational form are tightly intertwined. For example, he describes an advertising firm in California that has re-created the geography of a village in its headquarters, complete with notional streets, squares and neighborhoods. Gallison (1997, 1999) shows how the architecture of science is intimately bound with its epistemology in, for example, MIT’s Rad Lab in 1945. Whether a laboratory joins or separates the theoretical and experimental practices of physics reflects the view of science at the time and has an effect on the degree of collaboration that results. Enter the World Financial Center office complex. Take the elevator and go up one of its towers. As you exit the elevator on the 20th floor, a sea of desks with multi-colored Bloomberg screens opens up. The desks are occupied by relaxed traders clad in business casual wear. Unlike a standard corporate office with cubicles and a layout meant to emphasize differences in hierarchical status, trading rooms are open-plan surfaces where information roams freely. Instead of housing its support staff in the center of the floor, as corporations do, International Securities sits its manager at the center, where everyone can reach him. And instead of having its senior managers scattered on window offices around the exterior of the building (where the chance of bumping into them is minimized), the bank puts managers in the same desk as their teams, accessible to them with just a movement of the head or hand. Underscoring the importance of trust and sociability, the bank has limited the number of people in the room to 150 employees and has a low-monitor policy so that people can see each other.

What about the traders themselves, those privileged inhabitants of the trading room? Their outlook and personality has been radically altered by a silent technological revolution that swept over Wall Street in the last two decades. This revolution—the quantitative revolution in finance—was ignited by the rise of derivatives such as futures and options, of mathematical formulas such as Black-Scholes, of network connectivity to electronic markets such as the NASDAQ, and by high-powered computers.2To date, the leading analytic strategy by sociologists studying modern finance has been to focus on one or another of the key components of the quantitative revolution. Exemplary, in this light, is the recent paper by Bruegger and Knorr Cetina (2002) who analyze one of the key trends of the quantitative revolution, the rise of electronic markets, arguing that electronic trading has altered the relationship between market participants and physical space. MacKenzie and Millo (2001) focus on another leg of the quantitative revolution, the rise of mathematical formulae and their consequences for trading (see also MacKenzie 2002). As a result, finance is nowadays mathematical, networked, computational, and knowledge-intensive. (By January 2000, for example, the total notional amount of derivatives contracts outstanding world-wide was $108 trillion, the equivalent of nearly $18,000 for every human being on earth.) In this context, traders have evolved with the industry. Whereas the traders of the 1980s, acutely described by Tom Wolfe (1987) as Masters of the Universe, were characterized by their riches, bravado, and little regard for small investors, the quantitative traders of nowadays have MBA degrees in finance, PhDs in physics and statistics, and are more appropriately thought of as engineers. None of them wears suspenders.

The trading strategy of choice of quantitative traders is arbitrage in its different blends and styles (for a detailed treatment of valuation and arbitrage, see Beunza and Stark 2002).3The emerging field of “social studies of finance” brings researchers from the social sciences with an interest in the capital markets together with sociologists who were earlier established in the field of science and technology studies. Its classic studies include Baker 1984, Smith 1990; Abolafia 1996. More recent contributions include Uzzi 1999; Zuckerman 1999; Muniesa 2000; Lepinay and Rousseau 2000; MacKenzie and Millo 2001; Bruegger and Knorr Cetina 2002; Hagglund 2002; Muniesa 2002; Lepinay 2002; Preda 2002; Riles 2002; Scott and Barrett 2002; Zaloom 2002a, 2002b. Arbitrage produces profits by associating previously disparate markets and interpreting securities in multiple ways. For example, arbitrageurs associate the markets for the stocks of two merging companies when the merger makes their value momentarily comparable. Or they associate the stocks of two companies that are in the same index, and hence move similarly, or a stock and a bond of the same company, whose value is linked by a legal clause that makes the bond convertible into stock. The point in every case is to avoid the conventional route of valuing a company by its intrinsic value or by how hot it is with market speculators, and to choose instead a lens that produces an opportunity—a new, original valuation that differs from the value that the market assigns to a company. Thus, like a striking literary metaphor, an arbitrage trade reaches out and associates the value of a stock to some other, previously unidentified security. The two securities used for arbitrage have to be similar enough so as to hedge exposure, but different enough so that other traders have not seen the resemblance and realized the opportunity before. Each trade, then, is never exactly like the previous one. While alternative trading strategies, such as value investing or momentum trading, emphasize early access to information, arbitrage draws on novel interpretation. And whereas value trading is essentialist and momentum trading is extrinsic, arbitrage is associational.4Our theory of arbitrage (elaborated in Beunza and Stark 2002) contributes to debates in economic sociology. Economic sociology was founded through a pact with economics in which economists study value while sociologists study values; they study the economy, we study the social relations in which economies are embedded. Our work is part of a research agenda that breaks with that pact (Boltanski and Thevenot 1991; White 1981, 2001; Thevenot 2001; Stark 2000; Girard and Stark 2002; Callon and Muniesa 2002; Callon et al. 2002). To constitute economic sociology as something more than a sociology of business, its object of study should be the problem of worth. The first steps must be detailed accounts, across a range of settings, of how actors engage in such fundamental activities as calculating value and constructing equivalences. Trading provides such an analytically privileged case.

The trading room of International Securities buzzes with a variety of arbitrage styles. Each desk in the room corresponds to a different strategy, such as merger arbitrage, index arbitrage, or customer trading arbitrage. But the differences among desks are more than just operational: different desks have different principles of value, and the intimacy and repeated contact that they create leads to different social worlds. Traders at the merger arbitrage desk, for example, value companies that are being acquired in terms of the stock price of the company that is acquiring them. They specialize in asking themselves, “what is the probability that company X and Y might merge?” Analytical and calculating, for them companies are little more than potential acquirers and acquisition targets. By contrast, traders at the convertible bond arbitrage desk exploit the value of “convertibility provisions” embedded in some bonds that give the bondholder the option to convert his or her bond into stocks. To do so, they look at stocks as bonds, and focus on information about listed companies that would normally only interest bondholders. Traders at the customer sales desk, to use another example, take and give buy and sell orders to customers. Sociable and gregarious, they trade, talk on the phone and pass around Beef Jerky. The sound of their voices on the phone gives the rest of the room a window on the anxiety level of the traders’ customers and the sentiment of the market at large.

