1 Introduction

What is the connection between mass surveillance and institutions of individual agency, freedom, and self-governance? Recent literature on “surveillance capitalism” argues that, over the past two decades, the capitalist system has been transformed by companies such as Google, Facebook, and Amazon, which have commodified personal data for profit. This commodification goes beyond gathering information to directly improve the products provided by the collecting organization and entails using data to predict what people will do, selling that data to third parties, and using that data to modify the behaviors of unknowing consumers. According to critics of surveillance capitalism, this process erodes individual dignity and freedom and threatens democracy.

The formative work in this area is by Zuboff (2015, 2019). She defines surveillance capitalism along three dimensions. The first dimension is the collection of personal data as a by-product of interaction with devices (e.g., cell phones) and platforms (e.g., social media). The second is using computational tools, such as artificial intelligence and machine learning, to predict behaviors. The third dimension is “behavioral modification” (2019: 297), whereby third-party entities use the first two aspects to change how private people act to profit. In this context, surveillance capitalism is a negative-sum relationship between for-profit firms (collecting information and those purchasing it for use) and the members of the populace. For-profit firms engage in “domination” (2019: 10), “expropriation of human experience” (2019: 128), and a “war of extraction” (2019: 157), at the expense of the general populace whose welfare is harmed.

This paper offers an alternative framing of surveillancedescribes attempts at mass manipulation and data collection based on comparative institutional analysis. Following Buchanan (1949, 1954), and others in the tradition of Virginia political economy, we focus on the epistemic and incentive properties of different institutional settings. This approach assumes that people are goal-oriented, irrespective of whether they act in private or government settings. Variation is introduced by how different institutional environments lead to different incentives and different types of knowledge. In the context of data collection, different institutional arrangements will shape how data is collected and used for better or worse.

Our core thesis is that while data collection and attempts at persuasion are present in private and government settings, the welfare effects vary due to institutional differences. We leverage the comparative institutional framework to analyze private and government data collection differences. To differentiate between these two contexts, we follow Mosco (2016: 10) in referring to the former as surveillance capitalism and the latter as the surveillance state. Our analysis sheds light on how data collection and advertising in the private, for-profit sector has different welfare consequences from those in the surveillance state.

To date, the scholarship on surveillance capitalism is interdisciplinary and includes work in accounting (e.g., Andrew et al., 2021), legal studies (e.g., Cuéllar & Huq, 2020; Yeung, 2018), philosophy (e.g., Venkatesh, 2021, Williams & Raekstad, 2022), political science (e.g., Anison, 2022), sociology (e.g., Fourcade & Healy, 2017), and social work (e.g., Garrett, 2022). Our analysis contributes to this literature by offering an economist’s perspective, particularly emphasizing the difference between data collection in private and government settings. In doing so, we connect current critiques of surveillance capitalism to earlier critiques of advertising, and to provide responses emphasizing the often-overlooked benefits of contemporary data collection. We also add insight into the welfare effects of different types of government data collection in the surveillance state. Government data collection can be used as an input into the protective-productive functions of the state, but it can also be used for predatory purposes.

Our analysis focuses on the United States for two reasons. First, the Big Tech companies often associated with the rise of surveillance capitalism are U.S.-based. Second, a mature democracy poses an interesting case for studying potentially predatory surveillance states, given the existence of constraints and the rhetorical commitment to individual freedom and liberty. It is not surprising when authoritarian regimes act predatorily; it is more so when liberal democracies do so.

We proceed as follows. The next section provides an overview of surveillance capitalism, as presented by Zuboff (2019), and discusses the connection to earlier critiques of data collection and advertising. The central concerns of critics of surveillance capitalism reflect earlier arguments about markets, advertising, and consumer manipulation. Section 3 considers surveillance capitalism in the context of private markets. We focus on the institutional features of markets that direct data collection to activities that benefit consumers. In doing so, we also discuss concerns about preference manipulation and information asymmetries, which prevent private users from understanding how their information is being tracked and used. Section 4 analyzes the surveillance state. We differentiate between potentially protective-productive and potentially predatory behaviors in the use of surveillance by state actors. Section 5 concludes with the implications.

2 Surveillance capitalism and its economic precursors

2.1 What is surveillance capitalism

Surveillance capitalism refers to a new economic system, beginning in the early 2000s, “that claims human experience as free raw material for translation into behavioral data” (Zuboff, 2019: 8). This behavioral data is appropriated from private users and sold to third parties for profit. These third parties use the data to manipulate people without them being aware that modifications to their behavior are taking place. The underlying logic is as follows.

Technology firms—starting with Google and later followed by Facebook, Microsoft, Amazon, Apple, and third-party data brokers—realized they could profit, not just from providing services directly to consumers, but also by collecting micro-level data on users (e.g., online purchases and searches, health and activity information) and selling it to third-party buyers. While the collecting entity may use some of the data collected to improve the service they directly provide to consumers, much of the information collected is not, resulting in what Zuboff (2019) labels a “behavioral surplus” (8). Employing machine learning and artificial intelligence, this behavioral surplus is used to predict the actions of private people, permitting third parties to engage in “behavioral modification,” which allows third parties to “shape human behavior towards other ends” (8).

Among the examples that Zuboff offers are Facebook and Pokémon Go. In 2012, Facebook ran an experiment to see whether adding specific features to users’ news feeds—an “I Voted” button, the location of voting booths, and how many other users voted would increase user turnout. They found that those who received this information were more likely to vote relative to the control group (Zuboff, 2019: 298–308). Playing Pokémon Go, an augmented reality mobile game, requires users to search for creatures in different geographic locations. Private retailers partnered with surveillance capitalists to strategically place Pokémon near areas where players were likely to shop to generate foot traffic (Zuboff, 2019: 308–318).

