2015年11月4日 星期三

CHAPTER 29 FINANCIAL AUTOMATION, PAST, PRESENT, AND FUTURE

CHAPTER 29
FINANCIAL AUTOMATION, PAST, PRESENT, AND FUTURE
JUAN PABLO PARDO-GUERRA
Surely, nothing can be more plain or even more trite common sense than the proposition that innovation . . . is at the center of practically all the phenomena, difficulties, and problems of economic life in capitalist society.
Joseph Schumpeter in Business Cycles
Economic theory asserts that technological developments enhance the quality of mar­kets. Financial markets are no exception. Devices of different sorts are widely seen as factors that have allowed for the consolidation of capital markets throughout the world.
Indeed, for economists such as Robert Merton, the benefits of technology overflow the financial system, contributing to the general performance of the economy. “Innovations in telecommunications, information technology, and financial engineering,” argues Merton (with Bodie 2005, “offer the practical prospect for multiple channels for the |AQ: 1 financing of economic growth” Merton’s claims are reflections of a larger body of work dealing with the historical development of finance. The economic literature is rife with demonstrations of how the adoption of specific technologies positively affected market quality—from execution costs and bid-ask spreads to liquidity and systemic risk. Propelled by fierce competition, technology is thus presented as the undisputed hand­maiden of finance.
The standard economic approach to technology in financial markets, however, tends to reproduce a well-known model of innovation: technological change is rendered as a shift in the underlying production functions of the financial system; given a set of inputs, technology may well alter the way in which the different factors of production relate but is not deemed a transformative element of the fundamental mechanisms of exchange.
Markets boil down to exchange, and while technology may make exchange cheaper and more efficient, it does not alter its basic economic nature; calculation remains embodied in the economic agent, not her technological appendices. Technology is thereby cast as neutral and unproblematic.
Such understanding of technology is incomplete, though. The artifacts populating finance are not merely instruments used to optimize the utility of rational agents. Rather, technologies crystallize rules of exchange, formatting the possibilities of action of mar­ket participants; as means of visualization and interaction, they present and scope what is increasingly a disembodied and dislocated—though highly material—marketplace; technologies also constitute part of the distributed agencies of reflexion and calculation that keep markets moving; and as innovations, market technologies are objects of politi­cal, contested intervention. Technologies constitute an important part of the micro­structure of finance.
This chapter follows from Bruce Carruthers’ observation on the relative marginal­ization of finance in sociological thought. As Carruthers shows, while finance has not been completely ignored within the literature, analyses of this sphere of human activity are relatively dispersed and occur on the margins of core sociological con- |AQ: 2 cerns. Typically, finance is relevant only insofar as it alludes to other themes, be they capitalism (Arrighi 1994; Strange 1986), firms and corporations (Fligstein 2001), dis­asters (Perrow 2010), or globalization (Castells 2000). Recent contributions seem to challenge this omission, dealing with finance as a valuable topic in itself rather than an illustration of some broader social process (Knorr Cetina and Bruegger 2002; MacKenzie 2003; Muniesa 2003; Preda 2009. Tellingly, much of this literature con­sists of explorations of the role of market technologies. This is rightly so. The day-to- day operation of the market rests an intricate mesh of technological systems that define the spaces of possibilities of financial action and, as such, have concrete and enduring effects on the dynamics of the market. While rules of exchange (Fligstein
2000)  may inform macro-action, market microstructures are concretely manifested in technology.
This chapter deals with this sociotechnical character of finance. Much of the historical development of financial markets was framed by a type of technological change cap­tured under the rubric of automation. Accordingly, this chapter provides a historical- sociological sketch of the drivers, manifestations, and possible consequences of automation in financial markets. In providing what is an admittedly partial sketch, it captures recent insights of the sociological literature on the place of technology in finance, arguing that financial markets are intelligible only if both their micro- and mac­rostructures are framed in the language of sociotechnical dynamics.
The structure of this chapter is as follows. The next section briefly reviews the con­cepts of technology and automation in finance, and concludes by highlighting the differ­ent drivers and spheres in which automation occurs. This is followed by a description of how automation occurred in the different spheres of financial practice, from settlement and clearing to trade execution. Finally, the chapter offers thoughts on the scope of financial automation in the future.
Financial markets intermediate the movement of capital between notionally “unpro­ductive” savings and “productive” investments. As such, they have played a key role in technological innovation. By pooling and funneling idle money into novel ventures and established institutions, these markets typically allowed states and entrepreneurs to spread investment risks and pursue costly projects—from the transcontinental railway systems of the nineteenth century to the developments in pharmaceutics and biotech­nology of recent decades.
Financial markets, however, were traditionally spaces populated by relatively mun­dane technologies. For most of their history, the bricks and mortar of finance were liter­ally that: bricks and mortar, framing the creation of contractual agreements of different sorts. Indeed, even as late as the nineteenth century, the material cultures of stock mar­kets were formed by little more than chalk, ink, paper, marble, and wood, fueled by the authority of interpersonal trust and the courts, and the manual labor of armies of clerks.
The economic expansion of the nineteenth century generally and the growth of mar­kets in sovereign debt and joint stock shares particularly posed novel possibilities for entrepreneurs. As economic networks coalesced and financial institutions stabilized, it became increasingly possible to, for instance, buy and sell British sovereign debt in the exchanges of such cities as Paris and Amsterdam. This form of geographic fragmenta­tion of markets (that is, the fact that the market for a specific contract existed in multiple sites rather than in a single locale) entailed exploiting price differentials across exchanges—the price of a contract in London need not have been the same as in Paris or Amsterdam owing to the speed at which relevant information arrived at each city. Thus, sophisticated entrepreneurs enabled with access and efficient communications could make a profit through arbitrage: buying in the cheaper to sell in the dearer marketplace. The adoption of telegraphs among financial institutions largely responded to such profit opportunities.
How should we conceptualize these innovations and, accordingly, the sociology of financial automation? The adoption of telegraphs would seem to suggest that the canon­ical economic representation is correct: technology improves efficiency but does little else to the dynamics of the market. While Paris, London, and Amsterdam may have been better interconnected, their local practices were largely unchanged. Here, however, we can refer to three critical sociological perspectives in order to better understand the imbrications of technologies and finance. At one level, the now classical discussions in the sociology of technology of the 1980s shed light on the dynamics of technological change. Following a different tradition, debates on the role of technology within organi­zations allow us to gauge the manner in which specific systems structured practice in the financial sphere. Lastly, recent insights on the place of technology in distributed forms of economic action highlight the diversity of technology in the marketplace and thereby the different forms and logics of automation deployed by actors across finance.
The early literature in science and technology studies presents technological change as a contingent process requiring the alignment of, and successful negotiations between, different interested actors (Bijker et al. 1987; Bijker and Law 1992; MacKenzie and Wajcman 1999). Costs, efficiencies, and other seemingly objective measures of perform­ance alone do not determine the success of technologies; the meanings and contexts of practice associated to novel artifacts and systems have to be settled among all the rele­vant stakeholders before they become established facts. Further, the articulation of such meaning is often riddled with struggle and resistance—innovation is political—for tech­nology is not about materialities but about the social relationships they entail. Paraphrasing Bijker and Law (1992), markets are built along with their objects and artifacts.
The construction of market institutions is not subordinate to innovation, though. Social and technological change are better visualized as intertwined, codependent, mutually constructed trajectories. This intricate relationship is particularly clear in the way in which organizations adapt to innovations. As Wanda Orlikowski (1992) argued, within organizations technology both structures action and serves as a mechanism for the emergence of agency: artifacts may constrain human agency (a standard keyboard, for instance, is ill-suited for designing a book cover), but they also facilitate certain forms of action (the same keyboard allows interacting with a specific software in order to write a book chapter). Technology neither determines human action nor is it itself determined by intentional design. The specific dynamics of technology are influenced by prevailing institutional conditions which may well include normativities of proper action as well as attitudes toward technology and the organization of work itself. A change in such social conventions can become as radical in transforming sociotechnical trajectories as a “rad­ical” technical innovation.
Artifacts also congeal agency, serving as coordinating elements in networks of human and nonhuman actors. Pocket calculators shared by traders and programmed with spe­cific valuation formulas, as well as distributed software systems that present similar views of the world to agents in distant places, allow for specific forms of action coordina­tion to emerge in the marketplace: in the first, formulas may be performed into being (MacKenzie and Millo 2003) while in the second sophisticated forms of valuation are achieved despite the absence of copresence (Knorr Cetina and Bruegger 2002). Even something as seemingly innocent as the telegraph—a cable connecting two distant financial centers through information distributed piecemeal—can have profound reper­cussions in the way in which market participants act and conceptualize their worlds. In his meticulous account, Alex Preda (2009) showed how the adoption of the telegraph changed the understanding of finance in the late nineteenth and early twentieth centu­ries, leading to new forms of interaction between brokers, investors, and the market.
To these three perspectives we may add a fourth, comprised by a more general socio­logical understanding of automation. Given its connections to capital, labor, and work­place processes, automation is a recurrent theme within several traditions of sociological thought. In the early years of mechanization, machines were rendered revolutionary catalysts of change, be they virtuous substitutes for “the lowest operations of the human


