Protest of the World Trade Organization in Seattle, 1999
There’s a relatively familiar criticism of globalization that goes: unrestrained globalization is bad because it undermines national sovereignty and forces poorer countries to do things that aren’t in the best interests of (most of) their people. In its most sophisticated form, this criticism is associated with people like Joseph Stiglitz (Globalization and its Discontents) and Dani Rodrik (The Globalization Paradox). Recently, in assessing the effects of trade with China on post-industrial parts of the United States, many commentators (and both political parties, to varying degrees) have generalized the criticism, dropping the word “poorer.” Globalization can be bad for any country.
If you think about it, though, this critique shouldn’t be unique to globalization. It sounds like a specific case of the broader charge that market outcomes are often undesirable and should be regulated. In domestic politics, this obvious fact has been acknowledged (in general, if not in many particular cases) since the British Factory Acts of the early 19th century. Collective oversight of our international relationships is comparatively underdeveloped. As David Grewal puts it: “The central tension in contemporary globalization is that everything is being globalized except politics.” In this light, he observes, debates about globalization are really just extensions of the old debate about whether it is better for people to engage with one another as private individuals or, by contrast, as citizens with equal shares in a collective decision-making process.
Grewal’s book Network Power is an effort to give us a better framework for talking about globalization and connecting it to other forms of coordination (like markets and democracies). He defines globalization as “the rise to dominance of shared forms of social coordination or international standards that enable us to coordinate our actions on a global scale.” Prominent examples of such standards include: the English language, the gold standard, the World Trade Organization’s membership rules, and multilateral treaties like the failed Multilateral Agreement on Investment of the 1990s. The rise of such standards is often “double-edged,” Grewal observes, in that it allows diverse participants to cooperate through one channel while threatening the elimination of all others. Coordination standards become more valuable as more people use them, so it’s common to see a sudden exodus from a losing standard (e.g. a dying language) as everyone clusters to the presumptive winner. You’re just describing network effects, you might interject, and indeed that was one of my early reactions to the book. But I think what Grewal is really doing is turning network effects political.
A few chapters in, Grewal reveals his true ambition: to intervene in the oldest debate in social theory (itself the child of the oldest debate in philosophy), the one between structure-focused and agency-focused accounts of social life. The old way of thinking about power was straightforwardly agentic: I exercise power over you by making you do what I want. This is sometimes called the power of sovereignty, including by Foucault, who famously sought to replace it. Foucault argued that the more important form of power in modern life isdisciplinary power, which is diffuse, invisible, and all-pervasive, enacted not by the sovereign but by our collective circumstances, norms, and even our consciences. Grewal insists that we must make more room for individual action than this.
The problem is not with agency, he argues, but with assuming that relationships of power must always resemble political sovereignty. In the case of globalization, they distinctly do not. The concept of network power can “help us see how individual actions can create structures [e.g. networks] that in turn limit individual agency in a way that resembles the more familiar exercise of power by one person over another.” This type of power relationship he termssociability. Sociability is voluntaristic; it’s raw human interaction (fighting, cooperating, trading) prior to the construction of a political sphere. With this concept in hand, Grewal can give his best definition of globalization: that which “extends and deepens relations of sociability at a global level without the concomitant construction of a global sovereignty.”
The problem with network effects, in Grewal’s view, is that they give adopters little choice but to participate. He cites the philosopher Serena Olsaretti, who wrote: “A choice is voluntary if and only if it is not made because there is no acceptable alternative to it.” And once a standard is selected, there is no going back. The linguist Ferdinand de Saussure may have been the first to recognize this: “Once the language has selected a signal, it cannot be fully replaced by any other…The community, as much as the individual, is bound to its language.”