Together, these situations are taken as evidence that private corporations can modify the behavior of the masses in politics and economics. From this, Zuboff concludes that surveillance capitalism is exploitative: “It revives Karl Marx’s old image of capitalism as a vampire that feeds on labor, but with an unexpected turn. Instead of labor, surveillance capitalism feeds on every aspect of every human’s experience” (9).

From this perspective, surveillance capitalism is ubiquitous, parasitic, and directly threatens individual and social well-being. “Surveillance capitalism’s anti-democratic and anti-egalitarian juggernaut is best described as a market-driven coup from above. It is not a coup d’état in the classic sense, but rather a coup de gens: an overthrow of the people concealed as the technological” (Zuboff, 2019: 513, emphasis original). The result is tyranny, where a small group of surveillance capitalists controls the masses. “Surveillance capitalism’s tyranny does not require the despot’s whip any more than it requires totalitarianism’s camps and gulags. All that is needed can be found in the Big Other’s [Zuboff’s term for the overarching digital apparatus used by surveillance capitalists] reassuring messages and emoticons…” (514). Absent citizen pushback, this tyranny will continue to spread in step with technological advances that further empower surveillance capitalists.

2.2 Old wine in new bottles

The concerns with surveillance capitalism can be summarized as a dual information asymmetry problem. The first aspect entails private companies collecting large amounts of micro-level information on private users. These users are unaware of the extent to which their data is being collected or how it will be used. Given the profit-driven and exploitative nature of surveillance capitalism, it is argued that there is an incentive for firms to create and cultivate these asymmetries so that private users remain in the dark. The second aspect is using this data by third-party purchasers to modify the behaviors of private people. Unaware of these manipulations, private actors will change their behaviors compared to what they would do, absent the influence of third parties. These arguments have well-established precursors in the history of economic thought.

Best known for his work on “conspicuous consumption,” Thorstein Veblen offered one of the earliest critiques of advertising. In Veblen’s telling, businesses “gradually fashioned a new form of socially constructed, commercially functional scarcity by dint of advertising—a program of increasingly consequential ‘misdirection of effort’ from above” (Plotkin, 2014: 504). Advertising was exploitative in that it benefited businesses at the expense of consumers through deception—shifting perceptions from use value to symbolic value—and social waste. “Except for line and color, shape and surface, of the containers, and apart from verbal differences in the doctrinal matter which surrounds them, any distinctive character in these various articles of intimate personal use is something very difficult to get at” (Veblen, 1923: 301).

Beyond generating waste, Veblen (1904) believed that advertising fostered monopoly power. “The end sought by the systematic advertising of the larger business concerns is such a monopoly of custom and prestige” (155). Central to this process is the lock-in effects in consumer tastes and habits fostered by advertisers. “The aim of the advertiser is to arrest attention and then present his statement in such a manner that it is easily assimilated into the habits of thought of the person whose conviction is to be influenced. When this is effectually done, a reversal of the conviction so established is a matter of considerable difficulty” (Veblen, 1904: 56, fn 1). The pernicious effects of advertising, therefore, were twofold. In the immediate term, they led to behavioral modifications that did not improve consumers’ well-being. Over the longer-term, they reinforced monopoly power, further allowing firms to leverage their advertising capabilities.

Veblen set the stage for perhaps the most well-known economic critique of advertising by John Kenneth Galbraith. Galbraith’s (1958) argument for the “dependence effect” was that production increases create consumer wants. “As a society becomes increasingly affluent, wants are increasingly created by the process by which they are satisfied … Increases in consumption, the counterpart of increases in production, act by suggestion or emulation to create wants” (158). Advertising plays a key role in this process as “producers may proceed actively to create wants through advertising and salesmanship” (158). For Galbraith, advertising didn’t just influence consumers, but created wants anew for the goods and services being advertised. In his view, advertising undermined claims of consumer sovereignty because tastes, and hence market activity, were made by advertisers—not consumers. Because the wants are artificial, “it can no longer be assumed that welfare is greater at an all-round higher level of production than at a lower one” (158).

Today’s criticisms of surveillance capitalism mirror earlier critiques of markets and advertising. Today, technology firms collect information on people’s past and current decisions. This data, which comes from search histories, social media interactions, location tracking, and smartphone use, among other sources, is processed and sold to third parties for targeted advising. The concern, as in the past, is that these third parties use this data to manipulate consumers—through some mix of influence and want creation—to engage in behaviors that are not in their best interest. In doing so, consumer welfare is reduced while data-collection firms solidify their market power resulting in further welfare losses associated with monopolistic-oligopolistic market structures.

Concerns about privacy and exploiting the human experience are also not new. As Igo (2018) detailed, intense debates about privacy and data collection have existed in America since the late nineteenth century. These debates covered a range of issues and technologies, including journalistic and communication technologies, data-bank technologies, market research activities, workplace testing, and government programs such as social security. Six decades ago, Packard (1964) warned of the “naked society,” resulting from technological advances in widespread data collection, which threatened privacy, individual autonomy, and democracy.

To the extent surveillance capitalism is unique, it is the result of the technological advances that allow for greater micro-level data collection and attempts at prediction, which can then be sold to third parties. While these processes are new, and entail more effective targeting, the fundamental, long-standing concerns associated with data collection and privacy—manipulation and behavioral modification—remain the same. It may be a difference of degree, but not a difference of kind. There appears to be no distinct difference between prior and current discussions of advertising as it relates to the supposed establishment and maintenance of monopoly. As in the past, the welfare effects of monopoly power, data collection, and its use depend on the institutional environment.