intellect” (Babbage, quoted in Daston and Galison 2007), or perilous examples of “machinery dispensing] with muscular power” (Marx, quoted in MacKenzie 1984).
Later authors saw the spectre of automation as a corrosive force within the system of relations of production, one that subjugated workers under the gears of the machine. In his now classical analysis, for instance, Harry Braverman (1974) warned of the perils posed by automation to the skilled worker: with its increased technical sophistication, management could now do by mechanical means “that which it had previously attempted to do by organizational and disciplinary means” (Braverman 1999). De-skilling was thus an unavoidable consequence of the displacement of workers by machines, an example of the relentless technological logic of capitalist production.
For other authors, however, automation was less direct, possibly less nefarious. David Noble, for example, insisted on a more nuanced approach to the effects of automation. For him as for others (see Attewell 1987; Noble 1978), the dreams of management were tamed by the realities of the shop floor: de-skilling was less frequent and intense than Braverman anticipated; the embodied and often tacit skill sets of workers proved diffi­cult to replicate and substitute; and the new machineries of industry perennially required—at one level or another—some degree of manual intervention.
The trading floors of finance are not the shop floors of heavy industry, though, calling for a context-specific definition of automation. Although skilled shop-floor workers and some of the relatively wealthy financial intermediaries faced similar challenges and responded in similar ways to the rise of one form or another of automation, their logics and possibilities of action were quite different. Automation, in a sense, came to signify a specific form of finance and, consequently, was embraced or challenged in a distinct manner by financial practitioners. The relevant question is thus: How should we define financial automation?
Financial automation is a historically located process initiated in the mid- to late 1950s at the confluence of several independent trajectories. At one level, automation derived from the reorganization of the securities industry that resulted from a rise and reconfigu­ration of financial activity during the immediate postwar period, particularly in the United States (Friedman 1980). New traded instruments, a newfound affluence, and finance-friendly policies in America led the market capitalization of the companies form­ing the Dow Jones Industrial Average to grow from around $216 billion in 1949 to more than $13.5 trillion today (for a thoughtful account of financialization, see Krippner 2010). Onto this tremendous growth, regulatory changes in Western countries facilitated the introduction of specific forms of technological intervention in the market, including rules that allowed mechanized forms of clearing and settlement as well as electronic net­works for dealing in securities. In addition, the initial wave of financial automation fed from the expansion of information and communication technologies into general busi­ness applications (Cortada 2005) and the ensuing development of new organizational routines around such technologies. Thus, automation was a collusion of three broad fac­tors: increased trading volumes, new technologies, and friendly regulatory intervention.
While some authors would suggest that in its broadest meaning financial automation has roots in the development of tabulating machines during the late nineteenth century


