Introduction
Why do financial intermediaries exist? For an economist, the answer is perhaps obvious: they make financial markets more efficient by gathering and framing information that is incorporated into prices through trading. According to the efficient market hypothesis (EMH), all available information is reflected in prices because any mispricing is quickly exploited through arbitrage. Through this idealized lens, pure-information intermediaries have an ephemeral status: by fulfilling their function, they weaken their value proposition, and hence, their source of income and legitimacy. In an efficient market, no unpriced insight can be sustainably exploited to generate returns above the market rate, so the optimal strategy is to replicate the market instead of attempting to ‘beat’ it. Paradoxically, however, for the EMH to obtain, some actors must defy it, attempting to outperform the market by trading on unpriced information. If everyone believed the market was already efficient, it would never become so (Zuckerman, Reference Zuckerman, Knorr-Cetina and Preda2012: 231). In this sense, information intermediaries resemble Weber’s Protestants in Fredric Jameson’s (Reference Jameson1973) famous reconstruction: by helping the market become efficient, they diminish their own relevance, acting as vanishing mediators.
However, since real markets only tend towards efficiency, and financial intermediaries are rarely pure information brokers, these mediators never quite vanish. What emerges instead is a ‘thick’ social world of financial intermediaries with their own discourses, professional identities, hierarchies of power, and social ties. One might be inclined to analyze this field as possessing a set of social conditions conducive to (more) market efficiency (Zuckerman, Reference Zuckerman, Knorr-Cetina and Preda2012), in a broadly Weberian vein (recall, for instance, Weber’s analysis of conditions for the ‘maximal formal rationality of capital accounting’). While rooted in a storied sociological tradition, this approach would nonetheless retain the idea of financial market efficiency as its reference point, asking only which social and institutional variables allow intermediaries to fulfill this function more effectively. But what if, instead, one were to drop the idea of market efficiency altogether and approach intermediaries as ‘sui generis… social and cultural phenomena’, rather than analyzing them ‘in relation to the supposed functions they fulfill’ (Millo et al., Reference Millo, Spence and Valentine2025: 44)?
This is the key analytical move made by the three studies under review in this review article. Spanning the period from the burst of the dot com bubble (Ortiz, Reference Ortiz2021), through the post-2008 recovery (Leins, Reference Leins2018), to the volatility of the COVID-19 pandemic (Millo et al., Reference Millo, Spence and Valentine2025), they present an appropriately ‘oligoptic’ view of global finance, based on ethnographies and qualitative interviews conducted in some of its major hubs (Zurich, Paris, London, New York, and Chicago). Each study analyzes financial intermediaries without assuming that they contribute to market efficiency. Moreover, while acknowledging its field-defining influence, all three distance themselves from the performativity idiom, relying instead on the broader idea of reflexivity, which views financial intermediaries as capable of understanding and accounting for ‘various pressures and realities’ of the field, even if they do so selectively (Millo et al., Reference Millo, Spence and Valentine2025: 64).
Images of reflexivity
Leins focuses on buy-side analysts as ‘animators’ of financial markets. While analysts are often mistaken in their predictions, their narratives about the future help other actors to navigate uncertainty, creating a sense of agency in what otherwise appears as a random process (Leins, Reference Leins2018: 137). Leins is careful to emphasize that analysts make it possible for others to perceive the market as efficient, rather than performatively making it so (Leins, Reference Leins2018: 71). Analysts purposively cultivate their expert credibility and work to bring new business to their host banks in order to strengthen their positions internally. By spinning narratives that influence investment flows, they indirectly govern the economy. The more a narrative is accepted as legitimate, the more influence on the direction of investment it acquires, and the closer it moves to becoming a self-fulfilling prophecy (Leins, Reference Leins2018: 156). To accomplish this, however, analysts must become second-order observers, incorporating others’ opinions into their forecasts and engaging with information others are likely to see as relevant.