Big Tech firms today attain more detailed information about their current and potential consumers and can engage in more effective targeting, as a price of offering free content or services (Barnett, 2018). It should not be assumed, however, that consumers are passive agents within this process who exhibit no control over their choices or the products they elect to buy. On the contrary, consumers continue to play a critical role in the creation and shaping of goods, services, and the brands the provide them. Harper and Enders (2018), for example, note that consumers are not merely recipients of firm-led advertising. Instead, they are co-creators who play a critical role in the development and reception of brands. Importantly, they note that, even within contemporary technological structures, “[l]egal brand owners never have total control over their brands—customer networks often exercise substantial de facto control rights…over the use and transformation of brands” (Harper & Enders, 2018: 1071).

3 Surveillance capitalism reconsidered

In response to the critiques by Galbraith and others, economists began to explore various aspects of the economics of advertising in for-profit markets. These explanations, which remain relevant for understanding surveillance capitalism, focus on the role of advertising in an open-ended world with human ignorance of both the knowable and unknowable.

3.1 Advertising as information

In his groundbreaking paper on the economics of information, Stigler (1961) noted that a given good may have multiple sellers and prices. Absent perfect information, consumers will tend to be ignorant of the full range of prices and suppliers. Information is available to consumers, but it is costly to obtain. In this context, the role of advertising is to lower the cost of consumers obtaining information. “Advertising is, among other things, a method of providing potential buyers with knowledge of the identity of sellers. It is clearly an immensely powerful instrument for the elimination of ignorance” (220). In removing ignorance, advertising reduces the prices paid by consumers. This price reduction can occur through either lower monetary prices for the good itself, through reduced time costs for engaging in purchases (due to more information), or both (see Ekelund & Saurman, 1988: 54). Nelson (1970, 1974) further developed the information aspects of advertising.

This same logic applies to surveillance capitalism. Data-collection firms provide information to third parties who use it for targeted marketing based on predictions of what individual consumers will demand. One way of understanding these activities is that advertisers provide this information to consumers in order to reduce search costs. The fact that there are intermediary data-collection firms that collect and package that information for resale doesn’t change these welfare-enhancing benefits in reducing search costs.

3.2 Advertising as entrepreneurship and persuasion

While advertising offers information to consumers, it does not guarantee that they will pay attention. Because of this possibility, advertising is also an entrepreneurial element, which entails alerting them to the product’s existence and relevant characteristics. As Kirzner noted, “I have not only to produce opportunities which are available to consumers; I have to make consumers aware of these opportunities … An opportunity which is not known, an opportunity to which a consumer is not fully awakened, is simply not an opportunity” (Kirzner, 2018a: 241). This is more than reducing the cost of information as in standard search theory. “It is necessary for the producer to grab the attention of his prospective customers and somehow to get them to see the product he has prepared for them (or at least to see the information available concerning such products)” (Kirzner, 2018b: 249, emphasis original).

Persuasion or “sweet talk,” McCloskey (2021: 30–37) argues, plays a crucial, yet neglected, role in most aspects of economic life. Persuasion is more than the simple transfer of information—e.g., the price of a good or a description of its features. It is the language we use in our regular interactions with others in various settings to foster cooperation. Sweet talk can be deceptive—after all, human beings can, and do, lie—but much of it isn’t as evidenced by personal self-reflection of the variety of honest, positive-sum interactions we engage in with others daily. Advertising and salesmanship can entail the transfer of information, but they are also one manifestation of sweet talk—gaining people’s attention and persuading them of the benefits of a product or service through conversation and human relations.

Surveillance capitalism aims not only to inform, but to alert and persuade. From this perspective, a key component of surveillance capitalism is entrepreneurial and persuasive. Micro-level data allows third-parties to target their efforts at specific consumers in order to persuade them to shift their attention to information about a good or service. Effective persuasion advances consumer welfare by alerting them to information about new opportunities.

3.3 Advertising as competition

Rather than viewing advertising as a source of anti-competitive monopoly power, it can be understood as a crucial part of the competitive process. Hayek (1948) emphasized that competition is not an (equilibrium) state of affairs, but rather a process of discovery and learning. In this context,

Advertising, the activity of capturing consumer attention, is easily seen to be just one particular dimension of competitive activity. Competing advertisers, vying with each other for the consumer’s attention, are inspired to devise attention-grabbing strategies that no one could have foreseen, disseminating information that no one could have known to have been needed (or perhaps even to be available), and stimulating the provision of goods and services the importance to consumers of which no one suspected (Kirzner, 2018b: 252).

From this perspective, advertising is wasteful and anti-competitive only in a world where people know everything they need to know. That is, a world where either they already possess the required information, or they know it is available if they desire to obtain it and is accessible through a process of search. In a world where people face genuine uncertainty, however, advertising contributes to the competitive process of learning and discovery. Consumers learn about opportunities that they were previously unaware of. Producers learn about what consumers value and don’t value.

In contrast to the view that advertising creates a barrier to entry, Demsetz (1982: 50) argued that it is central to fostering contestability. Advertising creates competitive pressures by allowing new entrants and smaller firms to compete with established and larger firms. Advertising both informs consumers of the availability of an alternative product and aims to persuade them to try it. Absent advertising, incumbent firms would have a significant advantage over new and potential entrants because they are already established and known. Demsetz argued that limiting advertising would not foster competition but instead reduce it.

