Why have financial markets, by and large, evolved into electronic spaces where the majority of trading activity is performed by computer algorithms rather than human traders? It is tempting to explain these changes by reference to the superior efficiency of electronic markets. Paper-based trades must be manually confirmed and settled—an extremely labour intensive activity. Moreover, human traders are slower than algorithms and must be paid—typically a large salary. The digitization and automation of markets, this line of thinking suggests, can best be explained by the improved liquidity and reduced transaction costs created by electronic trading systems: a simple matter of the competitive logic of the market itself shaping the operation of finance. This view is perhaps best captured by Lawrence Glosten’s 1994 paper, “Is the Electronic Open Limit Order Book Inevitable?”Footnote ¹.

Juan Pablo Pardo-Guerra’s Automating Finance challenges this simplistic explanation of financial market automation through a fascinating study of the digitization of stock exchanges in both the UK and the US. In Pardo-Guerra’s account, automation was less a rational imperative driven by a straightforward logic of efficiency than it was a bottom-up process of bricolage by engineers and technicians—many of whom were marginal to the organizational culture of exchanges—who possessed different and competing visions of what an “efficient” financial market would be. These initially invisible actors gradually gained power over time, while their projects were shaped and coordinated by regulators at key moments, who sought to promote automation for their own political purposes.

Pardo-Guerra tells this story with a particular eye to the “infrastructure” of stock exchanges: assemblages of practices, devices, and standards that underpin their day-to-day operation but which tend to fade into the background and become invisible to most financial actors. The concept of “infrastructure”, in this extended sense, has its origins in science and technology studies (STS), particularly the work of the late Susan Leigh Star and her colleagues.Footnote ² An attentiveness to the infrastructures of financial markets helps to avoid some of the blind spots of much of the classic work in the social studies of finance, which tended to frame finance as an epistemic activity analogous to science, and overemphasize the importance of trading and other “front office” activities, relative to those in the “back office”Footnote ³. To this end, Pardo-Guerra studies the transformation of the stock exchange by means of infrastructural inversion: making visible the often-ignored technical systems built by technologists that played a central role in the digitization of the stock exchange, thereby challenging our view of stock exchanges as places defined by the trading floor.

One upshot of this is how it challenges familiar binaries from economic and organizational sociology. Consider, for instance, the distinction between markets, hierarchies, and networks. These are often seen as mutually exclusive forms of social organization. On the one hand, Pardo-Guerra’s focus on infrastructure makes clear that the smooth functioning of a stock exchange critically depends on a large administrative bureaucracy to carry out infrastructural tasks critical to the exchange’s operation, such as the settlement of trades. On the other hand, infrastructural relations also tend to transcend the boundaries of particular firms and organizations. Thus, in tracing changes to these infrastructures, Pardo-Guerra not only shows how stock exchanges themselves became digitized, but how they became increasingly central to the broader ecology of firms and organizations that make up the stock markets.

The book can be divided roughly into two parts. The first part, where Pardo-Guerra’s analysis is richest, focuses on the London Stock Exchange (LSE) as well as a spin-off from the LSE called Tradepoint that pioneered the use of the electronic order book in the London markets. Pardo-Guerra begins with the origin of the LSE—then known as the Royal Exchange—in the late 17th century, and explains the process of social closure that gradually allowed it to develop a distinctive organizational culture among its members. Pardo-Guerra theorizes the Exchange’s organizational culture as a form of kinship among its members held together by the physical structure of the Exchange itself (the “House”). This is an analytically interesting choice that works on multiple levels. On the one hand, it nicely captures the organizational culture of the Exchange, in which its “members” (i.e. jobbers and brokers) had something akin to a master-servant relationship with its engineers and technologists. On the other hand, it also challenges a long-held view originating with Ferdinand Tönnies that kinship and exchange relations are categorically distinct.

After describing the LSE’s origin and culture, Automating Finance examines the gradual digitization of the Exchange during the mid-to-late 20th century. Here Pardo-Guerra carefully uncovers the sedimentary layers of infrastructure—known by acronyms such as MPDS, TALISMAN, and CHARM—that were built up by the Exchange’s engineers and technologists and which gradually effected the transformation of the LSE from an organization defined by a physical place into an internationalized trading system. Pardo-Guerra nicely captures the technical culture practiced by these engineers. They had an affinity for open, hackable technologies (IBM was “too hierarchical”) and saw the Exchange as a complex information system. Yet their work was often driven by practical problems: settling trades in a timely manner was a constant problem for the LSE, one which became particularly acute by the 1960s. However, the efforts of these engineers were, as Pardo-Guerra explains, initially invisible to the Exchange’s “members” who formed the core of the organization. Their partial invisibility afforded the engineers a degree of agency to develop systems in line with their own visions of how markets should run which would eventually transform the core of the LSE itself.