(see, for instance, Yates 2005), as a category used by investors, regulators, and market participants “automation” only became relevant in later times. In effect, to restrict the definition of automation to the second half of the twentieth century brings to bear the |AQ: 3 historical dimensions of the concept: to automate is not merely to standardize and mechanize processes. Rather, automation implies an attempt to create automaticity in a system, to give it some form of endogenous automatic control. To automate is, in a sense, an explicit quest to construct an automaton, to create a sophisticated machine that sim­ulates, perhaps even stands in place of, reality (Baudrillard 1996; Bedini 1964). This dream, to a very significant extent, was only embraced by finance in relatively recent times.
Financial automation is, above all, a technological transformation, and, observed from the theoretical perspectives described in this section, it is revealed to be a multifac­eted creature. In engaging with this creature, it is convenient to deploy an analytical tax­onomy sensitive to the institutional conditions and spheres of financial practice as well as the drivers fueling technological development and adoption. A taxonomy addressing the historical development of financial automation is here defined in terms of two axes.
At one level, we must distinguish between the relevant spheres of financial activity within which technologies are adopted. In broad terms, financial activity can be concep­tualized as occurring in three overlapping domains of practice: execution, information dissemination, and settlement. Each of these domains of practice entails different insti­tutional actors and conditions. Thus, firms involved in settlement and clearing are nota­bly different from those involved in selling and distributing information to the marketplace. A different level is given by the distinct drivers of technological develop­ment and adoption. Such drivers capture the institutional rationale leading firms and policymakers to acquire or campaign for particular systems and technological solutions. Historically, we can speak of four such drivers, namely costs, efficiency, control, and sur­veillance. The drivers are further described below.
Projects of financial automation often occur in connection to the deployment of four specific logics. The first and perhaps most prominent is pecuniary, and this is largely compatible with canonical economic theories of technological change. For proponents of automation, developing and adopting new technologies responds to cost incentives that can materialize as reductions in skilled labor force and increases in the productivity of specific processes. A historical illustration is given by the automation of settlement and clearing. Described in Charles Babbage’s account of the “original” computer at the Bankers’ Clearing House in London (Babbage 1835), settlement is the process whereby the deals reached in the marketplace are formalized, and securities and cash are exchanged between buyers and sellers. But settlement was traditionally a labor-intensive activity, requiring a gargantuan computer formed by armies of couriers, skilled clerks, countless pencils, paper, calculating machines, and the occasional lawyer, all of whom
would work in unison to match, confer, and finalize every single trade in the market. Consequently, changes in settlement were driven primordially by cost. Current forms of settlement have substituted the humans and paper jams of the past with electronic rela­tional databases, telecommunication networks, enterprise management software, and technicians of several sorts that match and confer orders at much lower costs (Pardo- Guerra 2010; Wells 2000).
文字方塊: AQE5Although costs and efficiencies are related, the latter is not necessarily an implication of the former. Bearing a long and illustrious intellectual history (Alexander 2008), effi­ciency is much a techno-economic measure as it is an object of ideological intervention. Within finance, efficiency is often presented as a defining characteristic of virtuous mar kets (Fama 1970). Indeed, the discursive supremacy of efficiency has fueled some of the most historically relevant projects of automation. The emergence of the National Market System in the United States, for instance, was arguably influenced by ideals of efficiency as paramount to fairness and justice in the national marketplace (see the work of Mendelson and Peake 1979). In this sense, automation is a project informed by the nor- mativities and politics of efficiency, shared by market participants, regulators, and tech­nologists. Technology is not only a solution to the cost-benefit challenges of economic competition, importantly, it is framed as a worthy element that will make markets more transparent, efficient, and fairer. Technology is, in this sense, the means for the realiza­tion of an ideal (and ideological) marketplace.
The third driver is organizational: implementing automated systems is not merely a response to exogenous market signals, fundamentally it is also an intelligible solution to problems of governance, command, and control. Automated technologies are intro­duced to achieve fairness among a set of market participants by, for instance, giving equal access to market information (Preda 2006), but technologies are also introduced to configure regimes of governance and accountability within specific domains, as dem­onstrated by the hierarchical accounting systems of financial institutions (Hatherly, Leung, and MacKenzie 2008; Power 2007). A shared database is not only a convenient tool for collective work: crucially, it is a mechanism for managing a project and its par­ticipants; it is as much an instrument of power as it is one of distributed cognition.
Automation also entails shifting the boundaries of what and who is visible. Markets, however technological and automated they may seem, have always required some level of manual intervention, be it in the form of the trader’s judgment in calibrating specific automated trading systems or incarnate in the technologists who design and maintain the platforms of exchange. Behind the machines of automated finance one will always
find one form or another of human agency. Financial automation, in this sense, is cou­pled to specific forms of organizational self-representation that render some people invisible while making others more prominent. Like the chief engineer of heavy indus­try, elite financial managers gain visibility for their ability to coordinate the machinery of the market. And, like the skilled worker on the shop floor, “traditional” traders lose some of their political and symbolic prominence to novel types of intermediaries, such as individuals with training in science and engineering, who have an allegedly closer control of sophisticated market technologies. Automation can be, as Marx once argued, a means for dispensing with human power, for reconfiguring organizations around the needs of a specific group. Automation is as much about cost and efficiency as it is about control.
Automated technologies stand at the core of most of the practices in the financial serv­ices industry, embodying the equations, algorithms, and forms of visualization through which intermediaries engage with the market, gauge risk, and construct collective strat­egies (Beunza and Stark 2004; Knorr Cetina and Bruegger 2002; MacKenzie 2003; Mayall 2006; Preda 2009). But, owing to their capacity to record vast quantities of data, such technologies are also useful in monitoring the (mis)behavior of market partici­pants (Williams 2009). Support for automation has often come from regulatory agen­cies that see in electronic systems the possibility of complete surveillance. Indeed, the image of technology as means for achieving total control, command, and intelligence over the world is not new (see, for instance, Edwards 1996; Mirowski 2002) and has roots in technocratic discourses across the political spectrum (Boettke 2006).
These drivers had different expressions across the three spheres of financial activity (execution, information dissemination, and settlement). To illustrate the role played by these drivers in shaping the evolution of financial technologies, this section explores how automation evolved in different spheres of practice. We will start our exploration from the end of the trade (settlement) to its initiation (advice), reflecting to some extent on the historical diffusion of automation within finance.
Exchange ultimately requires the transfer of money and property rights between buyers and sellers. Keeping track of and successfully completing financial transactions there­


