How IT Mediates Organizations: Enron and the California Energy Crisis

1 Introduction

On May 24 1999, 6:10 a.m., Timothy Belden, working from a computer on Enron Corp.'s trading floor in Portland, Oregon, entered four bids to sell power to the California power exchange. His proposal to sell 2,900 megawatts on the next day, at $18 to $29 a megawatt-hours, was approved by the exchange authority. At 7:29 a.m., Belden announced to the state power grid that he would use the Silverpeak line, a terminal with a maximum capacity of 15 megawatts located in a Nevada desert town with the same name, to transfer the sold electricity. This single announcement confused the grid operators, caused a false congestion on the automated system that manages the state's transmission lines, allowed Belden to offer a series of counterbids to relieve the congestion, and raised the price of electricity by more than 70% across California. Californians paid about $5 million in added electricity charges that day, and Enron, which shared the profits with some of the local electricity operators, paid a penalty of $25,000 for its failed bid to the power grid.

A similar pattern of behavior was reportedly exhibited by Enron traders in other times and places, all modeled on Silverpeak, but variously given enigmatic names such as "Death Star", "Get Shorty", "Fat Boy" by Enron traders. Behind the enigma of the names, however, a common pattern of behavior was manifest: "the company pretended to flood transmission lines with excess power, and then earned fees for alleviating the nonexistent congestion by routing the electricity elsewhere". The aim of this paper is to try to make sense of these activities and the common pattern among them. The main question we want to answer is: "Why are some calculative agencies able to impose the events, actions and relations that other calculative agencies have to take into account in making their decisions?" (Callon 1998a). The purpose here is to apply the framework of actor-network theory (ANT) to better understand the dynamics of IT and organizational behavior, using the well-documented relation between Enron and the California energy market as a case study.

The above question, I believe, cannot be answered in terms of explanations based on simple causal accounts. Such accounts, depending on disciplinary and ideological affiliation of its originator, would single out one of the many possible causes (or a combination thereof) to answer the question: Belden's individual entrepreneurial/gaming behavior (economics/psychology), provisions/loopholes in California's energy regulations (law), Enron executives' genius/malice (culture/ethics), the overpowering capabilities/absurd fallibilities of computers (information systems), and so on. Examples of the latter type of explanation abound in both popular and professional accounts of information technology (IT), where utopian and dystopian views, respectively, invoke technology as the panacea or as the source of all evils (Kling 1997). Such accounts not only fail to answer the above question, they fall short of answering a more basic question about the possibility of calculation under uncertain circumstances: How can agents calculate when no stable information or shared prediction on the future exists? (Callon 1998a).

Broadly speaking, calculation, as used in the preceding paragraph, would involve the capability of agents to create:

1. A list of the possible of states of the market (in terms of participants, goods, and their distribution).
2. A ranking of these states in terms of their desirability.
3. The identification of the actions that lead to those states (ibid).

This notion of calculation, as Callon has argued, presupposes not only the availability of complete information about the above triple components, but also of the resources needed to attain the desired states of the world -- a presupposition that is hard to justify under the accounts mentioned previously, simply because they isolate the agents in closed worlds of individual competencies, culture, technology, and so on. An actor-network approach, to be advocated in this paper, can, however, justify that assumption because it postulates that:

If agents can calculate their decisions, irrespective of the degree of uncertainty concerning the future, it is because they are entangled in a web of relations and connections; they do not have to open up to the world because they contain their world. Agents are actor-worlds (Callon 1998a).

Under this conception, to make sense of incidents such as Silverpeak, one needs to take into account the network(s) that link Belden to a multiplicity of other players, from Alaska (and beyond)'s gas reservoirs, Portland's electric power generators, Nevada's power grid, and California's automated transmission system, through the state's legislation, its Independent System Operator (ISO), its residents, and its utility companies, to Wall Street media, banks, and analysts, Silicon Valley entrepreneurs and manufacturers, Enron executives, investors, and lobbyists, the technologies that they used or envisioned, and federal bureaucracies in Washington, DC. Far from being predetermined and essential, the identities of these actors "all depend on the morphology of the relations in which they are involved" (ibid). Furthermore, as we will see with many of these examples, the players not only cover vast spatial and temporal spans, they have transitory hybrid identities with technical, social, legal, financial, institutional, and technological dimensions. Bringing these variegated elements together calls for an explanatory framework that is not based on clear-cut dichotomies between the natural and the social, the local and the global, the micro and the macro, and so on -- a framework, in short, that admits and accommodates the proliferation of quasi-objects and hybrids (Latour 1993). Actor-network theory, as proposed and elaborated, among others, by Callon (1986a, 1998a), Latour (1987, 1988, 1993), and Law (1992) provides such a framework.