In light of these arguments, surveillance capitalism can be seen as advancing consumer welfare by contributing to the competitive process. Data is gathered and processed for forecasting purposes to target advertisements to specific consumers. The idea of a “behavioral surplus” that is “extracted” from consumers misses the role that learning plays in the competitive process. Consumer data is gathered and processed so that consumers learn about different products and services provided by both incumbents and new entrants. Through this process, incumbents and potential entrants also learn about consumer wants. This coordinating function of surveillance capitalism stands in contrast to the exploitative, anti-competitive view held by critics of surveillance capitalism.

In the past, the concern was that firms would use advertising to establish and maintain a monopoly. Today, the concern is that a small number of Big Tech firms—Google, Facebook, Amazon, and Apple—are using their ability to collect consumer data to maintain monopoly power while also providing data to third parties who can use it for anti-competitive purposes. However, in the past and the future, the contestability inherent in the market process serves as a check against monopoly power. AOL, BlackBerry, IBM, Myspace, and Yahoo!—firms that once held a dominant position in their respective markets—are all examples of the power of competitive entrepreneurship and the importance of market institutions for fostering contestability through time. In more recent times, many alternative social media platforms have emerged to compete with Twitter, Facebook, and Instagram due to some consumers’ dissatisfaction with their content regulation and with the key values displayed by other users on mainstream platforms (See Pew Research Center, 2022).

3.4 Advertising as revealing welfare-enhancing opportunities

Galbraith’s argument about the “dependence effect”—the idea that corporate advertising creates consumer wants (i.e., tastes) for the product being advertised—led to a response by Hayek (1961), who concluded that his position was a non-sequitur. Except for a few base needs, Hayek argued, all human wants are influenced by the cultural environment within which people are embedded. “To say that a desire is not important because it is not innate” Hayek concluded, “is to say that the whole cultural achievement of man is not important” (346).

Although not writing about advertising per se, Boulding (1961) made a similar argument. Human behavior, he argued, is grounded in a person’s unique “image” which reflects their subjective knowledge about the world. From the moment we are born until the day we die, our images are shaped by constant messages, of which advertising is one source, containing information. Many messages do not affect the image of the receiver. Others have marginal effects on their image, while yet others have dramatic effects. Boulding concluded, “The image not only makes society, society continually remakes the image” (64). From the perspective of Hayek and Boulding, it is not controversial to say that advertising influences or creates tastes. It is also not an argument against advertising, since this is one of many information and knowledge sources influencing people.

Kirzner (2018b) argued that critiques of advertising for preference manipulation were grounded in a “closed system” view of the economy. A closed system lacks genuine surprise. In this world, people may be ignorant of specific information, but they are aware that information is available if they wish to incur the cost. In a closed system, one can understand how advertising is viewed as manipulative. In the absence of possible surprise, the consumer is demonstrating their most-valued use of scarce time and resources, and advertisers are attempting to manipulate them away from their first-best path before the exogenously-induced change.

Things are different, however, in an open-ended system where genuine surprise is a possibility. Here, advertising alerts the consumer to an opportunity she was previously unaware that she could even find. “The consumer finds that his world, his range of options, is a little richer than he had dared anticipate” (Kirzner, 2018b: 251). This is not due to psychological manipulation, but to opening the recipient’s eyes to an opportunity they had not previously noticed. Rizzo and Whitman (2020: 17) argue that a crucial part of being rational is experimenting with options, discovering one’s tastes (which can change), and learning through experience. This can only take place in an environment that is uncertain and open to change.

These thinkers argue that, while external factors influenced people, they ultimately maintained agency over their choices. Boulding (1961) argued that one’s image is “characterized by a phenomenal capacity for internal growth and development quite independent of messages received from outside” (26) and that one’s “image is always the property of the individual person’s, not the organization” that sends messages (28). There is empirical evidence to support the existence of individual agency in consumer decision-making.

Studies of “behavioral targeting”—online advertising that targets consumers based on their online histories—find that effective targeted ads make people more likely to click through and purchase the product (Summers et al., 2016).Footnote 1 However, what makes an ad effective is crucial. These studies find that targeted ads are only effective when they align with a consumer’s prior behavior. For instance, if one’s search history shows concern for environmental issues, a targeted advertisement about an environmentally-conscious good makes it more likely that they will click through and purchase the good. But an advertisement about a random good or a good that cuts against environmental consciousness will not affect consumer behavior. What this suggests is that targeted advertisements may amplify a person’s pre-existing image by making them aware of new opportunities that fit with their identity, but they cannot create wants anew as desired by advertisers. This comports with contemporary research suggesting that mass manipulation is much less likely than what is commonly believed. Mutz (2012) highlights a profound disconnect between public perceptions of the power of media influence and the reality. Mercier (2020: 14) describes attempts at mass manipulation as follows. “Far from being gullible, humans are endowed with dedicated cognitive mechanisms that allow them to carefully evaluate communicated information. Instead of blindly following…we weight many cues to decide what to believe, who knows best, who to trust, and what to feel…[T]he recurrent failures of these attempts [at mass manipulation] attest to the difficulties of influencing people en masse.”

While recognizing the multitude of influences on tastes, Demsetz (1968) also recognized the possibility of deception but noted this was possible across settings, not just in for-profit markets. “The formation of wants is a complex process. No doubt wants are modified by Madison Avenue. They also are modified by Washington, by university faculties, and by churches. And it is not at all clear…that Madison Avenue has the advantage when it comes to false claims and exaggeration” (810). The reason is that markets contain mechanisms to ameliorate possible deception (outside legal recourse through courts) through entrepreneurial innovation.

In his formative paper on information asymmetries, Akerlof (1970) pointed out the challenge facing markets lacking perfect information. Using the example of the used-car market, he noted that consumers cannot judge high-quality from poor-quality vehicles. Because of this, there is the threat that the market will unravel because consumers are only willing to pay a price that averages the high- and low-quality cars that they cannot discern. This price drives out owners of high-quality cars who will refuse the lower price which further lowers the average price. This will result in the market collapsing.