fore requires vast calculative efforts: matching deals, correcting for errors, and control­ling the transfer of certificates and money across physical and electronic domains. Effectively, the earliest forms of financial automation occurred in settlement operations in the early twentieth century, with the application of tabulating machines initially (c.i94os) and computers subsequently (c.i96os) to the accounting systems of the back office. The criticality of these systems (and thus, the possibilities for automation) was made evident in a series of episodes of systemic stress, when trading levels exceeded the capacity of manual order processing and required more efficient settlement systems (e.g., Wall Street’s paper jam of 1963; see Wells 2000).
After the initial wave of settlement automation, back-office settlement and clearing operations moved away from in-house systems onto independent, often collaborative, ventures. This integration of settlement was driven not only by costs and economies of scale but, as importantly, by notions of control; while settlement was traditionally a responsibility of stock exchanges, it was taken under the control of consortiums formed by end users such as institutional investors and investment banks.
This reconfiguration, which started in the late 1980s, resulted from two interrelated processes. At one level, the dynamics of settlement responded to the physical possibili­ties presented by the internationalization of finance of the second half of the twentieth century: as firms extended their reach throughout the globe, high-reliability proprietary communications networks became increasingly attractive mechanisms for information exchange and made possible for things to be settled across borders (Loader 2005). The costs associated with maintaining the infrastructure of proprietary networks, however, were often prohibitive unless syndicated, leading to industry-wide initiatives that pro­vided a common platform for communication. Similarly, the fact that the shared net­works were based on standardized protocols of data inputting and message transmission allowed for reductions in operational risks associated with the use of free texts in finan­cial communications (these risks emerge due to differing interpretations of a specific input, which can be reduced through standardization; see the analysis of SWIFT by Scott and Zachariadis 2010).
The consolidation of clearing and settlement also follows from the imperatives asso­ciated with the mechanics of settlement, including reducing conflicts around error cor­rection and isolating risk along rational/functional lines. At the organizational level, for instance, clearing institutions provided a relatively neutral ground for matching trades, one which would have been difficult to internalize by the firms that originated orders. By centralizing the operation under a single institutional arrangement, clearing houses provide a legitimate mechanism for resolving conflicting trade reports. Such a role, fur­thermore, made clearing and settlement institutions repositories of specialized forms of legal and bureaucratic knowledge that would have been costly to create and maintain within traditional trading venues such as stock exchanges, and medium-sized broking firms and investment banks.
At a regulatory level, innovations such as the dematerialization of shares followed from attempts by policymakers to minimize risks borne out of the physicality of paper as a medium of legal instruments (Cerny 1994). These regulatory shifts were fueled by


notions of efficiency, deeming physical shares inadequate for a globalized economy. As a measure of financial policy, operational risks in settlement are often seen as incompati­ble with market risks, requiring the creation of distinct, legally independent, institu­tions. Although such rationalization of risks by no means implies the detachment of clearing and settlement from trade (Millo et al. 2005), it nevertheless implies a trans­formed industrial organization and, concomitantly, the emergence of novel organiza­tional cultures, expertise, and systemic interdependencies.
An interesting element of the technological trajectories of clearing and settlement concerns their relation to the temporalities of markets and how they reflect technologi­cal and economic notions of efficiency (Miyazaki 2003). Clearing and settlement are performed over relatively long temporal windows—for instance, regulations stipulate that settlement should occur within days of a transaction, rather than hours, minutes, or seconds—the systems implemented in this area tend not to be time-critical but, rather, to focus on achieving resilience, stability, and reach. Such characteristics are visible in the development and evolution of so-called large-value payment systems (LVPS). Designed as conduits for large monetary transfers between financial institutions (nota­bly, banks), LVPS constitute the technological backbone for mobilizing financial capital across the world.
Yet, a key aspect configuring the shape of such systems is their attempt to minimize “systemic risks” derived from the discontinuities of everyday financial exchange. The unsettled deals remaining at the end of a working day are sources of risks. An illustra­tion of this is provided by the fund transfer system of the United States’ Federal Reserve—Fedwire. Established in 1918 as a proprietary telegraphic network that allowed funds to be transferred between the 12 Federal Reserve Banks in America, Fedwire was opened to nonmember users in 1981 (FRBNY 2010). Since clearing is a discontinuous process, “financially healthy” users of Fedwire are allowed to incur in daylight over- |AQ: 6 drafts. Such overdrafts, however, can be considerable. While the value of transactions in Fedwire in 2003 was in the order of $704 trillion (Martin 2005), daily overdrafts stood at around $65 billion in 2008 (Federal Reserve 2009). As a form of debt, the overdraft bears a settlement risk. A default can have serious consequences on the stability of the market by unravelling a series of large unsettled positions.
With the growing volumes of international transactions and the overall organiza­tional complexity of clearing, settlement, and payment systems (BIS 2008), real-time communications and processing are increasingly forwarded as logical bridges between continuous intraday trading and seamless payment. In effect, under the normative umbrella of efficiency, a trend exists within financial markets to consolidate the indus­trial organization of settlement (BIS 2005) and accelerate its pace of operations in order to bring them into line with the speed of trading. Arguably, the technological develop­ments in clearing and settlement are making financial markets resemble their economic idealizations by reducing the separation between the trades as conducted on the screen and those conducted on the books: with real-time settlement, execution will be tanta­mount to exchange.


FINANCIAL AUTOMATION, PAST, PRESENT, AND FUTURE 577
A thread linking the evolution of different automated market systems is the provision of electronic forms of market information. Automation in finance is embedded in an ear­lier proliferation of electronic communication systems, which included entities such as the telegraphic networks of Reuters (Read 1992) that served as predecessors to the latency-sensitive networks of today’s trading venues. Not only do these forms of infor­mation dissemination shape the forms of apprehension of market participants (Preda 2009), but they also shape the possibilities of evolution for the system as a whole.
Market information is undoubtedly central to modern finance. Market information sets the pace for the market serving as the basis for future economic action and as an obligatory referent for justifying past behaviors. Market information, in particular, is the element that closes the self-referring loop of the financial space: it is created by the market for the consumption of market participants. Yet market information is a peculiar object of consumption. Its afterlife is null, its contents are varied, and its effects unpre­dictable by definition (Knorr Cetina 2010). It is, furthermore, customarily attached to the technicalities of financial communication, from the codes and protocols of account­ing that imbue company reports to the very architecture of the information networks that distribute data across the world. Indeed, the automation of the dissemination of market information has been the basis of much of the automation of trading and settle­ment services, since these two necessitate information in order to operate.
^                                      Owing to its importance, the main driver of the automation of market information is                                          ^
control: over its production, dissemination, and use. The initial waves of automation in information dissemination trace to the nineteenth century, when telegraphs were intro­duced to finance (Preda 2006, 2009). These technologies, however, were not truly auto­mated in a modern sense—they required armies of clerks to feed and interpret information.
Automated information systems emerged with the adoption of computer technolo­gies which, coupled to communication networks, allowed for new forms of data dis­semination. In London, for instance, the introduction of automated information systems in the late 1960s and early 1970s occurred as a result of the technological spillovers from settlement (that is, the investments in computers that were purchased for settlement but could be put to additional uses), as well as the realization that distributing information was a profitable venture in itself (Pardo-Guerra 2010a, 2010b). The control of informa­tion in London by means of sophisticated telecommunications systems represented not only a narrow commercial strategy but, more broadly, control over the marketplace; it was, above all, a demonstration of power of the main trading venue in Britain, the London Stock Exchange (LSE). It was only in the 1990s that regulatory intervention “unbundled” the stock exchange’s relative monopoly, obliging it to distribute market information to secondary resellers under principles of fairness, equality, and efficiency.
In America, the automation of market information was driven primarily by regula­tory concerns. The historical predominance of the New York Stock Exchange (NYSE)