1.1 Actor-network theory: a synopsis

Originating in sociology of science, actor-network theory is an evolving body of knowledge that is being applied to increasingly diverse fields of inquiry. The dynamic and "diasporic" character of ANT is indeed one of its hallmarks, to the extent that some of its originators renounce labeling it as a "theory" altogether (Law 1999, Latour 1999). But there are some core intuitions common to the different "versions" of ANT, a synopsis of which might be helpful here.[i]

In rough outline, ANT analyses socio-technical processes in terms of networks where resources are concentrated in a few places ("the knots or the nodes") connected with one another by "the links and the mesh" (Latour 1987). A network, according to ANT, is comprised of heterogeneous "actants" the most durable of which -- people, institutions, tools, texts, money, technologies, information, etc. -- flow through the network, in a way defining and creating the nodes. Depending on the relative flow and concentration of these "immutable mobiles," some nodes and actors may acquire a privileged status, rendering the situation "irreversible." That is, they find it possible to make autonomous choices that not only fall in line with those of the other actors but makes it impossible to go back to a point where alternative possibilities exist (Granovetter 1973; cf. Callon 1998a). As some theorists with views similar to ANT have argued, this margin of maneuver can become very large if an actor succeeds to situate itself at the intersection of two or more networks that hardly overlap elsewhere (Burt 1993). The crucial point to be borne in mind is that, according to ANT, the form and identity of the actors, far from being inscribed in their nature, is defined by "the relations in which they are located," and that "they are performed in, by, and through those relations" (Law 1999). To create and populate networks, actors often employ different devices -- e.g. problematization, enrollment, interessement, or mobilization -- the common objective of which is to "translate" an idea, through the identification of a problem or opportunity, into reality (Callon 1986a).

Of special relevance to this paper is the ANT perspective on technology (or, more accurately, on techniques). For this purpose, we are going to use the notion of "mediation" that plays a key role in the more recent ANT literature. In contrast to the notion of "intermediary", mediation is an event or an actor that cannot be exactly defined by its input and its output (Latour 1999). It is an occurrence that is "neither altogether a cause nor altogether a consequence, neither completely a means nor completely an end" (ibid). Latour discerns four meanings of technical mediation -- namely, mediation as interference, as composition, as reversible blackboxing, and as delegation. Interference takes place in a process of translation where one actor (e.g. a technology) displaces the original goals of another actor and is itself displaced in terms of its goals -- a person is different holding a gun and the gun is different when it is being held in a hand. Composition takes place when different actors, human and nonhuman, join forces (so to speak) in order to perform an action and to achieve a goal. Action, in this sense, "is simply not a property of humans but of an association of actants" -- Boeing 747s do not fly; airlines fly (ibid). Reversible blackboxing captures the ever-changing, unstable character of associations made among different actants, sometimes leading to an opaque picture of the joint production of actors and artifacts (blackboxing), and other times leading to the dispersal and disintegration of those same actants (reversal). One of the key steps in this process is the formation of an "obligatory passage point" -- that is, a point where anyone with a stake in the network would have to pass through in order to attain their goals. Finally, delegation is manifested whenever techniques find their own meaning and their own way of articulation, as when a speed bump (which in French is called a "sleeping policeman") silently invites drivers to slow down (whether to save human life or to protect their car's suspension). This paper shows how IT played a mediating role in all of the above senses in Enron's activities.

A number of ANT concepts have been applied specifically in this study. Table 1 illustrates some of these key notions and the way they relate to the current study.

1.2 On method: pasteurization of France and Enronization of the USA

The purpose of this study is to explain the role of Enron in the California energy market through a close examination of Enron's "networking" strategies of which IT was a major component. For this purpose, one needs to conduct a study that involves both historical and sociotechnical dimensions similar to, for instance, what Latour (1988) has done in the study of Louis Pasteur and his microbes. Despite great differences in actants and their motivations and strategies, I believe, Latour's method applies rather well to what, as we shall see, can be appropriately called the "Enronization of the USA".

Being temporally detached from the events under study, Latour draws upon available reports and documents to reconstruct the story by basically using text as data. I seek to do the same here. Although we are temporally much closer to the objects of our study, the complexity of the situation and the turn of events after Enron's debacle in December 2001 makes it very difficult, if not impossible, to acquire first-hand data about the situation in a systematic way (through field study, interviews, etc.). Therefore, despite our desire to do so, I decided to draw upon the abundant information about Enron, in professional and trade journals as well as in academic accounts and the public press. Enron has been under the spotlight for many years, originally for its highly rated performance throughout 1990s and then for its controversial collapse in 2001. I have taken advantage of these sources, compiling and studying hundreds of them, a sample of which is provided in the additional information sources. Almost all of these accounts represent Enron's greater (and later, lesser) ability to mobilize allies and to control the presentation of its image in the public and professional eyes. IT, as we shall see, has played a central role in these efforts, hence making the case an interesting one from the perspective of social informatics.