Reflexivity takes another form in Ortiz’ analysis of the financial industry as a political institution which creates and maintains the hierarchy of access to money. As a decentralized assemblage without unifying logic, this field is held together, Ortiz argues, by standardized valuation methods enshrined in legal contracts, educational curricula, and regulation. These methods ‘do not presuppose logical consistency’ and involve contradictory assumptions; their meaning is instead stabilized through ‘built-in’ moral and political imaginaries centered around value, independent investors, and efficient markets (Ortiz, Reference Ortiz2021: 52). Technically correct financial valuation is supposed to uncover the ‘true’ value of assets and thus contribute to market efficiency, assumed to be a socially optimal and politically desirable outcome (Ortiz, Reference Ortiz2021: 18). In effect, however, valuation practices provide access to money for some activities, while excluding others. Such valuation, moreover, is carried out not by free-floating investors maximizing shareholder value, but by salaried employees of bureaucratic organizations who merely purport to enact the imaginary ‘gaze’ of independent investors – Reference Graeberthe silent scandal of financialized capitalism previously noted by Graeber (Reference Graeber2015). In this context, reflexivity turns into self-vindication (Hacking, Reference Hacking and Pickering1992): ‘investors act according to the best definition of themselves only if they apply this theory, and when they do, they contribute to the market efficiency that the theory itself presupposes’ (Ortiz, Reference Ortiz2021: 197).
Finally, Millo et al., offer an image of reflexivity as inertia. Combining network-and field-theoretic insights, they analyze the mutual constitution of social and cognitive structures in the field of investment advice. The social relations in this field congeal over time, creating interpersonal and interorganizational dependencies and sedimented mental structures interiorized by the actors involved. While these structures underpin both disruptive and inertial tendencies in the field, the latter tend to dominate. Financial intermediaries cultivate their expertise through ‘dense networks and deep connections’, but outsiders see them through reductive metrics (Millo et al., Reference Millo, Spence and Valentine2025: 6). This combination of strong ties needed to maintain field position, and weak outward signals that confer status and legitimacy, creates tendencies towards herding, risk-aversion, conformity, and social and intellectual conservatism (Millo et al., Reference Millo, Spence and Valentine2025: 84). Social inertia makes the ‘epistemic regime’ of investment resistant to change, leading to intellectual ‘ossification’ (a notable trope from Schumpeter-inspired prognoses of generalized capitalist inertia). In line with the reflexivity argument, much of this inertia is deliberate, because ‘a lot of energy goes into defending and maintaining the status quo’ (Millo et al., Reference Millo, Spence and Valentine2025: 8), and habits and routines ‘are often purposefully enacted in ways that create future path dependencies in the field’ (Millo et al., Reference Millo, Spence and Valentine2025: 182).
Taken together, the three images of reflexivity surveyed here – capturing analysts’ second-order observations, the self-vindication of financial theory, and the socio-cognitive inertia immanent to investment advice – help to illuminate the practices of financial intermediaries and the social structure of the field. They also show how reflexive actors purposively reproduce social and cognitive structures that expose them to slippage into self-referential parlor games and the loss of relevance.
The contradictions of finance
Leins, Ortiz, and Millo et al.’s theoretical arguments are developed through a wealth of empirical material, focusing primarily on the ‘active’ investment community which believes that in-depth research yielding differentiated insights helps to sustainably generate ‘alpha’, or an above-market rate of return (Ortiz, Reference Ortiz2021; Millo et al., Reference Millo, Spence and Valentine2025: 26). Initially conducted by sell-side analysts, this research combines ‘small data’ about companies with idiosyncratic insights collected through ‘rich interpersonal interactions’ with sources within and outside of the field (Millo et al., Reference Millo, Spence and Valentine2025: 7–8, 24). These insights are subsequently interpreted with similar valuation methods by buy-side analysts, working for fund managers in charge of clients’ money (Ortiz, Reference Ortiz2021: 91). While the sell side also provides transactional services to the buy side, the field’s illusio assumes that research is key to generating alpha (Millo et al., Reference Millo, Spence and Valentine2025: 95–97). In this sense, analysts closely approximate pure information intermediaries, and are a good guide to the field’s key tensions and contradictions.