over the national securities market provided an incentive for the Securities and Exchange Commission (SEC) to endeavor to put in place a comprehensive program of automation that included the creation of a consolidated system for market information, one that would link different trading venues through electronic means. Thus, starting in the late 1960s, a series of technological systems were deployed to create a national infrastructure for the flow of market information. In America, the automation of market information was also a response to control, but, unlike Britain, control was not of an institution over the market but rather of the regulator over an institution, the NYSE. The politics of tech­nology in American markets are largely comprehensible in these terms, as a contest between the once monopolistic force of the NYSE and the regulatory imperatives of the SEC.
Importantly, the automation of information dissemination was crucial in allowing for the development of a third sphere of automated finance: execution. The automation of trades requires a stable framework of information dissemination, providing equal data on market events as well as a robust infrastructure for communication through which actors could engage in exchange. Indeed, some of the most salient examples of auto­mated trading were the product of the development of sophisticated informational net­works (such as Reuter’s trading systems, which became the standard for the foreign exchange market from the 1970s onward). Without a standard platform for information dissemination, automated exchange would be simply unfeasible.
The automation of trading is a recent example of changes introduced into the financial system by technological change. It is not surprising that the demise of phys­ical, face-to-face trading and the rise of new modes of financial activity through elec­tronic networks are canonically presented as the exemplars of financial automation (Castells 2000) and its organizational correlate, financial globalization (O’Brien 1991; Shiller 2003). The increased use of electronic systems in the everyday operation of finance is symptomatic of the institutional developments of the financial services industry over the past half-century. However, the introduction of these systems did not initially imply a complete automation and geographic de-location of trading. Most of the early waves of trading automation located in the 1970s and 1980s concen­trated on the provision of electronic price and quote dissemination services rather than on the automation of trading per se. Considerable technological, regulatory, and organizational efforts were required for these automating technologies to allow trades to move off the floor of stock exchanges and into the offices of investment banks and other intermediaries.
The consolidation of automated trading—visible in some recent trends in the use of computer systems that automatically create orders and submit them to the market with­out direct human intervention—was the product of developments that occurred prima­rily in the 1990s and early 2000s. For instance, the automation of execution largely


followed the development of novel investment strategies that were intensive on trading, requiring such things as adjusting portfolios in real time to changes in the market or in the expectations of market participants (examples of these are the strategies followed by hedge funds, which advocated automated trading since the 1990s). Conceptions of effi­ciency and control had to congeal in order for automated trading to become a common feature of financial markets: market participants had to develop new expectations of the speed of finance as well as instruments to deal with an increasingly technological and fragmented marketplace. The automation of trading was not merely a reactive adoption of ready-made technologies in response to competitive pressures. Quite the contrary, trading automation was fueled by numerous pressures, with economic incentives pro­viding but one of a constellation of justifications for technological change.
The historical trajectories of trading automation are thus varied. In some cases, such as that of the LSE, automation was almost serendipitous product of, first, the acquisition of computers to mechanize back-office operations and, second, the hiring of specialists to implement, maintain, and progress technological systems within the organization. For the leadership of the stock exchange up to the 1960s, technology was merely an aid to laborious tasks such as account keeping and data transmission. As a British broker of the 1970s warned those tempted to feed their business into a compu­ter, “dealing done that way will never be fun. Dealing certainly ought to be, and I think that between humans it usually is” (Paul Bazalgette, quoted in Kynaston 2001). This vision was contrasted by the views of the LSE’s technological experts. For George Hayter, one of the leading technology managers of the stock exchange, the market was “100 percent information. . . . Starting with market information to the broker and his client, generating an order, [and ending in the] contractual exchange of ownership and exchange for money. . . . The whole blazing thing is actually information flow, from start to finish” (Hayter, quoted in Pardo-Guerra 2010). A relatively small, stable, and politically autonomous group, the heads of technology departments within the stock exchange drove much of the policies of market automation. Undeniably fueled by the regulatory incentives of Margaret Thatcher’s government, the automation of the LSE’s front office in the late 1980s (in the form of automated quotations and execution sys­tems) was thus an organizational outcome of the empowerment of technologists and their conceptions of efficiency, derived from the departmental units created to ration­alize the adoption of information technologies for the back office in the 1960s (Pardo- Guerra 2010).
Other trajectories of market automation involved qualitatively different processes. The automation of the Paris Bourse, for instance, was framed by a constant negotiation of the technical translation of accuracy, fairness, and justice in the construction of its first automated quotation system, the Cotation Assistee en Continu (CAC). Introduced in 1986 for dealing in the less liquid securities, and implemented over the entire market in 1989, CAC was originally planned as a technological solution to the long-standing informational asymmetries of the open outcry system that traditionally characterized exchange on the Parisian trading floor. The development of CAC, however, encountered problems notably different from those that percolated systems production in the City of
London. Derived from Toronto’s Computer Assisted Trading System (CATS), the con­struction of CAC was not mired in the difficulties of creating a new system from scratch or of erecting a new and competitive mechanism on the basis of legacy systems as had been the case of the LSE. While the crux of British finance was grappling with the organ­izational politics of an expanding material market agencement, French finance was entangled in the politics of the algorithmic configurations that were to substitute face- to-face trading (Muniesa 2003). With automation, order queuing, trader anonymity, and the definitional and procedural aspects of informational intermediation became contested issues upon which actors had to settle in order to allow the symbols presented by the automated quotation system to be perceived as accurate and legitimate sources of economic action (Muniesa 2007). For machines to serve as the basis for exchange, the politics of meaning had to cool down.
Both Paris and London share the commonality that governmental intervention— whether in the form of specialized reports, as in the case of French financial reform, or the threat of anti-monopolistic actions, as was the case of Britain—influenced the pace and direction of automation. The guiding role of regulation, however, is visible in the paths taken by the technological infrastructures of the American financial sphere. Financial markets in America were configured by two different influences: competition and regulation. The American financial system reflected a fragmented history, with dif­ferent trading venues competing to capture a larger share of the market. For instance, although prominent and emblematic of Wall Street, the NYSE shared the financial space with regional exchanges—from San Francisco and Philadelphia to Chicago and Boston—national trading venues—including the American Stock Exchange and the National Association of Securities Dealers—and so-called “Electronic Communication Networks” like Instinet.
Competition was tempered by the centralizing agency of the SEC, which since the 1930s was the uncontested regulator for banking and financial industries. Through its political clout, the sec proved fundamental in catalyzing the automation of American finance. Authorized by Congress in 1961, the sec created a special group responsible for studying and investigating the “adequacy, for the protection of investors, of the rules of national securities exchanges and national securities associations” (HCC 1961). Published in 1963, the Report of the Special Study Group stressed the advantages of auto­mation and electronic quote dissemination in “vastly increasing the flow of market information and . . . insuring better executions for the public” (SEC 1963). Surveying the future, the Special Study prepared the grounds for digital finance. From their consulta­tion with specialists in electronic technologies, the Special Study Group noted
the potentiality of a system which would select the best bids and offers, execute orders, and clear transactions. Transmitting and receiving units would be installed in the offices of all subscribing broker-dealers. Wholesale dealers and other broker- dealer subscribers could enter quotations (and size of market) into a central compu­ter for indexing under the appropriate security and could interrogate the computer to determine the highest bid and lowest offer, selected by the computer, together with the number of shares bid and offered at such prices. (SEC 1963)