Grewal says there are two types of consequences of being forced into a popular network: interest concerns and identity concerns. These roughly map onto Charles Taylor’s phrase, “Freedom to get what you need / freedom to become who you are.” Interest concerns arise when less powerful actors are forced to accept a bad deal because it’s the only one on the table. This is one way to characterize the widespread criticism of investor-state dispute settlement (ISDS) provisions in trade agreements (like TPP), which give multinational companies the opportunity to sue countries for harming their investments (e.g. by enacting environmental protections). As much as lefty Americans complained about the ISDS clause in TPP, I bet the American trade representatives were the ones insisting on the clause. Lots of multinational companies are owned by American investors. I suspect the government of Peru, for example, would much rather have signed a version of TPP without any ISDS.
Identity concerns seem even more fundamental to the working of network power. Because networks are often winner-take-all, every alternative option will lose out. Few of us mourn Friendster and MySpace, but we do lament the decline of local and regional languages. In general, Grewal observes that globalization tends to enhance some forms of universalistic recognition (notably, the idea of human rights) while diminishing more local, solidaristic “communities of value.”
How to remedy the consequences of network power? One common approach you see is the effort to withdraw from networks. Grewal argues that this is unrealistic and even counter-productive: people (and countries) only enter these networks in the first place because they need each other. Instead, he concludes that the best bet is to provide multiple alternative channels of access. One way to build alternative channels is to make networks non-exclusive. Another way is to make networks compatible with one another, like how any two languages can be mediated by translation. In a world with ubiquitous, low-cost translation services, no one would need to give up the language of their grandparents.
Network Power includes commentary on topics you would expect like international trade agreements, investment standards, and technological standardization. On globalization, Grewal praises Dani Rodrik as one of the best critics of hardcore free trade ideology, which treats trade as an end in itself rather than a means to development. This reminded me to pre-order Rodrik’s upcoming book, Straight Talk on Trade. But in general, I found the theoretical framework I’ve been discussing to be the most interesting, original contribution. I’ve been exposed to social network analysis in a few classes, and one debate that always arises is whether there is any theory of networks or just a method of characterizing networks, nodes, and their position therein. Grewal doesn’t speak much to the literature on networks from sociology, probably because that literature is very concerned with an agent’s exact position within a network while Grewal is mostly concerned with whether you’re in or out. And while it may seem simplistic, I think treating network membership as dichotomous helps unlock much of Grewal’s insight. If the costs of staying out are high, people will sacrifice a lot to get in.
Grewal’s theory of networks is clearly applicable to antitrust policy in the context of platform monopolies: Google, Facebook, Amazon, and perhaps Uber. Each giant has achieved critical mass in its domain, initially for intrinsic reasons (quality) but now increasingly for extrinsic ones (it’s where all your friends/customers/drivers are). In the past I’ve discussed the idea of regulating these platforms as public utilities, e.g. by requiring nondiscrimination between their own products and others or by imposing common carrier obligations. Using Grewal’s framework, this approach would address interest concerns–everyone would get equal access to the dominant networks–but not identity ones. It is tempting to conclude that diversity is not a realistic value when it comes to platforms; that they work best when everyone gets together on a single protocol.
The parallel to languages and translation suggests one way out. Imagine a world with ten ride-hailing services, all linked on some underlying level thanks to data sharing and an API (indeed, providing the underlying infrastructure is the key to all of Amazon’s businesses). You might prefer to use a particular service–the one that pays it drivers best, the fancy one, the one that specializes in pooling–but if there were no cars available you could seamlessly expand your search to all ten, or some subset. Like with language translation, we might prefer to use our idiosyncratic protocol, but we also want access to everyone within arm’s reach. My ten-Uber hypothetical is probably flawed on many levels, but I share it as an example of how this concept of accessibility helps us think beyond the seemingly inevitable winner-take-all dynamics of networks.