The associations established by the arbitrageurs are shaped by patterns of association in the room. Each arbitrage strategy associates securities that share a common property that makes their value comparable such as convertibility, volatility, participation in a merger, liquidity, or optionability. Since, as noted above, each desk in the room corresponds to a different strategy, interaction across desks helps traders deconstruct the value of a stock or property into its constituent aspects, or properties. A merger arbitrage trade, for example, associates two stocks that share one property—a high probability of merger—but may be affected by a different property of the stock such as high volatility, a convertibility provision, lack of liquidity, or pressure from an index. Physical closeness to other desks helps merger arbitrageurs isolate the property of interest from unwanted ones; for example, overhearing nearby traders at the convertible bond arbitrage desk may make them aware of details of those provisions. In turn, traders at the convertible bond arbitrage desk may benefit from overhearing details about the volatility of a stock from traders at the nearby options arbitrage desk.

Co-location also allows traders to synthesize the strategies performed by different desks into original, innovative trades. At International Securities, for example, a desk called “special situations” recently designed a novel “election trade” by imagining themselves being merger arbitrageurs in a case that involved a stock swap. Looking at a swap as if it were a merger gave them a distinctive perspective, the best source of profits in an industry charaterized by electronic markets and instant diffusion of information. The traders could do so because of their closeness to the merger desk. A trader is not an isolated and contemplative thinker, but engaged in cognition that is socially distributed across persons and things.5The notion of distributed cognition was developed in the work of Suchman (1987) and Hutchins (1995). Hutchins (1995) showed how the cognitive process of navigating an American warship is distributed (i.e., spread) across the members of a team, its artifacts and internal and external representations. Similarly, Suchman (1987) showed that the actions of photocopier users emerge from contextual cues provided by the machine—they are situated in the process of photocopying.

Thus, in trying to understand the modus operandi of the trading room we came to see that its locus operandi was crucially important. The more we observed, the more we could not ignore claims that electronic trading would eliminate the importance of physical location. We found that the more that trading becomes virtual, the more it heightens the salience of physical proximity—at least at the elite level. The reason for this is that the more information is simultaneously available to nearly every market actor, the more strategic advantage shift from economies of information to socio-cognitive processes of interpretation (Brown and Duguid, 20001a, 2001b, 2000a and 2000b, 1998). This particular trading room makes profits—considerably higher than industry-average profits—not by access to better or more timely information, but by producing communities of interpretation.

In addition to distributed cognition through co-location, technology is another key source of competitive advantage. International Securities, for example, invests massively in Bloomberg terminals that allow traders to represent financial value in a thousand different ways such as spread plots, bond valuation models, or active spreadsheet links. High-bandwidth connections to the market give traders a crucial temporal edge over retail investors by providing them with price data almost in real-time. A computer platform (called the “trading engine”) automates all the clerical operations related to trading such as registering trades, breaking them into small pieces to avoid detection by rivals, etc. And numerous traders use computer systems (called “trading robots”) to automate the buy-and-sell process according to a logic codified in an algorithm.

However, mindless engineering alone does not give International Securities its edge over rivals. The key lies in an interaction between technology and humans and ideas, a socio-technical network constituted by all these three elements (Latour 1991; Callon 1998). Trading robots are a good example of that interaction. A robot is system made up of connections, algorithms, and computer hardware that receives market data and sends trading orders according to some theoretical principle of finance such as “mean reversion.” But there is a lot that is social among those cables, chips, and lines of code. In the development of the robot, for example, the algorithm is programmed collaboratively by computer programers and traders in a special meeting room designed for rapid informal collaboration (“the whiteboard”). The robot is monitored by a human trader, a so-called statistical arbitrage trader, whose job is to stop the algorithm whenever the market situation is no longer consistent with the theory that inspired the code. For example, when two companies merge, the principle of mean reversion no longer applies and the robot, if not turned off, would perform money-losing trades. To supervise the robot, the statistical arbitrage trader makes use of humans in the rest of the room. For example, the trader obtains crucial hints about which companies are about to merge by overhearing conversations at the nearby merger arbitrage desk. Similarly, the human monitor of the robot uses the room to find out whether the data arriving to the robot is delayed (and therefore a dangerous misrepresentation of real prices). This is done by paying careful attention to expletives or panic among the computer technicians that sit close by instead of relying exclusively on the dials and speedometers built into the robot. If the statistical arbitrageur hears expletives, it means that there are technical problems, even if the computer dials say “fine.”

The trading room of International Securities, as much as those trading rooms at the Trade Center that disappeared with the attack, assembled together an original set of social, spatial, and technical elements that need to be understood to appreciate what “finance” meant in Lower Manhattan. In this elite world of finance, social relations matter: the interpretive process that took place in it drew on non-hierarchical social relationships, trust, and lateral ties. Space also mattered: the room and its desk-based spatial configuration promoted communication and distributed cognition across teams. And technology mattered too. The room relied on highly automated trading technologies such as trading robots and trading engines, and these technologies took advantage of the constant communication across traders afforded by the co-location in space.

A desk on the 20th floor

On September 11th, a deafening explosion interrupted the work of the arbitrageurs at International Securities. As they rushed to the windows on the east side of their trading room they saw the adjacent building, the Trade Center, go up in flames as the first terrorist plane hit Tower One. The second plane crash brought terror and a tumultuous escape to the Hudson River. By the time the towers fell, the traders were already on the ferry to New Jersey. Fortunately, none of the employees at International Securities was harmed. The building, however, was badly damaged, making the trading room dangerous and inaccessible. The Trade Center had collapsed at its doorstep. Its windows were shattered with the explosion and pierced by debris from the fallen towers. Dust, ash and dirt, possibly containing asbestos and toxic chemicals, entered the room and penetrated the computers, clogging their fans, overheating them and rendering them unusable and unsafe for repair. The data they contained was lost. The building was deemed structurally unsafe, and access to it was prohibited for months. As a result, the lively trading room that had once supported the innovative work of interpretation became a dark hole with no electricity, no connectivity and no assurance of safety from toxic chemicals.

On the night of Sept 11th the management team of the equities trading room at International Securities regrouped in New Jersey and estimated that it would take them from three weeks to three months to be trading again. The bank had only one equities trading room in the US and there was no backup site to which they could go. The bank did have another available facility, a back-office in a suburb of New Jersey, but the only resource that the traders could count on was spare space in a basement where the firm stored corporate-style minicomputers for processing payroll data. The basement had no computers, no desks and no connectivity.