Akerlof recognized that this adverse selection problem offered a profit opportunity to the alert entrepreneur to develop what he called “counteracting institutions” (499). Examples include guarantees, warrantees, brand-name goods, and chains. These innovations emerge to reduce information asymmetries about quality and facilitate coordination in the market. They also can contribute to overcoming information asymmetries facing consumers about what data is being collected.

For example, in 2021, Apple introduced a privacy feature through its mobile operating system, whereby users are prompted as to whether they want apps to access Apple’s Identifier for Advertising (IDFA), which assigns a specific code to each mobile device for data collection and marketing. This, as well as other privacy features such as Mail Privacy Protection, were introduced by Apple’s advertising campaign to inform consumers and to differentiate Apple products from competitors. These features by Apple are in addition to the numerous third-party tracking blocker apps and browser extensions that exist on desktops and mobile devices to limit data collection. These trends represent the market’s contestable character that incentivizes incumbent firms to adopt counteracting institutions that reflect and appeal to consumers’ privacy wants.

Additionally, consider the emergence of various encrypted Virtual Private Networks (VPNs), offered by for-profit organizations to encrypt internet activity and online identity. Typically, an internet service provider (ISP) keeps track of an internet user’s internet protocol (IP) address. The IP address, used to determine a user’s identity, is then associated with the user’s internet activity. Behavioral capitalists may then buy or collect data about a user’s interests for advertising purposes. Encrypted VPNs, run by the private network host, hide users’ IP addresses by redirecting internet traffic to a configured remote server. The redirection of internet activity to the private network host means the VPN server appears to be the origin of the data, making it impossible for third parties to monitor or collect data about the specific user’s internet activity. VPNs have been used across the globe for identity protection purposes and can make third-party data collection on identified individuals more difficult. Moreover, personal data removal services offer consumers the service of contacting data holders to remove their customers’ sensitive and private information from the internet.

Also, consider the emergence of secure internet browsers that limit identification and data collection. For example, Brave Browser, launched in 2016, is a browser that blocks phishing, browser fingerprinting (gathering information from a user online to construct a unique identity), tracking and advertisements, and allows users to clean history and cookies after every session. Brave Browser has a feature of “debouncing,” which prevents attempts to overcome restrictions on third-party storage in privacy-focused browsers, otherwise called “bounce tracking.” Other secure web browsers that offer similar features of blocking phishing, fingerprinting, tracking, and advertisements, include Duck Duck Go, Waterfox, Epic Privacy Browser, and Vivaldi. Duck Duck Go also has a secure search engine that can be accessed on mainstream internet browsers. These innovations suggest that even where information asymmetries exist—i.e., users know that their data is being collected but lack specifics—there are entrepreneurial market forces at work that will offer consumers increased control over how their data is collected.

Finally, entrepreneurial innovations also play a key role in reducing deception and are at work in both easier and harder cases of possible deception, with the difficulty dependent on the nature of the good as emphasized by Nelson (1970) and Darby and Karni (1973). Easier cases include “search goods,” which refer to goods whose features can be directly evaluated by a consumer before purchase. For these goods, consumers can directly observe whether the product’s features align with advertised features.

Harder cases include both “experience goods” and “credence goods.” Experience goods are goods which a consumer can evaluate the good only after experiencing it, as in the case of a vacation, while credence goods are those for which a consumer cannot easily evaluate quality even after purchase because of a lack of relevant knowledge, such as in the case of repairs to specialized equipment. Markets address these harder cases through competition and consumer feedback. Rating portals, such as Yelp and Google Reviews, and word of mouth, as seen with direct discussions with others or community discussion boards, combined with multiple providers, serve as checks on deception through experimentation, learning, and information sharing.

Surveillance capitalists operating in liberal democracies face institutional constraints that limit their ability to engage in legal coercive force against consumers. This introduces two major checks on opportunistic behavior. First, freedom of entry by competitors and freedom of voice by consumers check the ability of entrepreneurs to deceive and influence consumers. These freedoms allow for consumers and market competitors to challenge manipulation and deception. Further, the existence of legal institutions—due process and state enforcement of judicial decisions—offers formal recourse against entrepreneurs engaging in false advertising or fraud. Second, the protection of person and property means that surveillance capitalists cannot prevent, via force, the emergence of counteracting institutions, such as VPNs and secure browsing technologies, that allow users to protect their information from collection.

4 The surveillance state reconsidered

The surveillance state refers to activities taken by governments to monitor their citizens. To understand the potential welfare effects of the surveillance state, we leverage Buchanan’s (1975) distinction between the protective-productive state and the predatory state. The protective-productive state enhances the welfare of citizens by protecting their constitutional rights and providing value-added services. The predatory state reduces citizen welfare by violating the rights of citizens or by extracting resources from citizens to benefit a small group of politically-connected elites. This harms individual agency, freedom, and self-governing democracy.

4.1 The protective-productive surveillance state

The provision of government services entails the collection and processing of information. In many cases, this information is provided by citizens. For instance, when citizens file their taxes, they provide identifying personal information and details on their various forms of income. But governments also collect data on citizens without their knowledge. In principle, this can be used for protective-productive purposes.

One example is smart cities, which refer to technologically integrated urban spaces. Smart cities collect large amounts of micro-level data from various sources—including citizen activity—which is processed and used, often in real-time, to improve the provision of goods and services. Data collection and processing can be used to improve the provision of utilities, critical infrastructure, transportation, and public services (Joshi et al., 2016; Samih, 2019).