The active orthodoxy is challenged by a smaller, more homogenous, and more academic ‘passive’ community. Relying on replication-based investment strategies, this community casts a skeptical eye towards the value of investment research. Indeed, an entire literature has emerged which focuses on exposing the failure of the active community to generate alpha long-term, portraying its members as ‘human equivalent[s] of the QWERTY keyboard’, save that the keyboard ‘at least was the answer to a particular mechanical problem at a point in time’ (Millo et al., Reference Millo, Spence and Valentine2025: 44). Importantly, skepticism about the value of active investing does not require commitment to the EMH. True, if financial markets are indeed largely efficient, there shouldn’t be ‘a place for those who routinely provide overpriced, underperforming services’ (Millo et al., Reference Millo, Spence and Valentine2025: 19). But even if financial markets are not efficient, the active community might not necessarily be able to predict which elements of the information that they collected will affect the markets and how (Leins, Reference Leins2018: 27–28).
Entrenched practices do not crumble if proven futile in theory, as the young Marx knew all too well. The active community persists despite its lack of legitimacy within either neoclassical or heterodox economics, and has effectively assimilated much of modern finance theory within its own rationality, claiming the price discovery function which finance theory describes as a justification for its own value and relevance (Leins, Reference Leins2018: 70; Ortiz, Reference Ortiz2021: 160–161; Millo et al., Reference Millo, Spence and Valentine2025: 162). Taken as a long-term approximation, the EMH gives meaning and urgency to analysts’ work, triggering constant suspicion about whether information is already ‘priced in’ (Leins, Reference Leins2018: 78–81). One of the reasons for this assimilation is the contradiction in the liberal imaginary of markets as simultaneously spaces of free exchange and as quasi-natural entities obeying eternal probabilistic laws (Ortiz, Reference Ortiz2021: 199–200).
This contradiction is reflected also in specific valuation practices. The common sense of the active community is embodied in the practice of ‘fundamental’ valuation, which assumes the existence of a ‘true’ value of a company that might not be captured in its price until the market ‘corrects’ (Ortiz, Reference Ortiz2021: 216; Millo et al., Reference Millo, Spence and Valentine2025: 160). This approach is routinely combined with ‘relative’ valuation, where a company (X), whose value is assessed fundamentally, is compared to a set of companies (Y), whose values are taken to be captured by prices. Paradoxically, this approach assumes that the market is inefficient for X, but efficient for Y, for otherwise, the prices of the latter would be a poor guide to gauge the value of the former. Finally, ‘speculative’ valuation rejects the idea of ‘true’ value and uses similar techniques to arrive at short-term approximations of other actors’ opinions. Despite their conceptual contradictions, all three are routinely combined in the ‘classical investment method’, whereby a portfolio reproduces the market, but the fund manager changes the weights in anticipation of price movements, implicitly assuming that the market can be ‘beaten’ (Ortiz, Reference Ortiz2021: 70).
Such contradictions are ubiquitous in a field of financial intermediation that is characterized by persistent mismatches between ‘the ostensible and the actual roles performed by financial intermediaries’, and the value attributed to them within and outside the field (Millo et al., Reference Millo, Spence and Valentine2025: 103). Three examples, discussed in the following sections, are particularly instructive here: the tension between the pressures to differentiate and to conform; the mismatch between the perceived and the actual value of sell-side services; and the way in which financial intermediaries justify their work in the face of critique.