As mentioned earlier, actor-network theory is taken up as a useful framework by researchers in different areas. What makes it a worthwhile analytic tool in information systems' research is its close attention to micro-analysis. Walsham (2001) reports a few studies of information systems that use the ANT framework and applies it to some case studies himself. The story behind Enron's involvement in the California energy market might well be more complicated in texture, more stretched in time and space, and more variegated in terms of players than typical cases studied so far, but we want to demonstrate that the perspective of social informatics in general and ANT in particular can provide explanations in these complex situations as well, if care is taken by the analyst in adopting a dynamic and historical perspective.

For the purposes of this study, historically one might as well go back in time to the 1880s, when Thomas Edison and a close circle around him embarked on a long-term project to delineate the boundaries of what was to become the electric industry of the later decades in the USA. However, thanks to numerous studies of the industry in the USA, many of them close in spirit to a social-informatics approach (Hughes 1983, Granovetter and McGuire 1998), we will begin much later, in the early 1990s, when energy (gas, oil, and electricity) markets began an integration process under new laws and when IT also played a central role in many different ways. It is in the same period that Enron Corp. adopts a strategy that would transform it from a gas and oil utility company to an asset-light, global trading enterprise.

2 Background: the electric industry, Enron, and IT

In their study of the making of the electric industry in the US in its formative years of the late 19th and early 20th centuries, Granovetter and McGuire (1998) have shown that the way this industry developed was one of several possible outcomes and not necessarily the most technically or economically efficient. They have forcefully argued that the particular form of the industry, which finally stabilized in private central station firms, "arose because a set of powerful actors accessed certain techniques and applied them in a highly visible and profitable way" (ibid). They have also illustrated that these techniques resulted from the shared personal understandings, social connections (friendships, clubs, professional associations), organizational conditions, and historical opportunities available to those actors. One such opportunity was provided by the policy of the federal and state governments that encouraged a strategy of "grow and build," that promoted mergers, and that rewarded firms with the greatest access to investment capital -- all of this at the obvious expense of decentralized alternatives and under the concept that electricity provision was a "natural monopoly" (ibid). This led to a regulated market, which was gradually institutionalized in later decades, guaranteeing profits for central station firms.

Such was the state of the electric industry throughout most of the 20th century. For reasons the analysis of which is beyond our scope here, however, the situation started to change in early 1990s. One of the reasons, but by no means the only reason, for this change was the emergence on the scene of new ICTs, which facilitated the integration of the energy markets in all areas of production, transmission, distribution, and consumption. In 1994, the Federal Energy Regulatory Commission (FERC) mandated, in separate orders, the use, respectively, of electronic bulletin boards (EBBs) and real-time information networks (RINs) to exchange information about gas pipeline trading and electric power trading among all players in the respective markets. The proclaimed goal of this initiative was "to squeeze the inefficiencies out of the marketplace" (Posposil 1995), and to "open up" the market by structuring exchanges and transactions in a so-called seamless fashion -- that is, to create a network that would allow suppliers, transporters, and consumers to visibly interact with each other "by punching a few computer buttons" (Durgin 1994). Introducing the impersonal style of interaction implicated by the electronic media was, however, a huge undertaking in an industry where the "close-knit network of relationships may slowly change" (ibid). And, to use Granovetter and McGuire's proposition once again, such an undertaking called for "powerful actors [who] accessed certain techniques and applied them in a highly visible and profitable way" (ibid). Enron Corp. was such an actor, or, more precisely, transformed into such an actor, as a 'brief history of Enron' (Appendix A1) illustrates.