Yet, six days later, as soon as the New York Exchange re-opened on September 17th, the traders at International Securities were trading again. We were privileged to witness how this was accomplished. Several days after the attack, we sent an email of concern to ask if everyone had escaped unharmed. To our relief, we learned that no one was injured. To our surprise, the return email included an invitation, indeed, an insistence, that we come over to New Jersey to witness the recovery process. “It is chaotic,” wrote the manager of the trading room, “but also very inspiring.” Our presence would be “a reminder of normal times.” As ethnographers, we felt enormously honored to be welcomed to document these extraordinary efforts.

Thus, on September 19th we were back among traders in our role as observers, this time in an improvised trading room in a converted basement warehouse of New Jersey. The temporary trading room was barely an hour’s drive away from Manhattan, but it felt a universe away from the excitement and activity of Wall Street. Located in a suburban corporate park, the building was surrounded by similar low-rise corporate offices, used by manufacturing companies such as Colgate or AT&T. Just around the corner, a farm announced “Hay For Sale.” The surroundings offered an endless succession of down-market shopping malls, Wal-Marts and Dunkin Donuts; one could drive around for an hour and never be able to find espresso coffee. What had been the back office of International Securities had, in effect, become its front office too. Our traders were Wall Street traders… in New Jersey.

The trading room was located in the basement of the building. To reach it we had to pass several rows of corporate-style cubicles and beige carpet; after the cubicles, we reached the trading room—perhaps the most unexpected sight in such an environment. A huge open-plan space, complete with traders, desks, computers, outsized TV screens, and multi-time zone clocks. The room had a makeshift feel to it: no windows, a low ceiling and walls painted in industrial yellow, more fitting for a storage room than a trading room. Indeed, one week before our visit the place was still being used to store the mainframes and tape machines used by the bank’s data center. The floor-level air-conditioning ducts used to cool the machines were still working on September 19th, chilling our legs from the shoes up. Inside the room, workers in the technology department constantly moved up and down among spare cables, keyboards, and mouses interspersed with empty cans of diet Pepsi and Mug root beer.

Our traders were not just makeshift arbitrageurs—they were survivors. “I don’t have to tell you how close we were,” one of them told us. “You’ve been there. You know it.” A huge American flag hung in the middle of a wall, and dozens of small ones colored the top of many traders’ screens like flowers in a green field. Of the three home cinema-sized TVs (typically used in the Financial Center trading room to get market news), one was switched from CNBC to CNN for news of the impending war in Afghanistan. The dress code had shifted from business casual to jeans and boots. The room was noisy, but the sound, as one trader put it, was “a wonderful sound of life.”

Our traders were in New Jersey, unquestionably in a basement storage room in New Jersey. But a sign taped prominently on the wall gave different bearings: “20th Floor, Equities.” In other parts of the same enormous room one could read other signs: “21st Floor, Fixed Income” and “19th Floor, Risk Management.” Our traders were still between the 19th and the 21st floors, but now horizontally rather than vertically. Moreover, within the constraints of those temporary quarters, they had arranged their desks to replicate the layout of the Financial Center trading room. For example, every trader in the “agency trading” desk remained together, sitting on the same desk. In the Financial Center trading room they sat on a spacious desk between the “stock-loan” and the “special situations” desk. In New Jersey, they camped on a table partly occupied by two photocopiers and three fax machines, in what used to be the fax station of the data center. But they stayed together. The desks also preserved their relative locations, reconstructing the cognitive order of the trading room at the Financial Center. When the managers of the agency and special situations desks found themselves sitting again in front of each other, they reverted to their old routine of checking perceptions against each other, probing into each other’s beliefs, and designing together new arbitrage trades. At some point, one of them exclaimed in exhaustion, “Everybody seems to be thinking with my brain today!” a reflection that the distributed cognition afforded by the desk pattern was again taking place.

The traders could replicate the floorplan of the Financial Center trading room, but not the technology. Direct data from the New York Stock Exchange was not available. “Trade Manager v1.4a,” the platform of hardware and software that registered and processed trades (also called the “trading engine”), was not working. The customary phone turrets with twenty lines each were also not available, and the traders had to make do with off-the-rack single-line phones (which they slammed with the usual energy). Instead of Sun workstations, they were working on Pentium IIs and laptops, some brought from the traders’ homes, some rescued from the data center, some hurriedly purchased in the days following the attack. Instead of having virtually unlimited bandwidth they now had to adapt to limited network connections that did not allow all desks in the room to trade simultaneously.

The trader’s response to September 11th contains important insights for a socio-technical view of organizations. In the sections above we have argued that arbitrageurs associate stocks by associating people, artifacts, and ideas in the same place. Conceptually, it is tempting to split this socio-technical network into humans and machines—people who think and talk vs. machines that obey pre-programmed commands. But such separation is misconceived. “Technology,” writes Bruno Latour (1991, p. 1), “is society made durable.” Yet, what happens when technology breaks down, when traders who were accustomed to twenty dedicated phone lines apiece must share phones, when traders whose style of trading is based on speed and volume must suddenly operate with minimal bandwidth? The breakdown of the trading technology at International Securities opened up for us a window on its socio-technical network – a network that operated seamlessly and invisibly in the Financial Center trading room. The breakdown of technology is society made visible.

The breakdown of technology revealed the ways in which people and artifacts are inextricably linked. For example, in describing the process whereby the bank established a connection to the NYSE, we noted that the head of technology at International Securities used “connection” and “relationship” interchangeably. On some occasions he would refer to “Mike,” and on some others to “the ISDN connection,” yet mean the same thing. The first attempt to connect was through electronic communication networks (ECNs) but the connection kept dropping every minute, which proved very problematic for the traders because they could not know their exposure. In the end, the bank only managed to connect to the NYSE through an ECN that brought their technicians to the trading room. And, in turn, the only reason the ECN invested its resources (technician, etc.) in this manner was that it had an on-going relationship with International Securities and was interested in having the bank trading through its system and providing volume and liquidity. Hence the tight link between social and technical ties: as the head of technology explained, “Once we establish a relationship with someone, it’s very easy to move on” to a connection. Companies with wide social networks, this implies, should recover more easily from problems with their technology.