Another example is the national-security surveillance state, which affects the entire populace of the nation. Surveillance of the general populace, it is argued, is crucial to preventing crime and threats to fundamental rights and institutions and pursing perpetrators who do so. This logic was used to justify the U.S. government’s domestic and foreign surveillance programs undertaken as part of its “war on terror.” To the extent that mass surveillance is effective—meaning it is limited in use and effective in data collection—it can serve as an input into the security and policing activities of the welfare-enhancing protective-productive state.

What about advertising? In the realm of the productive state, governments use advertising to inform current and potential citizens about candidates for office and about available public services. Political campaigns often partner with social media and Big Data firms to maximize the impact of political advertisements (see Lord & Potter, 2015). Proponents of these activities argue that targeted advertisements better inform citizens, allowing them to best satiate their preferences through the political process. Advertising can also be a source of competition between jurisdictions—e.g., a jurisdiction advertising its unique features to attract tourists or new residents away from competing jurisdictions. Interjurisdictional competition can potentially discipline governments and incentivize them to meet citizen preferences (Kenyon, 1997).

4.2 The effectiveness of the protective-productive surveillance state

The effectiveness of the protective-productive surveillance state in contributing to citizen welfare is a matter of public administration, which refers to the generation and implementation of public policy. As Boettke (2018: 949) noted, “property, prices, and profit and loss are absent in government activity, so other mechanisms are in operation to serve as proxies for the functions of incentivizing, guiding, luring, and disciplining decisions.” This means that the political process replaces the market process.

Competition exists in both the market and political contexts, but as Wagner and Yazigi (2014) contend, the selection mechanisms generate very different outcomes. Markets, they note, “are the domain of logical action where different producer offerings can be compared within an environment of voluntary exchange” (513). Here the term “logical” is borrowed from Pareto (1935), who defined a logical situation as one where performance can be gauged as a ratio of revenues to expenses. In contrast, “[p]olitics is the domain of non-logical action where there can be no direct comparison of options by referring to prices in relation to qualities” (Wagner & Yazigi, 2014: 513). Because political activity cannot rely on market prices and profit and loss, it cannot leverage the economic knowledge that emerges through the exchange process and is unavailable in its absence (Hayek, 1948; Mises, 1951).

Beyond these knowledge issues, public choice scholars have documented a variety of frictions and perverse incentives in democratic politics, which can result in inefficiencies. These include a variety of frictions and perverse incentives, including rational ignorance of voters, vote bundling, the limitations of a single vote, the timing between elections, principal-agent problems between voters and elected officials, the logic of special interest groups in spreading costs and concentrating benefits, and the incentives facing elected officials and bureaucrats to secure benefits now and shift costs onto others in the present and future (Buchanan, 1954; Reksulak et al., 2014; Tullock, 1965; Wagner, 2016). The incentives in politics can weaken democratic accountability and the link between political action and citizen preferences, eroding voter agency and self-governance. Importantly, these factors apply to national and local politics, albeit in different manifestations, which can undermine the ability of interjurisdictional competition to discipline political actors (see Boettke et al., 2011).

Also relevant is the type of good that is provided in political settings. In markets, there is a mix of search, experience, and credence goods. Many political goods are credence goods, making it difficult for voters to discern the quality of the product they are forced to pay for via taxes, whether they want it or not. “Political goods, described as different bundles of public goods and rhetorical platforms, are also a subset of the credence goods category as their quality cannot be determined after election” (Wagner & Yazigi, 2014: 508). For instance, politicians talk about the benefits of large-scale stimulus investments or national-security spending, but there is no way for citizen-voters to discern the counterfactual and accurately gauge the quality of the political product being provided or the specific factors that produced the observable outcome.

Together, the epistemic and incentive aspects of political institutions and the type of good produced will influence the behaviors of those in the protective-productive surveillance state and whether data collection, advertising, and persuasion improve, or reduce, citizen welfare. The exact manifestation of these factors will depend on the specific location of the relevant political institutions, the structure of those institutions, and the demands placed on those institutions. These factors will determine the extent to which political actors govern “over” citizens—i.e., top-down government by political elites—or “with citizens”—i.e., a self-governing system grounded in the art of association, individual agency, and self-determination (see Boettke, 2018: 952). Where these factors drive a wedge between citizen preferences and the behavior of political officials, it creates space for potential opportunism—narrow self-interest—by those in power, which is a core feature of the predatory state.

4.3 The predatory surveillance state

The paradox of government is that a government strong enough to protect property rights is also powerful enough to undermine those rights. Government surveillance can potentially be used to advance citizen welfare through protective-productive activities. However, it can also reduce citizen welfare while eroding individual freedom and agency. As Brennan and Buchanan (1985) note in discussing the granting of powers to the state, “[i]n the assignment of these powers, problems of control may arise, problems that are not amenable to easy solution. Once established as sovereign, government may not willingly remain within the limits of its initially delegated authority” (31). One recent example of this, which has deep historical roots, is the extensive surveillance of American persons, typically without a warrant and in violation of existing legislation, as part of the U.S. government’s “war on terror” (Bamford, 2008; Granick, 2017).

The activities of the national security state are especially conducive to potentially predatory behaviors. As Dorsen (1989) emphasizes, “Foreign affairs, and its close relation national security, has been a graveyard for civil liberties for much of our [the United States’] recent history” (840). To illustrate how the national-security surveillance state can undermine individual agency, freedom, and self-governance, consider the case of COINTELPRO, a covert surveillance program run by the Federal Bureau of Investigation from the 1950s through the 1970s. The domestic program infiltrated and collected information on members of various groups, including peace and anti-war activists, civil rights leaders, and political parties, among others, to influence and limit their actions. The U.S. Senate Select Committee to Study Governmental Operations (1976), which investigated these operations after they were made public, concluded that,

COINTELPRO began in 1956, in part because of frustration with Supreme Court rulings limiting the Government’s power to proceed overtly against dissident groups … the Bureau conducted a sophisticated vigilante operation aimed squarely at preventing the exercise of First Amendment rights of speech and association … Many of the techniques used would be intolerable in a democratic society even if all of the targets had been involved in violent activity, but COINTELPRO went far beyond that (3).