FINANCIAL AUTOMATION, PAST, PRESENT, AND FUTURE 581
The fragmented marketplace, the “lack of central location,” could thus be “overcome by the use of a single, central computer” from which information about trading would flow to both the professional dealer and the public.
Information technologies became instruments of control and regulation. By the mid- 1970s, the US Congress and the SEC extended their regulatory control through the transformation of the infrastructures of American finance. Amendments to the Securities Exchange Act in 1975, in particular, mandated the development of a National Market System (NMS), giving oversight and control over its establishment to the SEC. The objectives of the NMS resounded both with the ideals of economic theory and the regulatory imperatives of the SEC. The novel arrangement was to enhance the economic efficiency of transactions, ensure fair competition, provide broad availability of infor­mation, and—subjected to best execution policies—guarantee the possibility of auto­mating trades. Toward the end of the 1970s, a suite of technological systems gave shape to the NMS: while the Consolidated Tape System unified the reporting of trades, the Consolidated Quotation System created streams of market information from the trad­ing venues to data vendors.
Where are these innovations heading? In the past two decades, automation has spread throughout finance, redefining much of the way in which markets operate. It is now common to encounter discussions among both regulators and market participants that boil down to controversies on the technological infrastructure of the marketplace. The levels of technological sophistication of investment firms are historically unparalleled. And the very pace of innovation is increasingly hectic: the chip maker Intel, for instance, operates a research and development center in Slough, UK, that caters specifically to the needs of the high-speed, latency-sensitive sectors of the securities industry. The same drivers apply as before: costs, efficiency, control, and surveillance interact to shape the trajectories of market technologies in the present as much as they did in the past.
An interesting feature of current trends of technological innovation in finance that derives from the combination of these drivers and the sociological nature of economic knowledge is the emergence of what could be referred to as a pattern of fragmented innovation. Fragmented innovation may well constitute a descriptor for the future of finance, a future largely configured through the dynamics of technology in general and automation in particular.
Fragmented innovation derives from the interaction of three aspects of the financial system: competition, innovation, and knowledge. It is clear that economic considera­tions remain at the core of most of the developments in finance and, compounded with regulatory pressures toward increased competition, costs are paramount to the survival of firms and their financial innovations. Firms within the financial system—from tradi­tional investment banks to sophisticated funds—have to compete fiercely to attract


business and maximize returns. Such competition can occur in numerous ways. However, most forms of competition are based on either technological advantages (for instance, possessing a fast system that outpaces those of competitors) or technical advantages (such as having better trend-prediction algorithms or risk control models). Effectively, trading takes place in a knowledge-intensive domain in which technology and technique are paramount: information is not only the blood of the marketplace, it is the ultimate source of profits, making algorithms and a few microseconds of execution speeds features that separate winners from losers.
But in such a knowledge-intensive domain, secrecy is bound to emerge. Indeed, pre­liminary evidence suggests that in the most technologically intensive pole of the invest­ment domain (that of high-frequency traders, which are highly automated, sophisticated market participants who utilize algorithms to trade thousands of shares in fractions of a second), interfirm communication is particularly scarce: information about strategies, algorithms, and system configurations is worth its weight in gold. Hence, incentives do not exist for systemic coordinated innovation; innovation occurs largely in secretive silos, all of which are aimed in the same direction: increasing the speed and reducing the costs of the system.
The outcome of such a process is one in which structural uncertainty emerges: every actor in the marketplace uses its own standards, algorithms, and processes, hiding them from competitors. Furthermore, there is no incentive for transparency, given that trans­parency has a negative impact on profits. Under such structural uncertainty the market becomes more difficult to predict and regulate; it becomes impossible to know the mar­ket. The mechanisms of coordination that exist in an open market and which link dispa­rate sets of local knowledge (von Hayek 1949) are lost in a fragmented system, catalyzed by the pressure of competition. The market becomes less tractable among its partici­pants and, in the event of mistakes, more prone to oscillations. The extreme events of the so-called flash crash of May 6, 2010, demonstrated the fragmented technological fabric of finance and may be evidence of this pattern of innovation: despite months of work and gigabytes of data, the commission convened to analyze the flash crash was unable to produce an account of the events that convinced most market participants. Ironically, the same technologies once introduced as instruments of surveillance and control have colluded to create an increasingly unknowable system.
The pervasive presence of technology in the financial sphere may give the impression that markets are evolving toward an increasingly automated state. This may be true from one perspective, but such claim would miss a critical aspect of technological development in finance: manual intervention is always present as technologies remain irrelevant outside of the networks of humans and nonhumans in which they acquire meaning. In trading, decision-making remains a highly interpersonal activ-