As society made durable, the technology of International Securities also reflected the regulatory environment in which it was developed. In the process of re-connecting the New Jersey trading room to the NYSE, our traders experienced great difficulty in finding appropriate modems for their machines. The reason, it turned out, was that in the past regulatory requirements limited banks to slow 9.6 K baud rate connections to the NYSE in order to prevent speed races. Technology is also regulation made durable. Without modems specially configured in that manner, the traders in New Jersey were not be able to send and receive data to and from the NYSE. But by September 2001, modems old enough to crawl at 9.6 K baud could not be obtained through commercial channels. In order to be able to trade, the head of technology explained to us, he tried to rescue them from the Financial Center:

The modems were in the old Unix computers, and we could not find new modems for our computers. So I had to go back to World Financial Center to strip the computers, walking up twenty floors in a chemical suit and with a torch light, as there was no electricity.

A socio-technical network is far more complex than the simple sum of the social and technical ties in the organization. The severance of technical ties, for example, cannot automatically be fixed by new social ones. This became clear in the sign “20th Floor, Equities” placed on the wall, and its insistence in reproducing the old floor structure embodied in it. The sign not only reminded traders that the equities trading room was between risk management and fixed income, but it also led employees back to their jobs as traders: by reconfiguring the socio-technical network that had disappeared, it reduced the fundamental uncertainty that the traders faced. According to Callon, a socio-technical network

is not connecting identities which are already there, but a network that configures ontologies. The agents, their dimensions and what they are and do, all depend on the morphology of the relations in which they are involved (Callon, 1998, p. 15).

After the attack, traders were left wondering whether their firm would continue to exist, whether the trading room would operate again, what they should do, and even what they were. The basement turned those survivors back into traders. To the question of, “who am I?” the computers, desks, and open-plan spaced answered “a trader.” To the question of, “what should I do?” the “20th floor” sign answered: the same as in the Financial Center trading room.

The ontological character of the socio-technical network was also manifest in the discourse of the company’s traders. We found them engaged in a seemingly philosophical debate about the meaning of “real data.” The problem that they faced was that the proprietary direct data connections that linked International Securities with the NYSE ran directly to the Financial Center, and therefore could not be used in New Jersey. The traders had to rely instead on data from Bloomberg L.P. But, the traders complained, “Bloomberg data is not real data.” It had small, unannounced delays, which made it unsuitable for some trading strategies such as index arbitrage. If data cannot be made part of the network, the traders seemed to be arguing, it does not even count as real. The network, as Callon argues, configures the ontology of the actants.

The socio-technical network of people, machines and ideas at the Financial Center was restored, but not in its pristine form. Whereas the elements that remained turned the survivors back into traders, those that changed redefined for many the meaning of “trader.” In the agency trading desk, for example, the destruction of the trading engine turned junior traders into clerks. The operations that were previously automated by the engine such as booking trades, registering them, breaking them up, etc., had to be done manually, effectively taking the bank to the trading technology that it used five years before. Junior traders stood up behind those lucky senior traders who had a seat and a computer, ready to help. When, in the middle of a phone conversation, one such trader suddenly needed to record a transaction, at the shout of “gimme a ticket, somebody gimme a ticket!” three junior traders scrambled to offer tickets, paper and whatever else he might need. Another junior was told to “help with the tickets” and “relieve others”—but told with a sensitivity to the situation characteristic of International Securities: following the indications, the senior trader who gave them added “Oh, and this isn’t permanent, by the way.” So unusual was manual bookkeeping for junior traders —so radical the reconfiguration of the socio-technical network that it required —that some of them did not even know how to do it. “Can we do manual baskets?” a young trader asked a more senior one. “Of course you can!” went the answer.

The lack of direct data from the NYSE transformed the traders in the statistical arbitrage desk from monitor of the trading robots into active participants in the price mechanism. “Welcome to cut and paste land,” one stat arb said to us by way of greeting as we approached his makeshift desk in New Jersey. By “cut and paste” he referred dismissively to the operation he was constantly performing, transporting orders from the e-mail system to the trading engine by force of pointing and clicking his mouse. He labored in this fashion because the lack of price feed in the Unix system forced him to connect one interface to the other. As a result, he said, “I have very little time left to do anything else” such as monitoring the market and the speed of the price feeds, his typical jobs. The breakdown of technology had transformed these stat arbs from monitors of the trading robots into manual operators.

Whereas technical ties can make social ties more durable, breaks in technical connectivity also create strains in social ties. At International Securities, insufficient connectivity gave rise to bandwidth politics. This was exacerbated by the limited speed of the “stop-gap” machines the traders had, off the rack PCs instead of specialized Sun workstations. As a consequence, for eight weeks the bank traded only one third of its regular volume before September 11th. Furthermore, the reduction in volume was not the same across the board; the customer trading desk, for example, rapidly went almost back to its normal level of volume, but the (bandwidth sensitive) index arbitrage desk remained extremely hampered. These problems led to conflicts between statistical arbitrage and index arbitrage traders over who could connect, and ultimately over the profitability of their respective “books.”

High touch, low tech

Following our observations at International Securities, we felt that we had to gain a better understanding of how socio-technical networks that had taken an enormous blow and been badly damaged could be, if not healed, at least restored. We began to talk to numerous Trade Center firms that were directly affected by the attack about response and recovery. We made extra efforts to speak with the people responsible for technology and people responsible for contingency planning, preparedness and continuity management. We spoke with managers in large companies as well as in small- and medium-sized firms. As part of this effort, Columbia’s Center on Organizational Innovation, joint with Columbia’s Interactive Design Lab, held a roundtable discussion on December 5th with senior information technology and communications executives from key Trade Center firms as well as major consulting and technology firms. The companies included Merrill Lynch, Cantor Fitzgerald, Deustche Bank, Sun Microsystems, Guy Carpenter, Accenture and Fred Alger Associates. We told them we would only report here their comments, without attributing them to specific persons or companies.6More information on the roundtable with Trade Center companies affected by September 11th, is available at: http://www.coi.columbia.edu/pdf/infrastructure_interface_program.pdf

What did they tell us? No one said, “David, technology saved us.” Or, “Daniel, our plan really worked.” Despite being technology officers, they all pointed to the human element in organizations. Of course, they did talk about contingency plans and about technology. They told us, for example, that it mattered that the Trade Center had been bombed once before in 1993 and that their planning and preparation subsequently made a difference, or that extra back-ups in preparation for Y2K had proven for them to be really important in recovery. But sound planning was not sufficient in dealing with the uncertainties created by such a disaster. According to three executives,

Well, yes, we could not have done without the corporate technologies that we had in place. But what surprised us, in the initial hours and days after the attack, was how important were the technologies that the company hadn’t invested in.