By 1995, Enron managed to attain an overwhelming dominance in the energy market, to an extent that, Enron Power Marketing Inc. (EPM), made 42% of transactions in the whole industry. This dramatic ascendance to the top was sometimes attributed to such parameters as "the companies financial resources [that] set it apart from the increasingly crowded field" (Pospisil 1995). However, stories based on IT capabilities and the Internet were central among the others, and were indeed the ones most publicized by Enron itself (PR Newswire 1997). For instance, Enron Capital & Trade -- where traders immersed themselves in an IT-intensive environment, often having as many as four different types of wireless devices and twelve $2000 flat-screen panels on their desks (Berinato 2002) -- was reported to be the most successful Enron subsidiary, apparently because of the speed of its operations. The speedup derived from extensive use of IT was highlighted as the central feature of company life. Arguing that the high cost of information transaction costs is the reason behind vertical integration of the type pursued by major oil companies, Kenneth Lay, Enron's chief executive, advocated a "virtual integration business" model for the energy industry similar to that of Dell Computers Inc. (Poruban 2001). "The internet -- and IT in general -- has turned this movement [toward vertical integration] "on its ear" and has helped to drive information transaction costs to "virtually zero"" (ibid).

These techno-centric views were widely received and promoted by both public and professional media. What made them attractive was not only their resonance with the "spirit" of the times (which was, by the way, created by other actors such as Enron) and their immense promotional power, it was also the (too) easy explanations they provided for basically any success story. The oddity is that sometimes the same reasons are invoked, in a magical reversal, to explain failures. Thus, Enron's bankruptcy in December 2001 was also attributed by some to excessive use of IT (Berinato 2002), or as a lesson about its importance in corporate governance. These oddities illustrate the need for more plausible explanations. The actor-network framework, as we have mentioned earlier, affords such explanations.

2.1 Enron: the intersection of multiple networks

The ANT approach seeks to explain the capability of actors for maneuvering and for making profits in their capability to situate themselves at the intersection of otherwise disparate networks. This, we believe, is what happened to Enron. To act as a power marketer, Enron needed to position itself as the representative or "spokesperson" of many other players who had a stake in this market and with whom Enron were to establish various alliances. It had to define itself, in other words, as a common node among many different local and global, financial and political, social and technological networks. Ekbia and Kling (2004) have argued why it is impossible to understand the rise and fall of Enron without expanding the scope of study to include these broader networks. We have shown how the interaction among the entities of such networks are constantly negotiated in a continuous fashion, giving rise to multivalent and nuanced links that vary from time to time, place to place, and actor to actor. Here we want to show the process of creation and evolution of such networks by taking a closer look at one specific aspect of Enron's activities -- namely, its involvement in California's energy market during the 1990s.

This process begins early in the decade with the attempts of top Enron executives to open the way for the company's involvement in the electricity market, not so much as a newcomer but as a trader. These attempts -- which were often characterized by Enron officials and economic experts as the opening of a market "that needed to be opened" (Wayne 2002) -- are better described, in our view, as attempts to disseminate and delineate multiple networks within and around the market in the name of perceived common interests. A rough look at the 'energy battles of the early 1990s' (Appendix A2) illustrates the nature of such attempts. (See Klashner (2004) in this issue for details.) One of the key devices in this process was the formation of an obligatory passage point.

2.2 Deregulation: an obligatory passage point

An obligatory passage point is a node in the network that all the actants who have a stake (or a perception thereof) in the problem would have to go through. In the 1990s, deregulation of energy markets was turned into such a passage point by actants such as Enron (Figure 1).

Figure 1. Deregulation as the obligatory passage point in Enron's strategy

For this to happen, different allies had to be convinced that deregulation would serve their interests, generating the image of what seemed to be a "united front":

Enron representatives often came in cold, without any contacts in government or the lobbying community. To compensate, they lined up experts to testify, hired local lobbyists and joined with consumer groups and some local utilities to present a united front for deregulation. (Wayne 2002)

The notion of a "united front" used in the above passage is helpful shorthand insofar as we do not lose sight of the fact that each of the members of the front pursued their own unique objectives. As Figure 1 demonstrates, deregulation for consumers was perceived as a way to become independent of a single supplier, accessing multiple sources and "stylized" products (such as a combination of peak and off-peak rates), paying less energy fees, and even "hedging their bets" in future markets (Pospisil 1995); for the local governments, deregulation promised a way of guaranteeing a reliable supply of energy to residents, of saving money in energy, and of spending the savings for other purposes; to the politicians, in addition to providing a plentiful source of campaign money, deregulatory policies were an effective demonstration of their concern for public good; for industry and business, deregulation was a promise of more opportunities and a more profitable future; for journalists it was a topic for exciting stories; and for various others (experts, charities, universities, etc.) it was simply a source of income; and so forth. Without an active involvement on the part of Enron and other powerful players, however, these actors could have remained unknown to each other, ignorant of their putative common interest. However, deregulation, as an obligatory passage point, brought these otherwise-unrelated actors together in one single mesh.