Conformity and differentiation
Despite the assumed value of research as the primary alpha-generating activity, good analysis is insufficient for professional success without strong social ties (Millo et al., Reference Millo, Spence and Valentine2025: 62–63), the ability to create convincing narratives (Leins, Reference Leins2018: 131), and peer recognition (Ortiz, Reference Ortiz2021: 32). In a classic triadic scenario (Stark, Reference Stark2020), actors in the investment field compete for the attention of clients rather than striving for accuracy per se (Millo et al., Reference Millo, Spence and Valentine2025: 53). Hence, analysts’ performance is measured by their readership size, publication frequency, and their number of client touchpoints (Leins, Reference Leins2018: 141–142; Millo et al., Reference Millo, Spence and Valentine2025: 129–130). A distinctive profile and analytic ‘style’ – of which forecasting accuracy is only one element – is what captures the client’s ‘mind share’.
Because valuation techniques at once presuppose market efficiency and inefficiency, whether the price of a particular company reflects its fundamental value is always uncertain (Leins, Reference Leins2018: 75; Ortiz, Reference Ortiz2021: 69). This uncertainty opens a space for financial analysis, wherein interpretative flexibility is the name of the game (Collins, Reference Collins1981). It allows analysts to build individualized profiles, relying on their cultivated ‘market feeling’ to develop narratives of possible futures. Narrative is prioritized over calculation, which is often delegated to junior colleagues or outsourced; only the ‘right’ numbers are incorporated into the story (Leins, Reference Leins2018: 72–73; Ortiz, Reference Ortiz2021: 104). Analysts are thus keen on choosing from a ‘repertoire’ of calculative approaches and combining them into an individualized ‘style’ of analysis (Leins, Reference Leins2018: 73–75). By doing ‘interpersonal work’ (Millo et al., Reference Millo, Spence and Valentine2025: 96, 61) with company management, analysts seek differentiation through access to dedicated information asymmetries. Yet, as in the case of quantitative analysis, the information used is often selected because of its narrative fit, rather than its potential to move the market, partly due to the difficulty of predicting how any given piece of information will affect prices (Leins, Reference Leins2018: 144, 84). Critically, the same information can be analyzed in many ways, with different importance afforded to timeliness, applicability, credibility, and originality (Leins, Reference Leins2018: 70).
There are several reasons for this imperative towards differentiation. The first is the field-level belief, shared across buy and sell sides, that analysts’ core skill is the ability to generate original forecasts through information asymmetries and that those who can do this should enjoy a better reputation (Leins, Reference Leins2018: 92; Millo et al., Reference Millo, Spence and Valentine2025: 26). Second, differentiation is valorized by the political imaginary of finance: applying highly standardized valuation procedures, employees must nevertheless act as independent investors exchanging opinions in a free market (Ortiz, Reference Ortiz2021: 32–34). Finally – and more pragmatically – individualized forecasts can be customized by salespeople and buy-side analysts for their clients and thus dilute responsibility if predictions prove inaccurate (Ortiz, Reference Ortiz2021: 90; Leins, Reference Leins2018: 131).
At the same time, differentiation has its limits. Like other experts, analysts need to balance exclusivity and broad appeal (Eyal, Reference Eyal2013). On the one hand, to ‘become reflected in the market’, analysts’ forecasts must be ‘in line with others’ (Leins, Reference Leins2018: 86). To spot emerging consensus, analysts incorporate pre-framed (and unoriginal) information others will likely attend to. On the other hand, the application of valuation methods is validated through peer recognition, which confirms that ‘the valuation achieved was indeed the act of a maximizing investor’ (Ortiz, Reference Ortiz2021: 87). Finally, the ‘social nature of evaluation’ in the field pushes actors ‘to revert to an intellectual mean rather than acting on conviction’ (Millo et al., Reference Millo, Spence and Valentine2025: 84). This intellectual mean-reversion partly results from the ‘social stickiness’ of relationships in the field.