A business plan is one thing but you need a people plan, and everybody needs a responsibility.

Without that human element of commitment to the task, commitment to each other, preparedness wouldn’t have done anything. The best plan would have never opened up.

Our analytic findings are straightforward. Recovery was a combination of planning and spontaneity, of redundancy and self-organization. To give texture —organizational and human texture—to these abstractions, we present some accounts of recovery in the words of our informants. The first, appropriately, is a story about stories from an executive at a major bond trading firm in the Trade Center that suffered terrible casualties. On the evening of Sept 11th, the survivors of the leadership group met, knowing that they had to be trading when the bond markets opened in the same week. The firm had followed all the guidelines for contingency planning. They had backed up their data – at not just one but, in fact, two off-site locations, one across the river, one across the Atlantic. But they could not access the system. As the executive recounts:

We had 47 hours to get [ready for] September 13th, when the bond markets reopened and there was one situation that our technology department had that they spent more time on than anything else…. It was getting into the systems, [figuring out] the IDs of the systems because so many people had died and the people that knew how to get into those systems and who knew the backup…. and the second emergency guy were all gone. How did they get into those systems? They sat around the group, they [technology officers] talked about where they went on vacation, what their kids’ names were, what their wives’ names were, what their dogs’ names were, you know, every imaginable thing about their personal life. And the fact that we knew things about their personal life to break into those IDs and into the systems to be able to get the technology up and running before the bond market opened, I think [that] is probably the number one connection between technology, communication, and people.

For this organization, as for many after September 11th, it was personal knowledge that made the difference. The organizational forms that responded well were those in which employees shared non-professional knowledge – not knowledge about careers or accomplishments, but the kind of knowledge that could only come about by close contact in relationships of trust. Without this knowledge, as we saw in the case of the system passwords, you had no technology and you had no information. 7For a more elaborated discussion about how response and recovery revealed that the interface between humans and data is socio-technical, see Kelly and Stark 2002. This, in turn, reveals the organizational form that supports such knowledge. Employees knew the names of their colleagues’ spouses, or where they went on vacation, or their favorite movies or music—but not because the organization had formally inquired and entered this information into some central database. The key in this case was how well they knew each other personally—details of private life that are, in the strict sense, irrelevant to their status as co-workers. With so much attention on how to create resilient, self-healing technological networks, here was compelling testimony to the resilience and self-healing capacity of social networks. We heard similar things from many representatives of directly affected firms—about the importance during crisis of personal commitment, feeling like a family, individual empowerment, and “lateral teams.”

The importance of the human side of socio-technical networks in crises was echoed in many companies. “Without that human element of commitment to task, commitment to each other,” one executive noted, “preparedness wouldn’t have done anything. The best plan never would have been opened up.” This is illustrated in detail in the following quotation, from the head of communications of an investment bank:

You realized that the buildings had gone down. There’s a moment where you really do believe that you are the only person left in this company alive and right from the beginning I think it was more instinctual than it was ever organizational. Within an hour and a half after the first plane hit, the four remaining members of my team (I had ten at the Trade Center) were at my front door. They had come from downtown, from wherever they were, whether it was in a subway on the platform, in the concourse, and they showed up at my door for no reason other than we had to do something.

What to do? Just as the traders at International Securities reverted to trading, these engineers reverted to communications:

We had to do what we knew how to do, which was communicate, and I think that’s the core of this whole tragedy is that the people in this sort of environment, you always go back to what you know best, and we were a communications team… within three hours we had a call center set up.

Personal contact was critical to recovery because firms, it is important to recall, were managing people who were in fear and grief. In these circumstances, “what made the difference, for every company that came back successfully, one executive observed, “[was] that kind of touch, high-touch, low-tech solution.” The following illustrates the dramatic context in which such a solution is required:

This was not a fire in a building which just destroyed 2 floors… Most everybody lost people they knew. They were traumatized, there was fear of war. Nobody knew if the next day there was going to be more. I had a guy walking around with a picture of his wife and kids in his pocket and he was looking at it every two minutes because he was afraid he was never going to get home again.

The need for high-touch, low-tech was shown in the case of a company that found itself relying too much on the efficiency of a website, to the detriment of personal communication:

With regard to the family and the emergency communications, we really relied on a crisis communications website that we launched on September 12. It began really as a list of safe and accounted-for people with call center phone numbers to check with us, contact information, and then sadly and quickly became a place to distribute memorial service information, benefits information, aid for the families and friends of the victims.

The company was at first forced to rely on the website by the circumstances of the attack. As the executive painfully recalled:

Without an HR department—which we lost our entire HR department— (…) the website became our only ongoing method of communications and as such was so important to us. In terms of design and interactivity we kept it deliberately simple.

But the company eventually realized it had relied on the website too much:

One thing that I do want to say though is that in hindsight, that we over-relied on that website because I think the most important and crucial part of the tragedy is not to forget the benefits of face-to-face communication. You could have told these people go to the website a million times and we had so much information on the the website… [But] you’re grieving, you’re not going to read the information on the web site, and you know we could have posted stuff there every second and it wouldn’t have mattered.

The importance of social ties and personal knowledge was also manifested, in the first days after the attack, in the problem of locating people. Personal knowledge was key because corporate means of communication had broken down. In the first moments after the attack when the phone lines were swamped, survivors would try to locate the others by e-mail sometimes, as our informants narrated from Internet cafes somewhere in Manhattan. But frequently corporate email was not working. In this context, do you know your colleague’s personal email account, for example, an AOL or Yahoo account? When you obviously could not reach someone at their office phone, do you know the home phone? If not, do you have the beeper, pager, Blackberry, Palm PINs? None of these? Do you know how to get to the home of your supervisor?

Similarly, porous organizational boundaries that led to strong social ties with clients, vendors, and consultants also proved crucial. According to another Trade Center executive,

Vendors and suppliers in our information technology areas, in communications and almost across the board really were absolutely outstanding. It’s very easy to criticize these people routinely. They’re the brunt of bad jokes. It’s sort of corporate yucks to go around and make fun of the infrastructure and who supplies it. But in this case it was exceedingly generous. I can’t begin to tell how much we could count on the relationships we had with vendors, consultants, and clients. People were willing to do whatever they had to do to reconnect to us and whether that meant working around the clock so that we could be open on the 14th, they were there. You know, those relationships can never be replaced with anything.