FINANCIAL AUTOMATION, PAST, PRESENT, AND FUTURE 583
ity, populated by a plurality of instruments and devices that are utilized in judging the state of the market and constructing positions in the system. And despite the growth of algorithmic systems that automate some of the decision-making process, fine-tuning parameters and responding to market surprises is an activity in which expertise and experience remain key referents. Even in the highly electronic domain of market information, there remain numerous instances of manual intervention. Documents have to be read, interpreted, and reformatted before they become infor­mation and are distributed to users. Automation is always incomplete, and it is pre­cisely its inherent incompleteness which fuels, to an extent, further technological change in the marketplace.
Automation has important consequences on the evolution of markets. At the end of the day, automation requires armies of technologists, who through the years have become central to the growth and reconfiguration of finance. Expert technologists involved in the automation of financial centers are, in a way, the human tails of the mar­ket’s microstructures. Traditionally conceptualized in terms of the formal rules of exchange that define the interactions of the marketplace (O’Hara 1995), market micro­structures are also comprehensible in terms of the material arrangements that shape, constrain, and make possible the practices of financial agents (cf., Beunza and Stark 2004; Zwick and Dholakia 2006). Here, technologists—from engineers to managers, from programmers to vendors—must be understood not as providers of instrumental platforms, of crystallizations of rules of exchange, but as active shapers and modifiers, expert constructors of the platforms of interaction, of the microstructures of the market itself.
Considering engineers and technologists—that is, the key agents of automation—as active participants in the history of finance implies embracing their heterogeneous prac­tices and promissory expectations as elements that shape the evolution of finance. Finance is not merely driven by a quest for profits. It is as much driven by pragmatic ide­ologies of efficiency as by the political discourse of control. Indeed, the forms of decen­tralized innovation that characterize the development of financial systems are central to an explanation of the dynamics of financial change. There is, in this sense, no single tra­jectory, no standardized route for market evolution, but rather an ecology of instru­ments, platforms, techniques, and worldviews that define the past, present, and future of action in the financial sphere.
Alexander, J. K. (2008). The Mantra of Efficiency: From Waterwheel to Social Control. Baltimore,
MD: Johns Hopkins University Press.
Arrighi, G. (1994). The Long Twentieth Century: Money, Power, and the Origins of Our Times.
New York: Verso.
Attewell, P. (1987). “The Deskilling Controversy.” Work and Occupations, 14/3: 323-46. Babbage, C. (1835). On the Economy of Machinery and Manufactures (4th edn). London:
C. Knight.


Baudrillard, J. (1996). The System of Objects. New York: Verso.
Bedini, Silvio A. (1964). “The Role of Automata in the History of Technology.” Technology and Culture, 5/1: 24-42.
Beunza, D. and Stark, D. (2004). “Tools of the Trade: The Socio-Technology of Arbitrage in a Wall Street Trading Room.” Industrial and Corporate Change, 13/2: 369-400.
Bijker, W. E. and Law, J. (1992). Shaping Technology/Building Society: Studies in Sociotechnical Change. Cambridge, MA: MIT Press.
------ , Hughes, T. P., Pinch, T. J., and American Council of Learned Societies. (1987). The Social
Construction of Technological Systems: New Directions in the Sociology and History of Technology. Cambridge, MA: MIT Press.
BIS (Bank for International Settlements) (2005). New Developments in Large-Value Payment Systems. Basel: Bank for International Settlements.
------ (2008). The Interdependencies of Payment and Settlement Systems. Basel: BIS.
Boettke, P. 2006. “Hayek and Market Socialism,” in E. Feser (ed.), The Cambridge Companion to Hayek. Cambridge: Cambridge University Press.
Braverman, H. (1974). Labor and Monopoly Capital: The Degradation of Work in the Twentieth
Century. New York: Monthly Review Press.                                                                                      |AQ: 7
------ (1999). “Technology and Capitalist Control,” in D. MacKenzie and J. Wajcman (eds.),
The Social Shaping of Technology. Maidenhead: Open University Press.                                       |AQ: 8
Castells, M. (2000). “Information Technology and Global Capitalism,” in A. Giddens and
W. Hutton (eds.), On the Edge: Living with Global Capitalism. London: Jonathan Cape. |AQ: 9
Cerny, P. G. (1994). “The Dynamics of Financial Globalization: Technology, Market Structure, and Policy Response.” Policy Sciences, 27/4: 319-42.
Cortada, J. W. (2005). The Digital Hand, Volume 2: How Computers Changed the Work of American Financial, Telecommunications, Media, and Entertainment Industries. New York:
Oxford University Press.
Daston, L. and Galison, P. (2007). Objectivity. Brooklyn, NY: Zone Books.
Edwards, P. N. (1996). The Closed World Computers and the Politics of Discourse in Cold War America. Cambridge, MA: MIT Press.
Federal Reserve (2009). Fedwire Funds Transfer System.                                                                     |AQ: 10|
Fligstein, N. (2001). The Architecture of Markets: An Economic Sociology of Twenty-First Century Capitalist Societies. Princeton, NJ: Princeton University Press.
FRBNY (Federal Reserve Bank of New York) (2010). Fedwire" and National Settlement |AQ: 11 Services.
Friedman, B. (1980). Post-War Changes in the American Financial Markets. Cambridge:
National Bureau of Economic Research.
Hatherly, D., Leung, D., and MacKenzie, D. (2008). “The Finitist Accountant,” in T. Pinch and R. Swedberg (eds.), Living in a Material World: Economic Sociology Meets Science and Technology Studies. Cambridge: MIT Press.              |AQ: 12
HCC (House Commerce Committee) (1961). Securities Market Investigation.                                  |AQ: 13|
Knorr Cetina, K. (2010). “The Epistemics of Information.” Journal of Consumer Culture, 10/2:
171-201.
------  and Bruegger, U. (2002). Global Microstructures: The Virtual Societies of Financial
Markets. American Journal of Sociology, 107/4: 905-50.
Krippner, G. R. (2010). Capitalizing on Crisis: The Political Origins of the Rise of Finance. Cambridge, MA: Harvard University Press.
Kynaston, D. 2001. The City of London: A Club No More, 1945-2000. London: Chatto &
Windus.