3 Multiple mediations

The main focus in this paper is to illustrate the different roles that new ICT played in the Enron story, and we have adopted a social-informatics perspective for this purpose. Unlike techno-centric or human-centric approaches, social informatics encourages us to view IT as a socio-technical network, where "technological effects are indirect and involve different time scales ..., articulation work is often needed to make IT work, ... [and] politics are central and even enabling" (Kling and Lamb 2001). That Enron did, in fact, use the most advanced computer technologies of the day in its various activities -- from trading, credit evaluation, and advertisement, to finance, modeling, and risk management -- could hardly be disputed. Enron spent an average of $250 million on IT each year (Dow Jones News Service 1999). In addition to basic IT applications for accounting, human resources, inventory, and so on, examples of IT applications used by Enron as related to this study are:

• Client/Server software (D&B Software): for basic operation (Business Wire 1992)
• SCADA and Lap Link: Remote control operation of pipelines and compressor stations (monitor readings, reset alarms, and operate controls in real time), which resulted in huge savings in field operations (Rood 1999)
• PageWriters (Motorola): to remotely monitor the readings off power meters (Lorek 1998)
• ClickAtHome: a company-wide program to provide all employees of Enron (and its wholly owned subsidiaries) with computers and broadband access at home (Denne 2000)
• Tops 2000 (Savvysoft): Risk management modeling (using ASP services) (Middlemiss 2000)
• Real-Time Information Network (RIN): for power marketing (International Gas Report 1995)

Looking at this (admittedly curtailed) list, it might not be inaccurate to say, as an Enron executive did, that, "Enron no longer considers IT to be a support function. It is part of the core business" (Natural Gas Week 2000). The interesting question, therefore, is not whether Enron used IT to the extent possible or it didn't. Nor is it even whether it used it in the most cost-effective and optimal manner. The questions are:

• What role, if any, did IT play in the translation process?
• In what terms should this role be explained? (In terms of an irresistible enabling role, as dominant narratives tend to show, or in terms of a web of networks populated with heterogeneous players of which IT was but a single instance?)

We would like to argue for the latter view. That is, we want to show that computer and communication technologies, as elements of socio-technical networks spawned by Enron (and other players), played multiple roles. What follows is an elaboration of this point. We have found the notion of "technical mediation," developed in Latour (1999) useful for articulating this perspective. As mentioned in section 1.1, Latour discusses four meanings of technical mediation, and we would like to show how these meanings, and maybe more, are manifested in the way IT mediated Enron.

3.1 Interference: making an image

As mentioned earlier, interference takes place when one actor (e.g. a technology) displaces the original goals of another actor and is itself displaced in terms of its goals. Such was the relation between Enron and the public at large. To enroll the general public as a potential ally, Enron had to establish itself not only as a reliable source of electricity -- consumers needed to be convinced "why they should buy their [power marketers'] electrons instead of the others" (Pospisil 1995) -- but also as the representative or "translator" of consumer interests. But to convince consumers that Enron could represent their interests, a long process of image-building was required through which the public were to be persuaded that Enron does, in fact, have the prerequisite conditions, qualifications and resources to act as a translator of their interests. Translation, we should remember, is a process before it is a result.

To this end, in addition to traditional public relations tools (Appendix A3), there was the (often premature) announcement of high-profile projects that could attract the attention of the public. Prominent among these was a $350 million contract with Sun Microsystems to buy 18,000 Sun servers together with the required data storage (estimated at 500 Gbytes per POP) to build Enron Intelligent Network (EIN), a proprietary network to provide broadband transmission (The World Communication News Report 2000). Another 20-year contract was announced with Blockbuster Entertainment Corp.[ii] to deliver movies on EIN on demand to consumer's televisions or computers -- a project that, if successful, would have turned Enron into a household name in the USA. ( Gosmano et al. 2000, Kranhold 1999). The announcement of the first of these contracts drove up Enron's stock value by about 20% in a single day ( Gosmano et al. 2000).

A brief look at the above methods shows that IT played a central role in the image-making strategy of Enron. It also shows the importance of having "image" as a crucial social dimension of business practice, in the past and present -- this is not new by any means. What might be new are the possibilities generated by the new electronic media in the making of the image. IT constituted a recurring rhetorical theme in Enron's public-relation campaign. Rhetoric, defined as "the art of building alliances to establish a favorable balance of power," was an effective tool in this whole process (Callon 1998a). The most common rhetorical terms employed by Enron and its allies were IT-enabled transparency, speed, and cost saving (Shook 2000), themes that found eager ears within the industry as well.[iii]