We see the response to the crisis as an instance of innovation, as a deviation from established routines. What mattered were lateral ties that criss-crossed the vertical ties of corporate hierarchy and self-organization of the teams. The tragedy underscores the adaptive advantages, under conditions of uncertainty, of non-bureaucratic organizations. Non-hierarchical organizations, or “heterarchies” (Stark 1999; Girard and Stark 2002; Grabher 2002), are characterized by relationships of interdependence, by lateral and distributed authority, and an internal diversity of organizational forms. Like the trading room of International Securities of the previous section, heterarchical networks exhibit qualities of emergent self-organization, strong lateral ties, and a diverse distribution of knowledge and empowerment at lower levels of the organization.

The importance of heterarchical organization is relevant in the light of the ongoing discussion over the importance of “preparedness” against disasters. As one executive said,

The most forward-thinking and strategic of the things that we have learned is that it’s now our view that going forward companies are going to be valued on something that I would call preparedness. I think that that is going to become an integral part of how investors, employees, fiduciaries, everybody, counterparties, everyone looks at a companies’ worthiness. It will become analogous and perhaps even part of your credit rating, and these preparedness issues come down to some of the most mundane things… like mail processing, air travel, workplace security, data security, and data infrastructure, personal identification and personal accountability, and then even things like new accounts and sources of funding.

The traditional view on preparedness is based on contingency planning and what we call replicative redundancy, that is, having more than one of each thing in case the artifact in use breaks down—more than one phone line, one computer system, one trading room, etc. (Kelly and Stark, 2002). These are important; for example, having a redundant trading room, as some investment banks did as a result of contingency plans following the 1993 Trade Center attack, became key to guarantee continuity of business operations. But we would like to question the traditional view that there is a trade-off between preparedness and competitiveness, or that the need for preparedness will give an advantage to companies that are larger and more bureaucratic. The pressure to have redundant systems and facilities betrays an almost-exclusive attention to creating resilient, self-healing technologies that overlooks the human side of socio-technical nature of the networks. Worse, it does not take into account the resilience and self-healing capacity of the human ties in these networks, as manifested in the compelling testimonies of the Trade Center companies above. The commitment to the team, imaginative use of new technologies and creative responses exhibited by the companies show a different, generative redundancy that allows ties across socio-technical networks to be regenerated in a broader set of circumstances than those envisioned in a narrow recovery plan. As one executive put it,

I’ll stress that leaders lead in times of crisis. We had some excellent examples of very focused decision making. People from all parts of the organization really did think outside of the box. By ensuring people had the right focus, we were able to achieve some sort of miracle. We weren’t able to do this in our traditional modes of thinking and the last thing I think I guess I really wanted to stress is that if you empower people to think outside the box, you give them the authority to solve a problem, they will solve it. I can’t stress that enough. We came up with this whole new sort of terminology of crisis achievement. You know what we did in those first two weeks was absolutely outstanding. I just wish we could repeat it going forward.

If the factors that explain a successful response are similar to those that explain innovativeness, perhaps the metrics that explain preparedness could look very similar to the metrics for innovation.

Dispersion

In April 2002, the traders from International Securities returned to the World Financial Center. Other companies such as Merrill Lynch, Commerzebank and American Express have also returned, contributing to the future of Lower Manhattan as a business district. But not all are returning. Officials at the city and state levels have put together a program of economic incentives such as cash grants and tax breaks to keep companies in Lower Manhattan, but already an exodus seems to be in place, with companies such as Lehman Brothers, Aon, Pillsbury Winthrop, Dresdner Kleinwort Wasserstein, and ABN Amro leaving the area to more expensive locations in Midtown, or more distant offices in Brooklyn and Jersey City.8Charles V. Bagli, “Downtown, An Exodus That Cash Can’t Stop,” The New York Times, July 24, 2002.

What about the companies at the Trade Center? The one approach to preparedness on which Trade Center executives at the Roundtable most clearly coincided was the need to disperse operations geographically. “Are we going to have a single operation, a single site in New York City?” asked one executive. “The answer clearly is no, we’re not going to…” was his reply. Another executive explicitly linked preparedness and dispersion:

So when I think about measuring preparedness, one of the things that… I heard loud and clear this morning again is an organization’s ability to operate geographically dispersed, effectively, and those are two very distinct concepts that have to mutually exist.

Our Trade Center executives believe in dispersion for different reasons. The obvious reason is to spread important operations in different places, as one executive suggests:

Number one, we have decidedly rethought our strategy of having all mission-critical applications and functions, whether they be electronic or human, in one location, and that just is not restricted to buildings but geography.

But other companies discovered that dispersion brings in additional advantages, such as increased adaptiveness:

We have a number of organizations within our agency. The organizations that had a culture of dispersed employees, dispersed human capital if you will, functioned a lot better during the emergency than those that were traditionally organized along the office structure.

For two other executives the culture of dispersion is not just an advantage, but a challenge of corporate change:

There’s a lot of managers who are not yet in line with the idea of having a group, a team, your partner or whatever you want to call the organizational unit that’s in six different locations.

How do you disperse intellectual capital [and] …have a management team that can deal with the realities of trying to work in that environment and still be as productive? I think that’s a key issue.

In addition to a culture that promotes change, our executives were also interested in the technological transformation that dispersion will demand. According to one executive, “the push now [is] for video collaborative interactive tool sets. To be able to really take the model and decentralize it is a very big push for us.”

Trade Center, trading rooms

Given these trends in corporate dispersion, Lower Manhattan cannot count on getting back the full complement of companies that filled its Trade Center before the terrorist attack. For several decades the district had already been losing a competition against Midtown Manhattan as the location of choice for financial companies, and forced relocation has accelerated that trend. The debate about the redevelopment of the World Trade Center site must consider these historic trends as well as take into account the rise of electronic trading, a force that, according to some, removes the needs for a district in finance. Our research on trading prior to September 11th and on the dynamics of recovery afterwards bears directly on this debate.