Social stickiness and relational infrastructures
Despite academics’ celebration of analysts as market soothsayers, field insiders see their value differently. Analysts can spend months producing research reports that reach into the hundreds of pages but often remain unread, a phenomenon observed in Switzerland and across the Atlantic (Leins, Reference Leins2018: 10; Millo et al., Reference Millo, Spence and Valentine2025: 57; Ortiz, Reference Ortiz2021: 91). Often, the value of sell-side services is not research but the relational infrastructure that it provides, which the buy side sees as ‘a portal’ into investment banking services and brokerage activities, affording privileged access to IPOs and seasoned equity offerings (Millo et al., Reference Millo, Spence and Valentine2025: 70–71, 82). These ‘concierge’ services generate much more sell-side income than research – an ‘open secret’ to both sides (Millo et al., Reference Millo, Spence and Valentine2025: 73). As a result, analysts can become ‘conduits for volume’, whereby their paychecks are effectively ‘written by the companies about which they were supposed to be writing critical assessments’ (Millo et al., Reference Millo, Spence and Valentine2025: 70–71). To avoid this, the buy side tries to diversify their information sources and outsource transactions to unrelated brokers to encourage their sell-side partners to speak freely and sincerely, like independent investors (Ortiz, Reference Ortiz2021: 119–120). Nevertheless, the reality of power asymmetries and ‘social stickiness’ between the two sides regularly defies these expectations.
The buy side is well aware of the social forces that compromise the quality of sell-side analysts, who ‘could be wrong and make a ton of money’, especially at the top of the field, where there is little need to generate research revenue and thus little interest in hiring top analyst talent (Millo et al., Reference Millo, Spence and Valentine2025: 74–76). Moreover, since the sell side needs access to detailed and timely information about the companies they cover, they cultivate relationships with their management, facing ‘an implicit normative demand to put a positive spin on things’ (Millo et al., Reference Millo, Spence and Valentine2025: 66). Analysts generally tend to cover companies with good prospects. Meanwhile, company management prefer working with analysts whose coverage will likely be positive, and with whom they can maintain cordial relationships (Ortiz, Reference Ortiz2021: 89) and can share insights that might be packaged into a good story (Leins, Reference Leins2018: 145). This leads to biases rooted in ‘the sticky nature of social relationships in the field’ (Millo et al., Reference Millo, Spence and Valentine2025: 182). Thus, for instance, a Chicago investment bank had an unwritten policy of never issuing a ‘sell’ recommendation to maintain relationships with the companies that might bring transactions in the future (Millo et al., Reference Millo, Spence and Valentine2025: 66–67). Similarly, though buy-side analysts were prohibited from covering their own banks’ business interests, they maintained at least a 50% rate of ‘buy’ recommendations across the companies they covered because ‘disinvestment is bad for the economy’ (Leins, Reference Leins2018: 142).
Overall, the field is characterized by large ‘areas of interdependence’ between career trajectories on both sides (Millo et al., Reference Millo, Spence and Valentine2025: 59). Salespeople spend months winning the attention of fund managers by providing unsolicited (and initially unpaid) investment advice, then years building a ‘shared way of thinking’, so that the relationship becomes part of each party’s field-specific social capital, transferable across organizations (Ortiz, Reference Ortiz2021: 114–115). The quality of advice provided in this context is endogenous to the ongoing personalized relationship between salespeople and fund managers who engage in the ‘dialogic production’ of financial valuation, assumed to be ‘good enough’ so long as the relationship (and payment) continues (Ortiz, Reference Ortiz2021: 116–117). This dialogue takes place in informal ‘outings’ where the sell side tries to find the right tone to make the buy side ‘susceptible to listening’ (Millo et al., Reference Millo, Spence and Valentine2025: 61) without endorsing too much of the latter’s taste in drinks, music, or women, for fear of ‘looking like a clown’ (Ortiz, Reference Ortiz2021: 125). The buy side reluctantly attends despite the risk of paying for entertainment rather than investment advice, precisely to synchronize their thinking with the sell side through more personalization (Ortiz, Reference Ortiz2021: 123). Thus, if financial theory and the employee practices are chronically prone to slippage into self-referentiality, the social structure of the field makes this outcome more likely still.