1. Prime time for centralization
Speaking of monopolies, here is Stratechery on the Amazon / Whole Foods acquisition. It is the only thing you need to read to understand the logic of the deal. Amazon is not so much buying a retailer as buying an anchor customer for its impending Food Supply Chain Logistics business. Once it builds out the back end of a food supply chain well enough to service Whole Foods, it will start selling to other grocers, direct to consumers, to restaurants, to caterers, etc. This is how Amazon functions in every market: build out a service (Prime, AWS) and then open it up to third parties. So from an antitrust perspective, the FTC should ignore the red herring of centralization in the grocery industry (the much-repeated stat is that Whole Foods only sells 2% of groceries) and think about centralization in the food supply chain.
2. Dream of Cali jurisdiction
I follow a bunch of people on Twitter who share my concerns about monopoly power. Many of them were disappointed by this Supreme Court decision limiting the ability of plaintiffs to sue a company in a given state’s court even if most of them were allegedly injured in a different state. A bunch of people–most of whom are not California residents–sued Bristol-Myers Squibb in California court after being injured by a blood thinner. The Supreme Court held 8-1 that “what is needed is a connection between the forum and the specific claims at issue,” and that there is no such connection between these people and California.
Justice Sotomayor, along with my Twitter friends, bemoaned the likelihood that this decision will make it impossible for far-flung defendants to hold corporations accountable for their nationwide activities. I am inclined to agree, and so this case seems one of those frustrating examples where the desirable outcome can’t be justified by the law and precedent. Sotomayor acknowledges this in taking a realist tack: “What interests are served by preventing the consolidation of claims and limiting the forums in which they can be consolidated?” The majority reassures us that plaintiffs can still bring a mass action against a corporation in its “home” state–in this case New York or Delaware. In response, Sotomayor raises several troubling questions: Where can plaintiffs sue multiple defendants who are headquartered in different states? What about a nationwide mass action against a defendant not incorporated in the U.S. at all?
3. Racism for realists
Scott Alexander defines liberalism as “a technology for preventing civil war.” “Popular historical strategies for dealing with differences have included: brutally enforced conformity, brutally efficient genocide, and making sure to keep the alien machine [liberalism] tuned really really carefully.” This comes in the context of a controversial essayabout racism. I’m tempted not to share the post because I think it minimizes racism, but it’s important to talk about this stuff. On a charitable reading, I think there’s a valuable idea at the conclusion of the essay. Basically: we should avoid defining racism as pure evil, unintelligible, and unresponsive to reason. Because once you define it that way, violence is the only appropriate response. It’s more useful to treat racism as something very human, present in all of us to varying degrees, that can be changed and mollified over time. We might call this the “realist” approach to racism. I’ve seen it articulated by a number of anti-racist writers; one example is Beverly Tatum, famous for the “smog” and “moving walkway” metaphors. Thanks also to Kate Selker for helpful comments on this view of racism after a long-ago newsletter. I’d be interested to hear your views.
4. Pedagogy and pasttimes
Wonderful old essay from Alan Kay on science education and the way our pedagogy does a better job inculcating an appreciation for stories than an appreciation for logic. On being told that it’s too hard to get children thinking scientifically, he says it’s harder to learn to hit a baseball. “In fact, what really seems to be the case is that children are willing to go to any lengths to learn very difficult things and endure almost an endless succession of “failures” in the process if they have a sense that the activity is an integral part of their culture.”
5. Capital allocation is a myth, volume 481
John Kay: “The paradox of modern capital markets is that although there is less and less need for market activity from the point of view of either the end users of finance, or the investors who are the ultimate beneficiaries of finance, the volume of market activity has increased exponentially.” In this context, it was funny listening to Will Thorndike talking about the most effective CEOs of all time (from a financial perspective) on the investing podcast. In his book, Thorndike profiles eight CEOs who excelled at capital allocation, meaning they were very good at timing dividends and buybacks. There was almost no talk of CEOs making smart capital investments–i.e. the stuff that financial markets are supposed to be facilitating. And to the extent that Thorndike praised acquisitions, it was exclusively for finding opportunities to “reduce headcount.” None of this is how a society advances in the long run.