3.2 Composition: "Firms of the Future"

Composition takes place when different actors, human and nonhuman, join forces (so to speak) in order to perform an action and to achieve a goal. This is what happened between Enron and the rest of the electric power industry. Within the industry and business at large, a great deal of effort was spent to introduce and establish Enron as a pioneer in what was portrayed as the inevitable direction of the future. The broad strategy was to align Wall Street banks, analysts, and brokerage and consulting firms on Enron's side (Sherman 2002).[iv] More specifically, however, the industry had to be mobilized around the model of business represented by Enron, securing as such a leadership role for the company. Different devices were used for this. One was direct invitation or forewarning: "We'd love to see more companies trying to do what we are doing", declared an Enron manager in Europe (Petroleum Economist 1999); "Any company that has not adopted online services is risking losing its markets," said another executive (Shook 2000). The other device was to portray a strong picture of Enron in terms of its resources and capabilities. On the premise that power marketing requires heavy capital investment -- "for extensive back-office capability and information technology systems," for instance (Pospisil 1995) -- Enron was promoted as a company that could leverage its "considerable assets" to achieve this. Enron could also apply, it was claimed, its vast experience of the gas market in the electric industry.

Among these, arguments based on IT use were the most common and probably most effective. A recurring theme in relation to IT (and, in later years, to the Internet) was the deterministic argument about the accelerating diffusion of technology in society. Kenneth Lay, using the same forum that he had used to problematize deregulation seven years earlier, made such an argument: "A common measure [for asessing the rate of diffusion] is how quickly 50 million customers adopt a new technology ... If you go back to radio, it took 38 years to get those 50 million customers ... television 13 years, cable 10 years; if you go to the Internet, 5 years" (Poruban 2001). Lay also predicted a proportional growth in e-commerce, from $650 billion in 2000 to $6.5 trillion in 2004 (ibid).[v] In addition to Lay and other Enron executives, the top executives of some high-tech companies also played a critical promotional role for Enron. The chairpersons of multinational manufacturers, who had signed contracts to sell metering equipment to Enron, had virtually turned into a mouthpiece for Enron in advocating deregulation (PR Newswire 1997). A speech by Sun Microsystem's CEO, for instance, is believed to have been instrumental in helping Enron enhance its market valuation.

As a result of all this, in the late 1990s, an emergent consensus of opinion about the viability of Enron's business model was shaped among a diverse group of industry and business actors, from oil and gas to electronics, computer, and finance. The goal of establishing Enron as "the firm of the future," which creatively used advanced technology, was accomplished towards the end of the decade, mobilizing a multitude of industry leaders, bankers, analysts and stock traders behind Enron. So successful, indeed, that "On Wall Street, where analysts often look askance at big, risky trading operations, Enron's stock valuation isn't penalized for its trading units" (Dow Jones News Service 1998).

3.3 Blackboxing: turning competitors into allies

In broad outline, the enrollment of industry and public allies, as tedious and challenging as it was, can be deemed successful for Enron by the early 1990s. By the mid-1990s, the company had enrolled enough allies and spawned enough networks to be considered an unrivaled leader in both gas and electric industries around the globe (Denne 2000).[vi] This was especially true in the western parts of the USA, where either the market was deregulated, as in California, or Enron owned and operated generating plants in states like Oregon where deregulation was not achieved. In fact, the restructuring plan in California (Appendix A4) facilitated Enron's involvement in the electricity market of the region in a direct way.

The implementation of this restructuring plan was largely premised on the capabilities generated by new information technologies. What these technologies did was to "fold time and space" (Latour 1999) in ways that were unheard of before; they facilitated the process of blackboxing that is crucial for almost any technoscientific mediation (Latour 1987). Looking again at the kinds of technologies used by Enron in its activities as an energy broker (section 3) might reveal this aspect more clearly. What is important is that all of these technologies were literally Enron's allies in traversing spatial and temporal boundaries in an entirely opaque manner. The story with which we opened this paper is a case in point. Without these instruments, it was impossible for Timothy Belden (on behalf of Enron, of course -- remember Belden does not deceive; Enron deceives) to offer multiple bids, to get approval, and to create false congestion on the network. And all of this happened, not because the computers, sensors, and instruments were broken or malfunctioning, but because they were functional.