During the last decades of the 20th Century, Wall Street has gone through a veritable quantitative revolution based on three legs: high-speed network connectivity, high-powered computation, and the development of mathematical finance. Along important dimensions, quantitative finance reduces the salience of physical proximity. The NASDAQ, for example, has long operated as a virtual market. As an electronic exchange, it is far from an isolated example since the inexorable trend in cities around the globe is from physical to electronic exchanges.9Frankfurt’s DTB, Paris’ MATIF (both merged into Eurex in the year 2000), London’s LIFFE, and Stockholm’s and Madrid’s exchanges have already migrated to an electronic form. Other markets such as the recent ECN-turned-exchange Island Futures Exchange LLC, began in electronic form. And others, most significantly the Chicago Mercantile Exchange (CME), have developed a dual physical-electronic system by keeping its pits and developing an electronic system, Globex, that complements rather than threatens it (Milo, 2001; Scott and Barrett 2002; Zaloom 2002; Muniesa 2000). Many of the largest hedge funds are located in Connecticutt, and the largest mutual funds such as Fidelity, Janus or Vanguard are as distant from Lower Manhattan as their respective locations, Boston, Denver and Philadelphia.10Proximity does seem to matter a great deal for venture capital firms, but for them it is proximity to their new start-up firms and to knowledge-generating sources such as universities, resulting in concentrations in Silicon Valley, Boston, Austin, and so on. See Powell and Owen-Smith (1998). According to The Economist,

Lower Manhattan… may be the world’s largest single electronic marketplace. In the days when bank’s vaults were full of bearer bonds and stock certificates transferred by “runners” after trades were done, trading firms had good reason to cluster together. Yet proximity is little or no help in implementing trades. Anonymous, “The Markets Rewired,” The Economist, Sept. 22nd, 2001, p. 68-69.

But some aspects of quantitative finance heighten the salience of proximity. As electronic markets make hard information instantly available to everyone, knowledge in soft or more tacit forms of interpretations, impressions, and perceptions of others become the key source of competitive advantage. As we saw at International Securities, proximity of traders to each other in the room was key in creating the interaction across desks that lead to the most original and profitable trades. And the deliberate way in which they attempted to reconstruct the layout of the trading room in New Jersey revealed how acutely they were aware of these dynamics. Similarly, proximity to other financial firms was crucial for some arbitrage strategies such as merger and convertible bond arbitrage. In the case of merger arbitrage, for example, traders bet on the likelihood of a merger. As part of their strategy, they must determine the commitment of two firms to merge, and to do so they find it crucial to attend companies’ presentations. As the traders told us, it’s not enough just to hear the meeting webcast on the Internet—one needs to be there, to see the faces in the audience or around the table as firms make their bold claims, to bump into ex-colleagues in the corridors, and to have lunch with the people involved. To do so, it is important to be close to other financial firms. In itself, however, this does not imply being in Lower Manhattan: if enough firms leave the district for, say, Midtown, then after a while it will be in every firm’s incentive to follow suit.

While following discussions about the future of finance, it is clear that many academics, policy-makers, and even people in business identify Wall Street with the New York Stock Exchange (NYSE). This identification would have been correct for most of the 20th Century: trading rooms began as extensions of the NYSE that investment banks built inside their corporate skyscrapers in order to carve out and better process the information they obtained from it. Nowadays, however, equating Wall Street with the NYSE amounts to the same superficial approach that equates the Trade Center with its facade and never goes inside. Our research indicates that the real locus of modern finance is not the Exchange but the trading rooms. As a result, we should abandon visions of finance in Lower Manhattan as having a radial or mono-centric urban form—the NYSE surrounded by trading rooms—and embrace instead a multi-centric understanding of Wall Street. The Trade Center was not some sort of back-office to the NYSE. Wall Street is better thought of as a web of trading rooms in which each node is anchored to the area by its proximity to others, rather than to the Exchange. And instead of being miniature replicas of the NYSE, they are more like scientific labs.

How then to keep finance in Manhattan? One obvious answer is to replicate what was in place before the destruction of September 11th. Build new towers (two or twenty) with exactly the same square footage of office space and hope that the financial firms will return. To adopt this replicative strategy would ignore that we are rebuilding not for the next months or even years but for the coming decades and would neglect the important changes that have taken place within the sector. The opposite strategy is to anticipate what finance is going to become twenty or thirty years from now, and attempt to design accordingly. The problem with this latter approach becomes apparent when we reflect on the revolution in quantitative finance that has swept the industry. Thirty years ago, almost exactly when the Trade Center was being built as a “vertical port,” no one anticipated that it would become a center of financial trading. And no one is likely to have predicted that finance would go through the tripled features of the quantitative revolution. The Black-Scholes formula for pricing derivatives (one of the key applications of mathematics to finance) was developed in 1973 and was hardly on the radar screen of policy-makers. In 1973, computers were those things the size of a room that were used to process payrolls. And to speak about the “World Wide Web” and “high bandwidth Internet connectivity” in 1973 might have provoked suspicion that one’s connection to reality had been clouded by too many highs on recreational drugs. In short, the quantitative revolution in finance would have been difficult to anticipate and even more difficult to design for. With this retrospection as a cautionary note, who can say with confidence what finance will be thirty years from now?

In place of predicting the future or of replicating the recent past, the citizens of New York City should encourage their representatives to rebuild Lower Manhattan with an emphasis on increasing diversity of types of organizations – not simply more large corporations but medium size and start-up firms, not simply in financial services but a broader sectoral range, not simply businesses but educational and cultural institutions. Diversity accomplishes two tasks. First, it would make Lower Manhattan a more vibrant and exciting locale, and thereby more attractive to the knowledge-intensive firms that will be a source of economic vitality for the city. Why do energetic, ambitious, young people come to New York City? Because other young, ambitious, energetic people like them come here too. The more a city, or a district in a city, is a place of wonder and excitement, the more it can stimulate these tipping point effects. Second, greater diversity among the types of organizations produces a broader “gene pool” out of which innovative recombinations can emerge. Will such a strategy come from the financial sector itself? Will it come from a city administration lead by a businessman whose Bloomberg terminals populate the trading rooms? Like the generative strategies—with their lateral, heterarchical ties—that proved so resilient and effective in the first days and weeks of recovery, a generative strategy for rebuilding Lower Manhattan will require broader, horizontal ties actively involving citizens and civic associations. Like the trading rooms themselves, the new associations that will make for innovation in redevelopment will require distributed intelligence and the organization of diversity.