Pardo-Guerra refers to this as a process of “capture”. While he is careful to note that the ability of the LSE’s engineers and technologists to capture the cultural core of the organization was enabled by the LSE’s unique history and the infrastructural relations between its members and technologists, the concept is a potentially useful theoretical contribution, even to researchers outside the social studies of finance. Consider, for instance, the general problem of explaining institutional change. Organizational sociologists often point to “institutional entrepreneurs” as essential in bringing new institutions into being Footnote ⁴. But this leads to what is known as the “paradox of embedded agency”: how can institutional entrepreneurs change the institutions in which they are embedded when their practices and identity are shaped by those institutions? Automating Finance can be read as sketching a new way to resolve this paradox: it suggests that in trying to explain institutional change, we should look to those professional groups that build systems that are critical to a field but which tend to be overlooked or ignored by actors at the core of that field—in short, those that are treated as infrastructure. Under the right conditions, those professional groups can bring about significant institutional change.

While the transformation of the LSE initially emerged from infrastructural projects within the Exchange itself, the automation of the American stock markets instead originated with efforts to challenge the supremacy of the New York Stock Exchange (NYSE), the country’s dominant stock exchange. The second part of Automating Finance tells that story. These challenges both came from upstart exchanges such as Instinet, a company that commercially implemented the first electronic order book, as well as from the Securities and Exchange Commission (SEC), the US regulator in charge of securities trading. Pardo-Guerra shows how regulatory and commercial actors who pushed for increased automation were motivated by competing visions of what it meant for a stock market to be “efficient”, each of which were associated with distinct moral imaginaries of how those markets should, ideally, function. One such vision was of a single centralized order book, which would be composed of bids and offers submitted by both professional market makers and investors throughout the entire country. For various political reasons, however, the SEC came to conceptualize efficiency in terms of “price efficiency”, which instead entailed promoting competition between exchanges rather than the creation of a single, centralized exchange where all orders would reside. This led the agency to prioritize the development of infrastructure that would stitch together the price quotation and transaction recording systems of exchanges throughout the US via a consolidated quotation system and a consolidated “tape”. As Pardo-Guerra explains, the lasting consequence of these choices has been a proliferation of exchanges and other trading venues within the US and the emergence of high-frequency trading, much of which involves exploiting price differences between these venues.

Automating Finance’s dual focus on both the US and UK equity markets invites comparison between the two cases. However, drawing comparative inferences between the two parts of the book is made somewhat difficult by the fact that each uses distinct forms of data. Part one of the book draws extensively on interviews conducted with engineers and technologists who worked within the LSE as well as companies connected to the Exchange, such as Tradepoint. Not surprisingly, this portion of the book largely approaches market automation from a “bottom-up” perspective: infrastructure development essentially originates within the LSE itself and moves outward, with the state playing a secondary role in this process. Part two instead draws more heavily on secondary data and interviews with individuals involved in the SEC’s efforts to develop a National Market System, or who worked at challenger exchanges like Instinet. This leads to a more top-down perspective, which emphasizes the state as the primary actor in bringing about the automation of markets. Automating Finance makes clear that the NYSE was not a passive actor in this transformation; however, we see much less of the internal organizational politics of the NYSE and how they shaped the eventual transformation of US equity markets than we do with the LSE. Thus, while Automating Finance suggests that the state played a much more central role in promoting the digitization and automation of stock trading in the US than the UK, it is not clear to what extent this is due to limitations of the data. I expect that future work on stock market automation in the US will help resolve this question.

While Automating Finance’s major intellectual contribution is to the history of stock market automation, it is also likely to be an important piece of work for those with broader interests at the intersection of technology and finance. In prioritizing the infrastructure of finance, the book may serve as a useful model for how to approach the study of FinTech innovations. Like the automation of stock exchanges in decades past, the growth of the FinTech sector is often seen as an inevitable outcome due to the superior efficiency of these services over those offered by traditional providers, such as banks. But a key lesson of Automating Finance is that simplistic appeals to “efficiency” often conceal competing moral visions that motivate the development of—and in many cases are reinforced by—particular infrastructures. Moreover, unlike more traditional financial services firms—such as banks, hedge funds, or insurers—many FinTech firms frame their activities and ambitions in explicitly infrastructural terms. (Stripe, a prominent payments platform, describes itself as creating “payments infrastructure for the internet”.) Because the products of these companies become embedded into other companies’ own product offerings, following the infrastructural linkages that these platforms generate across firms is necessary to understanding their significance. For these reasons, Automating Finance is not only likely to be useful to those who study the history of financial automation, but also researchers trying to make sense of its present and future.