financial automation, past, present, and future 585
Loader, D. (2005). “The Role of the Clearing House,” In Clearing and Settlement of Derivatives.
Oxford: Butterworth-Heinemann.                                                                                                   |AQ: 14
MacKenzie, D. (1984). “Marx and the Machine” Technology and Culture, 25/3: 473-502.
------ (2003). “An Equation and its Worlds” Social Studies of Science, 33/6: 831-68.
------  and Millo, Y. (2003). “Constructing a Market, Performing Theory: The Historical
Sociology of a Financial Derivatives Exchange” American Journal of Sociology, 109/1:
107-45.
MacKenzie, D. and Wajcman, J. (1999). The Social Shaping of Technology (2nd edn).
Philadelphia: Open University Press.
Martin, A. (2005). “Recent Evolution of Large-Value Payment Systems: Balancing Liquidity and Risk.” Federal Reserve of Kansas City Economic Review, 2005/1: 33-57.
Mayall, M. (2006). ‘“Seeing the Market’: Technical Analysis in Trading Styles” Journal for the Theory of Social Behaviour, 36/2: 119-40.
Mendelson, M. and Peake, J. W. (1979). “The ABCs of Trading on a National Market System.”
Financial Analysts Journal, 35/5: 31-42.
Merton, R. and Bodie, Z. (2005). “Design of Financial Systems: Towards a Synthesis of Function and Structure” Journal of Investment Management, 3/1: 1-23.
Millo, Y., Muniesa, F., Panourgias, N. S., and Scott, S. V. (2005). “Organised Detachment:
Clearinghouse Mechanisms in Financial Markets” Information and Organization, 15/3:
229-46.
Mirowski, P. (2002). Machine Dreams: Economics Becomes a Cyborg Science. Cambridge:
Cambridge University Press.
Miyazaki, H. (2003). “The Temporalities of the Market.” American Anthropologist, 105/2:
0 255-65 # Muniesa, F. (2003). “Des marches comme algorithmes: sociologie de la cotation electronique a la Bourse de Paris” PhD thesis, Ecole des Mines de Paris, Paris.
------ (2007). “Market Technologies and the Pragmatics of Prices” Economy and Society, 36/3:
377-95.
Noble, D. (1978). “Social Choice in Machine Design: The Case of Automatically Controlled Machine Tools, and a Challenge for Labor” Politics & Society, 8: 313-47.
O’Brien, R. (1991). Global Financial Integration: The End of Geography. London: Pinter Publishers.
O’Hara, M. (1995). Market Microstructure Theory. Cambridge, MA: Blackwell Publishers.
Orlikowski, W. J. (1992). “The Duality of Technology: Rethinking the Concept of Technology in Organizations.” Organization Science, 3/3: 398-27.
Pardo-Guerra, J. P. (2010a). “The Automated House: The Digitalization of the London Stock Exchange, 1955-1990,” in B. Batiz-Lazo, J. C. Maixe-Altes, and P. Thomes (eds.), Technological Innovation in Retail Finance: International Historical Perspectives. London: Routledge. |AQ: 15|
------ (2010b). “Creating Flows of Interpersonal Bits: The Automation of the London Stock
Exchange, c. 1955-1990” Economy and Society, 39/1: 84-109.
Perrow, C. (2010). The Next Catastrophe: Reducing our Vulnerabilities to Natural, Industrial, and Terrorist Disasters. Princeton, NJ: Princeton University Press.
Power, M. (2007). Organized Uncertainty: Designing a World of Risk Management. New York:
Oxford University Press.
Preda, A. (2006). “Socio-Technical Agency in Financial Markets” Social Studies of Science,
36/5: 753-82.
------  (2009a). “Brief Encounters: Calculation and the Interaction Order of Anonymous
Electronic Markets” Accounting, Organizations and Society, 34/5: 675-93.


586 JUAN PABLO PARDO-GUERRA
------ (2009b). Framing Finance: The Boundaries of Markets and Modern Capitalism. Chicago:
University of Chicago Press.
Read, D. (1992). The Power of News: The History of Reuters, 1849-1989. Oxford: Oxford University Press.
Scott, S. and Zachariadis, M. (2010). “A Historical Analysis of Core Financial Services Infrastructure: Society for Worldwide Interbank Financial Telecommunication (S.W.I.F.T.).” Information Systems and Innovation Group, Department of Management, LSE, Working Paper No. 182.
SEC (Securities and Exchange Commission) (1963). Report of Special Study of Securities Markets. Washington: U.S. Government Printing Office.
Shiller, R. J. (2003). The New Financial Order: Risk in the 21st Century. Princeton, NJ: Princeton University Press.
Strange, S. (1986). Casino Capitalism. Oxford: Basil Blackwell.
Von Hayek, F. A. (1949). Individualism and Economic Order. London: Routledge & Kegan Paul.
Wells, W. (2000). “Certificates and Computers: The Remaking of Wall Street, 1967 to 1971.” The Business History Review, 74/2: 193-235.
Williams, J. W. (2009). “Envisioning Financial Disorder: Financial Surveillance and the Securities Industry.” Economy and Society, 38/3: 460-91.
Yates, J. (2005). Structuring the Information Age: Life Insurance and Technology in the Twentieth Century. Baltimore, MD: Johns Hopkins University Press.
Zwick, D. and Dholakia, N. (2006). “Bringing the Market to Life: Screen Aesthetics and the Epistemic Consumption Object.” Marketing Theory, 6/1: 41-62.
AQ1: Pl. add page no. of quote.
AQ2: Pl. specify if 2009a or b (see comment in References).
AQ3: Pl. reword for clarity. Brings to bear the concept on what?
AQ4: Please specify whether the reference “Pardo-Guera 2010a or 2010b.”
AQ5: Pl. add to References.
AQ6: Pl. clarify: does this mean ‘allowed to incur overdrafts during daylight hours’? AQ7: Pl. add page range.
AQ8: Pl. add page range.
AQ9: Pl. add page range.
AQ10: Pl. add publication details.
AQ11: Pl. add publication details.
AQ12: Pl. add page range.
AQ13: Pl. add publication details.

AQ14: Pl. add page range.

沒有留言:

vedio transcript

 00:13 vì có một số nguyên vật liệu cần đăng ký mua, vậy nên chúng ta sẽ bắt đầu nói về việc mua mặt hàng này trước. 00:23 bộ phận thu mua...