This isn’t a single occurrence. Following the September 11, 2001 attacks, Muslim Americans were broadly surveilled without cause, with some studies suggesting this led members of that community to alter their daily behaviors out of fear (Sidhu, 2007). The national-security surveillance state has unique institutionally-driven features that make it conducive to predation and differentiate it from surveillance capitalism.

One is that national security is carried out by the national government, which maintains a monopoly over security activities. This reduces the threat of exit by citizens as a disciplinary mechanism. It also concentrates political power, which is discretionary and extra-constitutional, in the hands of a small group of people, the justification being that those in charge need to be able to address unique and unforeseen circumstances quickly. The powers granted to members of the national security state are intended to control other human beings through direct and indirect manipulation and violence. In this regard, the activities of the national security-surveillance state are fundamentally different from the use of surveillance technologies in private markets, which are to make consumers aware of an array of alternative products and to persuade them to voluntarily choose one option over others.

In principle, this political power will be checked by the courts, voters, and legislature. But, historically, that has not been the case. As Corwin (1947) noted that “in total war the Court necessarily loses some part of its normal freedom of decision and becomes assimilated, like the rest of society, to the mechanism of national defense” (177). Further, citizens and the legislature face information asymmetries that limit their oversight capabilities (Coyne et al., 2019).

The activities of the national-security surveillance state are purposefully shrouded in secrecy in the name of protecting national-security secrets. At the same time, this secrecy allows those within the surveillance state to keep the nature of their actions hidden from voter and legislative scrutiny. This isn’t a matter of rational ignorance, in terms of voters not obtaining information because the perceived costs outweigh the benefits; instead, it is the inability of citizens to obtain the relevant information if they desired to do so. Moreover, congressional oversight can be circumvented as illustrated by the U.S. government’s war on terror.

In 2005, The New York Times reported that the Bush administration had authorized a secret surveillance program through the National Security Agency (NSA), which engaged in warrantless monitoring of calls between those in the United States and those in foreign countries (Risen & Lichtblau, 2005). The Bush administration classified this program as a “covert action” instead of as “intelligence activities” (Kitrosser, 2007), which meant they only had to inform a small group of congressional leaders (the “Gang of Eight”) rather than the full congressional intelligence agencies. The spirit of this more limited classification (“covert action”), intended to be used only in extraordinary cases, is to inform leaders of Congress instead of seeking approval. Even after the surveillance program was made public in 2005, there were few calls for congressional oversight because Republicans controlled both the legislative and executive branches, reducing the likelihood of legislative pushback in national security (Sugiyama & Perry, 2006).

Beyond this single case, Zegart (2011) argues that weak congressional oversight of intelligence activities results from a lack of expertise and insufficient budget power over intelligence community members. This creates significant space for members of the Intelligence Community to act with discretion, absent checks to ensure alignment with the citizenry’s well-being. Together, the limits on citizen and legislative oversight reduce the effectiveness of voice as a mechanism for disciplining the behaviors of those in the national security surveillance state.

Another key institutional feature of the national-security surveillance state is entanglement between government and private institutions. These entanglements take on two forms. One form of entanglement is parasitic, with state institutions taking the data collected by private firms, sometimes knowingly and other times not, for their use. This method has a long history in America that well pre-dates any notion of surveillance capitalism (see Hochman, 2022). These parasitic entanglements were evident in the public revelations by Edward Snowden regarding the U.S. government’s surveillance activities during its war on terror. Among other things, Snowden revealed that: (1) The PRISM program existed to request data from at least nine major technology companies (including Apple, Google, and Microsoft), which were required to comply, (2) The National Security Agency (NSA) tapped directly into the data centers of major data companies (including Google and Yahoo) without their knowledge, (3) The XKeyscore program allowed the NSA to search through comprehensive databases of metadata and online content without authorization, and (4) The NSA actively weakened internet security, which included requiring technology companies to install “backdoors” to ensure widespread government access (Greenwald, 2014).

Under this parasitic scenario, private data firms become purposeful or unwitting appendages of the national-security surveillance state. The data these firms collect to facilitate coordination in private markets becomes a tool for broad-based state surveillance and tracking of private persons for purposes of control (see Cox, 2020; Cameron, 2023; Davis, 2023). This fundamentally changes the nature of surveillance capitalism activities and the concomitant benefits discussed in the prior section. In the context of the national-security surveillance state, the concerns about privacy and abuse raised by critics of surveillance capitalism come to the forefront due to the purpose of data collection (general surveillance and control), the capabilities of government (the ability to access data from private firms), and the aforementioned gaps in oversight of these powers.

The second form of entanglement is through government funding for the services of private firms. The military funding of technology firms has a long history in the United States (see Saxenian, 1996; Nicholas, 2019; González, 2022). This funding offers a source of revenue outside of private markets and creates an environment conducive to “political capitalism,” characterized by rent seeking, rent extraction, and regulatory capture (Holcombe, 2018). Big Tech companies such as Amazon, Palantir, Google, and Microsoft rely on military contracts for part of their revenue. Some companies have created standalone divisions—e.g., Google Public Sector—to enhance the company’s efficiency in securing government contracts.