Daniel Beunza is a PhD candidate at NYU’s Stern School of Business and a Research Associate at Columbia’s Center on Organizational Innovation. David Stark is Arnold A. Saltzman Professor of Sociology and International Affairs at Columbia University and an External Faculty member at the Santa Fe Institute. He is currently a Visiting Scholar at the Russell Sage Foundation in New York City.

References

Abolafia, Mitchell Y. 1996. Making Markets: Opportunism and Restraint on Wall Street. Cambridge: Harvard University Press.

Brown, John Seely and Paul Duguid. 1998. “Organizing Knowledge,” California Management Review Spring, 401(3): 90-111.

—-. 2000. The Social Life of Information. Boston, Mass: Harvard Business School Press.

—-. 2001. “Knowledge and Organization: A Social-Practice Perspective.” Organization Science, 12(2) 198-213.

Bruegger, Urs and Karin Knorr Cetina. 2002. “Global Microstructures: The Interactional Order of Financial Markets.” American Journal of Sociology, March.

Callon, Michel. 1998. “Introduction: the embeddedness of economic markets in economics.” In The Laws of the Markets edited by Michel Callon. Oxford: Blackwell Publishers.

Callon, Michel and Fabian Muniesa. 2002. “Economic markets as calculative and calculated collective devices.” Manuscript. Centre de Sociologie de l’Innovation, Ecole des Mines de Paris.

Castells, Manuel. 1996. The Rise of the Network Society. Cambridge, Mass: Blackwell Publishers.

Galison, Peter L. 1997. Image and Logic: A Material Culture of Microphysics. Chicago: University of Chicago Press.

Galison, Peter L. and Emily Thompson, ed. 1999. The Architecture of Science. Cambridge, Mass.: MIT Press.

Girard, Monique and David Stark. 2002. “Distributing Intelligence and Organizing Diversity in New Media Projects.” Forthcoming in Environment and Planning A.

Gladwell, Malcolm. 2000. “Designs for Working.” The New Yorker (December 11): 60-70.

Grabher, Gernot. ed. 2002. Special Issue: Productions in Projects: Economic Geographies of Temporary Collaborations. Regional Studies 36(3) (May).

Hagglund, Peter, 2002. “Making a firm investment object. The importance of good connections in the stock market.” Paper presented at the New York Conference on Social Studies of Finance, Columbia University.

Heath, Christian, Marina Jirotka, Paul Luff and Jon Hindmarsh (1995). “Unpacking Collaboration: the Interactional Organization of Trading in a City Dealing Room.” Computer Supported Cooperative Work 3 (1): 147-165.

Hutchins, Edwin. 1995. Cognition in the Wild. Cambridge, Mass: MIT Press.

Hutchins, Edwin and Tove Klausen. 1991. “Distributed Cognition in an Airline Cockpit.” in Y. Engestrom and D. Middleton, Distributed Cognition and Communication at Work. Cambridge: Cambridge University Press.

Kelly, John and David Stark, 2002. “Crisis, Recovery, Innovation: Learning from 9/11.” Forthcoming, Environment and Planning A.

Latour, Bruno. 1991. “Technology is Society Made Durable.” Pp. 103-131 in A Sociology of Monsters: Essays on Power, Technology, and Domination, edited by John Law. London, Boston: Routledge & Kegan Paul.

Lepinay, Vincent and Fabrice Rousseau. 2000. “Les trolls sont-ils incompétents? Enquête sur les financiers amateurs.” Politix, 13 (52): 73-97.

Lepinay, Vincent, 2002. “Finance as Circulating Formulas” Paper presented at the New York Conference on Social Studies of Finance, Columbia University, May.

Lewis, Michael. 1999. “How the Eggheads Cracked.” New York Times, January 24.

MacKenzie, Donald. 2000. “Long-Term Capital Management and the Sociology of Finance.” The London Review of Books, 13 April.

—. 2002. “Risk, Financial Crises, and Globalization: Long-Term Capital Management and the Sociology of Arbitrage.” Manuscript, University of Edinburgh.

MacKenzie, Donald and Yuval Millo. 2001. “Negotiating a Market, Performing Theory: The Historical Sociology of a Financial Derivatives Exchange.” Manuscript. Department of Sociology, University of Edinburgh.

Milo, Yuval. 2001. “Safety in Numbers: How Exchanges and Regulators Shaped Index-Based Derivatives.” Paper presented at the New York Conference on Social Studies of Finance, Columbia University, May.

Muniesa, Fabian. 2000, “Un robot walrasien. Cotation électronique et justesse de la découverte des prix.” Politix, 13 (52): 121-154.

—. 2002. “Reserved anonymity: On the use of telephones in the trading room” Paper presented at the New York Conference on Social Studies of Finance, Columbia University, May.

Powell, Walter W. & Jason Owen-Smith (1998) ” Universities and the Market for Intellectual Property in the Life Sciences” Journal of Policy Analysis and Management. 17(2): 253-277.

Preda, Alex, 2002. “On Ticks and Tapes: Financial Knowledge, Communicative Practices, and Information Technologies on 19th Century Financial Markets.” Paper presented at the New York Conference on Social Studies of Finance, Columbia University.

Smith, Charles. 1990. Auctions: The Social Construction of Value. Berkeley, CA: University of California Press.

Suchman, Lucy 1987. Plans and situated actions: the problem of human-machine communication. New York: Cambridge University Press.

Stark, David. 1999. “Heterarchy: Distributing Intelligence and Organizing Diversity.” Pp. 153-179 in The Biology of Business: Decoding the Natural Laws of Enterprise edited by John Clippinger. San Francisco: Jossey-Bass.

—. 2000. “For a Sociology of Worth.” Keynote address, Annual Conference of the European Association of Evolutionary Political Economy, Berlin, November 3.

Wolfe, Tom.1987. The Bonfire of the Vanities. New York: Farrar, Straus and Giroux.

Zaheer, Srilata A. 1997. “Acceptable Risk: A Study of Global Currency Trading Rooms in the US.” Working paper, Wharton Financial Institutions Center, 97-22.

Zaloom, Caitlin. 2002. “Ambiguous Numbers: Trading and Technologies in Global Financial Markets.” Paper presented at the New York Conference on Social Studies of Finance, Columbia University, May 2002.