The future of intermediaries
The self-referential quality of investment practices takes a different form when investment professionals are challenged to justify their work. While they can be surprisingly critical of capitalism, employees of the financial industry tend to interpret bubbles, crises, technological disruptions, and other elements of the ‘eventful punctuation’ of capitalist development (see Sewell, Reference Sewell2008) as temporary distortions which might be fixed by market ‘correction’. Far-sighted investors who did not follow the herd into an erroneous valuation are thus seen as heroes (Ortiz, Reference Ortiz2021: 216–217; Leins, Reference Leins2018: 102). Similarly, investment professionals’ own high salaries and bonuses are justified by the market efficiency that they help to bring about. Professionals ‘create value’, it is argued, through rising asset prices and the profits of the financial industry (Ortiz, Reference Ortiz2021: 230, 191).
These justifications are becoming increasingly thin with the rise of big data-and AI-enhanced passive investing, which is seen as an ‘existential threat’ for the active community (Millo et al., Reference Millo, Spence and Valentine2025: 8). Despite the perceived gravity of the challenge, however, active intermediaries have tended to double down on existing price discovery claims and engage in ever more elaborate ‘distinction work’, which argues, for instance, that active investing works better for small portfolios and long-term horizons, while replication-based investment worsens bear-market losses (Millo et al., Reference Millo, Spence and Valentine2025: 162). Others resort to justifications ‘in terms of their own individual lifestyle benefits rather than contribution to generating alpha’ (Millo et al., Reference Millo, Spence and Valentine2025: 136, 151). Few in the active community even consider shifting to passive investing for fear of disturbing long-lasting ties with other players and losing their hard-won field position. This inertia translates into a preference for intelligibility over efficiency in the use of technology. Astonishingly, some of the top employees at a major London firm ‘don’t know how to use the computer’ (Millo et al., Reference Millo, Spence and Valentine2025: 108), while in Switzerland, the prioritization of shareholder value over infrastructure investment forces ‘very well-paid people’ to ‘work in relatively threadbare conditions’ (Leins, Reference Leins2018: 51).
Nevertheless, potential technological displacement of the active community will not be the end of financial intermediation. Rather, it will displace the competition for clients’ attention onto the construction of indexes and software design and replace the extractive tendencies of the active community with the oligopolistic tendencies of the passive one (Millo et al., Reference Millo, Spence and Valentine2025: 189). Until the financial industry loses its preeminent position within global relations of power, reflexive and purposive actors will struggle to maintain their position, legitimacy, and relevance despite technological disruption (Ortiz, Reference Ortiz2021: 242). The emerging sociology of financial intermediation teaches us to see these actors as sticky and persistent rather than ephemeral; sui generis rather than functionally necessary or performatively made up; as reflexive, yet prone to slippage into self-referentiality and inertia. It also offers a powerful criticism of the ‘scholastic fallacy’: the tendency to assume that what seems ‘interesting, valuable, and important’ to scholars is also perceived the same way by the actors involved (Millo et al., Reference Millo, Spence and Valentine2025: 33–34), or ‘to identify what’s interesting with what’s important… [assuming] that places of density are also places of power’ (Graeber, Reference Graeber2012: 111).
The biggest omission in the reviewed work – perhaps by design – is its relative lack of historical depth and its absence of a longer-term view of the professional ecologies of finance. To move forward, the sociology of financial intermediation will eventually need to go to the archive. A broader historical perspective might illuminate not only the operations, but also the genesis of financial intermediaries. As William Sewell (Reference Sewell2008: 526–527) notes, capital is characterized by a certain ‘stillness-in-motion’ at its most abstract level, endlessly reproducing the Marxian sequence of M-C-M’ and seeking gains that are ‘necessarily self-canceling’. It thus creates the need for a dialectical counterweight. It seems, therefore, that – until historical capitalism passes into history – niches will be continuously recreated for the animators of capital, acting against this self-canceling dynamic, attempting to reenergize capital’s drive to accumulate through narratives and other devices (Leins, Reference Leins2018: 157–158). The role of intermediaries in the larger structures of capitalist reproduction calls for new, historically grounded exploration.
Open access