Among the effects is a shift in the nature of competition from market activity to political activity. Firms that successfully build political relationships and secure contracts also have an advantage over new entrants. Government demand for private services is limited by budgets for the relevant period. Moreover, demand is likely to be relatively price inelastic due to incentives in political institutions—e.g., contracting process, disincentive to save, and absence of a clear residual claimant. This means that new entrants will not be able to rely on expansions in the market by competing on price as in private markets; instead, they need to displace established incumbents to secure a portion of fixed government funds. This will limit the competitive pressures that are inherent in private markets.

Since these entangled private firms straddle the public–private fence, these reductions in competitive pressures adversely affect private markets. What appears to be a lack of competition in Big Tech due to the existence of large-scale data firms may very well be the result of government privilege. For example, “Amazon has attracted a great deal of scrutiny for its enduring march toward retail monopoly, but in reality, the company’s future now rests as much in securing the top-secret data of CIA activities as it does in delivering groceries—a spectacular broadening of its reach…” (Silverman, 2018).

Government funding shifts the alertness of for-profit firms from satiating the wants of private consumers to satisfying political gatekeepers, while insulating them from market competition and allowing them to scale beyond what would otherwise be possible. The broader welfare effects of an entrepreneur relying on the political means of earning wealth are fundamentally different from the economic (market) means of earning wealth; the former contributes to increasing overall wealth, while the latter redistributes wealth (Olson, 1982).

Finally, in contrast to surveillance capitalists, the surveillance state can engage in legal coercion, resulting in two major consequences. First, the surveillance state can use legal coercion to stop competition and limit cultural influences that compete with the narratives or information it provides, creating barriers against challenging manipulative or deceptive information. For example, the Chinese government extensively censors the Internet (Fedasiuk, 2021). The Biden administration pressured social media companies—Facebook, Twitter—to censor COVID-related content that it deemed to be disinformation (Klar, 2021). Second, the surveillance state can use legal coercion to restrict and suppress the emergence of counteracting institutions meant to circumvent governmental surveillance. Examples include the shuttering of Silk Road (a darknet marketplace) by the U.S. Federal Bureau of Investigation (Flitter, 2013) and the Chinese government’s punishment of VPN providers (Feng, 2021).

5 Conclusion

Data collection exists both in private and government settings; it is not inherently “good” or “bad.” Whether it enhances or reduces citizen welfare depends on the institutional context within which data is collected and used. The rise of surveillance capitalism does not change this, although it reflects technological advances that allow for greater data collection and targeted advertising. The key issue is whether effective feedback mechanisms, and the incentive to act on that feedback, exist in the different contexts to discipline circumstantially welfare-reducing activities like surveillance (Boettke & Coyne, 2022).

Applying the comparative institutional framework to surveillance has several implications for the welfare of the surveilled. First, in the contemporary context, surveillance capitalism and the surveillance state raise issues of weak agency, which refers to the ability of people to act independently to pursue their well-being and interests (Collins, 2022). Weak agency is relevant because the surveilled lack essential knowledge about how their behavioral surplus is used, even though it may indirectly affect them (Satz, 2010: 96). Even though many internet users do not consciously read the terms and conditions of the platforms they use, there is reason to believe that weak agency is likely to be less extreme in surveillance capitalism for two reasons.

First, at the minimum surveillance capitalists offer disclosures regarding their use of consumer data. Second, there is rivalry in private markets which creates space for innovation through time to increase the agency of private people. Comparatively, the existing public choice literature provides many reasons to think that principal-agent problems are pervasive in government, including the operations of the surveillance state. The secretive character of the surveillance state for national security purposes means that users may be surveilled without knowing or consenting to any terms or conditions. Moreover, the national-security state has a monopoly control of the surveillance states which prevents rivalry.

The second implication relates to how privacy rights are defined and enforced, which will influence data collection by private and government entities. The weak agency in the context of surveillance capitalism results from the lack of detail regarding digital property rights. Therefore, it is not legally clear how or whether surveillance capitalists should be penalized for using user data. In a context where digital property rights are established and enforced by government, weak agency in surveillance capitalists’ markets would diminish since capitalists would theoretically be penalized for property rights violations. However, the secretive and monopolistic character of the surveillance state limits its vulnerability to any feedback mechanisms from citizens, making it unlikely to be discovered or penalized for welfare-reducing operations even with digital property rights. Thus, there is less of a theoretical reason to conclude that weak agency in markets where behavioral surplus is sold to governments would diminish.

Third, there is reason to believe that the harmful consequences of exchanges in government markets are magnified and sticky, relative to purely private markets (Alshamy et al., 2023). The reason is that private institutions cannot independently legally engage in violence or coercion to execute their goals, whereas government institutions can. So, the entangled exchange relationships between private and government institutions in the United States exacerbate the potential for harming individuals in ways that exchanges between surveillance capitalists in purely private exchange cannot.

Finally, the comparative institutional framework can help someone to think about how we might address the possibility of the predatory state, given the prevalence of surveillance. For instance, Brinn (1998) argues that there is no way to avoid advances in surveillance technologies, which are ubiquitous. The default response often calls for laws restricting surveillance for private use with enforcement by political authorities. But this can create an illusion of privacy—per the logic of the predatory state—since authorities have access to the surveillance capabilities. Brinn argues that the solution is not less transparency, but more in the form of “reciprocal transparency.” In this scenario, citizens can watch authorities just as authorities can watch citizens. If secrecy is the rule, then the elite will still be able to surveil; if non-elite can surveil the elite, however, it serves as a powerful check on abuses of power. The comparative institutional framework can shed light on the feasibility of diverse approaches to surveillance and privacy.