Another aspect of blackboxing is also worth mentioning again: blackboxing is always reversible because the associations made among different actants are often unstable. For example, blackboxing was successfully implemented by Enron in its relation to other players in the California energy market such as El Paso Electric. By placing itself, almost literally, between El Paso and the rest of the world, especially consumers but also the industry and oversight bodies, Enron weakened the ties between them, securing an active mediatory role for itself. This, of course, had payoffs for El Paso, which, according to the FERC report, earned in excess of $7 million from one month's joint dealings between the two companies, invoking its top executives to express their appreciation of "what is possible when teamwork, knowledge, initiative and accountability all come together" FERC Report 2002, p. 26). Blackboxing had turned a competitor into an ally but, as the later turn of events demonstrated, this was an unstable alliance that could not be maintained in the face of adverse circumstantial conditions in 2001, something that was also manifested in Enron's relation to its other allies such as investors, banks, politicians, and so on. One of Enron's "mistakes" might be that it viewed blackboxing as irreversible.

3.4 Delegation: let IT speak for itself and for us

Finally, delegation as the last kind of technical mediation is manifested whenever techniques find their own meaning and their own way of articulation. Returning to the questions posed in section 3, we want to show that this happened on many dimensions. As we have seen, some of the IT technologies employed by Enron, indeed, played an important role in the operation and functionality of the company, and hence in its transformation. However, other IT-enabled approaches played more of a delegation role for IT. We want to distinguish among these by considering three such roles:

• Proliferation: to multiply the links and interactions within the networks, complicating them in unpredictable ways;
• Amplification: to increase the calculative power of Enron as a major possessor of technologies, and, through that, to amplify the asymmetry between Enron and other players.
• Delegation: to provide new meanings and new meaningful articulations

The proliferation of links and interdependencies is the direct outcome of the ubiquitous use in advanced societies of new technologies and the multiplication of their products (Callon 1998b). Through these technologies, Enron, as we have seen, was linked up to many other players who would have otherwise had no direct stake in the deregulation of energy markets. The example of manufacturers such as Motorola or Sun Microsystems is a case in point. The executives of these companies became supporters and spokespeople of deregulation not because there was a direct relation between deregulation and the demand for their products, but because this had to go through Enron, and Enron was interested in deregulation. Furthermore, IT also proliferated connections in a more obvious (and well-known) way, and that was by making possible action at a distance in ways that could not be thought of previously (Latour 1987). Without IT, it would have been much more difficult, if not impossible, to control operations and transactions from remote sites, and to link up with actors in remote locations. Enhancing the capability to operate from a distance was, in fact, a major focus of Enron's activities, a fact that can be readily confirmed by a brief look at the above list of IT applications by Enron.

The amplification of calculating power is the second major function of IT in the translation process. Callon (1998a, p. 45) writes:

The probability of gain is on the side of the agency with the greatest powers of calculation, that is to say, whose tools enable it to perform, to make visible and to take into account the greatest number of relations and entities.

This is the aspect that is usually of most interest in information science research. However, there is another kind of IT-related activity, which goes beyond calculation power. An early example of this prevailed in the course of debates about the standardization of EBBs in the gas industry, an issue that had arisen due to interfacing problems among different products. In reaction to FERC's plans for such standardization, and despite the interest of other small marketers in it, Enron pushed for its alternative. One obvious reason was that Enron had already invested in the equipment ("the medium"), and was not willing to reinvest. But the more important reason, as one Enron spokesperson stressed, was the need to concentrate on "standardizing the message, not the medium" (International Gas Report 1995). What Enron sought, in other words, was to let the (standardized) technology take over the role of a delegate of the industry.

A recent example of technical mediation as delegation comes from activities such as selling video on demand over the Internet, which was quietly abandoned but contributed to Enron's image-making and the enrollment of various allies. In a strange way, though, the consumers of these latter IT systems in their "symbolic" mode tend to be outside the company, rather than within it, an observation that could only make sense from an ANT perspective, which opens up the boundaries of agencies and erases the inside/outside dichotomy.

These considerations expand the range of "users" for a given IT application. Traditionally, "users" of an IT application are those who interact directly with it. In this case, Enron does include such users at the micro level, for example, its energy traders. However, some of Enron's IT projects also had important symbolic dimensions that were aimed at participants who were not Enron employees, but who were enrolled by them at a macro level: journalists, investment bankers, and stock investors. Various IT applications helped to actor-network Enron well beyond its organizational boundaries. Again, the removal of such dichotomies as physical/symbolic and micro/macro makes sense within the ANT framework.

A more recent example has to do with Enron Online's trading platform, which was presumably accessible to customers through a kind of subscription. In actuality, however, this was a "one-to-many site" set up in such a way that Enron was always involved as the intermediary in transactions among buyers, transporters, and producers (Maselli 2001). The long-term vision among Enron executives was, in fact, to establish the company, through the construction of a fast proprietary global network, "as the preferred platform for e-commerce around the world". An extensive effort was spent to problematize the Internet by highlighting its unreliability as a medium for commercial transactions. The failure of this effort and the reasons behind it would provide the material for another study from an ANT perspective.

4 Performativity: making things happen

Before concluding we would like to discuss the ANT notion of performativity as it relates to the current topic. Traditionally, as we said, the energy market in the USA has been in the hands of local utilities regulated by state governments. Changing this situation called for a broad mobilization that Enron undertook at both the federal and local levels. At the federal level, most of the activity consisted of lobbying at the US Congress and soliciting political support by making generous campaign contributions. At the state level, however, in addition to such contributions, a much more fine-grained activity was taking place that involved, among other things, the hiring of local lobbyists, experts, and ex-commissioners. As a result of these mobilizations, between 1997 and 2000, deregulation laws were passed in 24 states. As the director of a national institute had put it, "They were trying to make something happen" (Wayne 2002). California provides a good example of the dynamics of this process of making things happen.

As early as 1994, Jeff Skilling, later an Enron executive who headed an Enron subsidiary at that time, testified to utility commissioners in California that deregulation could save the state $8.9 billion in costs, an amount of money that, according to Skilling, "can triple the number of police officers in Los Angeles, San Francisco, Oakland and San Diego" (ibid).[vii] Kenneth Lay, a lobbyist himself in the 1970s, was probably the most active advocate of deregulation. One of the strongest political ties was between Lay, Texas Senator Phil Gramm, and his wife, Dr Wendy Lee Gramm. It is said of these three, all of whom emerged from modest backgrounds and went on to earn doctorates in economics, that "Their professional interests meshed with their philosophical sharing of a passionate distaste for government interference of any type with commercial enterprise" (Fleck and Wallstin 2002). Like other advocates of deregulation, they believe that over-the-counter derivatives constitute capitalism at its purest.[viii] The theory is that if a player can turn something into a commodity, the buyers and sellers will inevitably appear.

The confluence of these economic theories, and the mobilization and enrollment efforts that we have described thus far, finally led to the deregulation of the electricity market in California and 23 other states. This is an excellent example of performativity: the turning of theories and ideologies into reality.

5 Conclusion

In order to prepare the market relationship, it is first necessary to relate, connect and associate. (Callon 1998a)

Having gone through the mediation processes that led to Enron's involvement in the California energy market, we can now make better sense of the story with which we opened this article. To understand Belden's behavior, we need to know how Enron transformed its business. A magic-bullet theory of this transformation process, as pointed out at the beginning, would invoke a number of reasons: the genius of new executives, special resources, legal loopholes, or IT. Such explanations are inadequate because they fail to take into account the crucial role of various actors -- people, technologies, information, documents, money, etc. -- in shaping the transformations. In particular, they fail to provide a full understanding of the role of IT in this process. The approach adopted here makes such an understanding possible. Our study shows that IT played a mediation role in Enron's activities, and in several ways:

• Interference or image-making in the public arena;
• Composition or enrollment in the business arena, e.g. the rhetoric of IT-enabled speed, efficiency, and transparency that resonated with the dot-com culture of the mid-late 1990s in the USA;
• Blackboxing or enrollment of competitors, e.g. taking advantage of the computerized power transmission system in California;
• Proliferation of links, e.g. in accessing remote points and actors via IT;
• Amplification of roles, e.g. in setting the rules of the game;
• Delegation or finding in IT new ways of meaningful articulation, e.g. high-profile projects such as the "Enron Intelligent Network" and movie "on demand" sales on the Internet.

We are not suggesting that every ANT analysis should be rooted in IT use. However, in our era, many firms that adopt a high-technology strategy are also likely to foreground their IT capabilities. This would often accord a pivotal role for IT in their interorganizational behavior (and thus in an ANT analysis). This study is an example of the dynamics of IT development and use and organizational change from an ANT perspective. In this particular case, many IT applications were part of Enron's expansion as a business. In addition, we added a new concept, amplification, to the conceptual lexicon of ANT.

The case of Enron may be problematic for some readers, because the company has declared bankruptcy and also because some of the practices described here (such as the energy market manipulation) were probably illegal. Ironically, it is Enron's high-profile behavior that leads to numerous news stories that we could use as data about complex organizational behaviors. However, many of the ways that IT applications actor-networked Enron can be readily observed at other organizations, but with a less comprehensive overall view because of the difficulties of collecting such data for vast multi-national firms. As many stories have now emerged, we should keep in mind that Enron's ethical transgressions might only be baby steps away from what goes for legitimate trading practices in other arenas. The recent controversy about the management of the power grid in North America is a case in point. Thus we should take the case of Enron as a serious point of departure in future IT research. Enron can actor-network us, even as it may be dying!