When Amartya Sen was ten years old, a desperate day laborer burst into his family’s garden, screaming in pain and pleading for help. The man, named Kader Mia, had been stabbed by some local thugs for the crime of being a Muslim in a Hindu neighborhood. As Sen and his father took him to the hospital where he would shortly die, Kader Mia explained that his wife had warned him not to go looking for work in a predominantly Hindu area, but there had been no other option. Sen tells this story eight pages into Development as Freedom, and extracts this lesson:
The experience was devastating for me. It made me reflect, later on, on the terrible burden of narrowly defined identities, including those firmly based on communities and groups….But more immediately, it also pointed to the remarkable fact that economic unfreedom, in the form of extreme poverty, can make a person a helpless prey in the violation of other kinds of freedom. Kader Mia need not have come to a hostile area in search of a little income in those terrible times had his family been able to survive without it. Economic unfreedom can breed social unfreedom, just as social or political unfreedom can also foster economic unfreedom.
Don’t worry, this isn’t That Kind Of Book, where every other page contains a tenuously relevant anecdote meant to Add A Human Touch. It’s hard to say exactly what kind of book this is. Sen claims to aim at a popular audience, but Development as Freedom lacks the highly-produced punchiness of most contemporary popular nonfiction. This isn’t Freakonomics, or even Why Nations Fail.
I’m not even sure this book is really about economic development, although I’d be a bit pressed to say what its true subject is. There comes a point in an academic’s career when they have finalized their library of favorite facts, ideas, and frameworks. From that point on, their intellectual output is like the menu at Taco Bell, each item a combination of the same basic ingredients in varying shapes and proportions. Amartya Sen had reached this point in his career when Development As Freedom was published. The year before, in 1998, he had won the Nobel Prize in Economics (sorry, the Bank of Sweden Prize in Economic Sciences in Memory of Alfred Nobel), which is exactly the sort of event that precipitates this transition.
The book originated in a series of lectures Sen gave at the World Bank in the mid 1990s, which he later decided to expand and publish as a book aimed at general audiences. This was the high post-Cold War period, when international development was on many people’s minds. History had just ended, after all, and what remained after the battle against communism was over was simply to make the prosperity and freedom of the first world universal.
This was also a time when words like “neoliberal” were acquiring their modern pejorative meanings. The World Trade Organization meeting in Seattle would be beset by anti-globalization protesters in 1999. People started to complain about the “Washington Consensus” imposed by organizations like the IMF and World Bank on struggling developing countries. Worries about job destruction had led to Ross Perot and protests of NAFTA, activating a strain of populist politics that would eventually lead to Donald Trump.
In the preface, Sen notes that he has had serious public differences with the World Bank, but it’s clear which of these two sides he feels more closely aligned with. Development as Freedom is neoliberal in a way that anticipates that word’s recent reclamation by the wonkish center-left (catalyzed by people who post economics memes on Reddit).
Getting to the putative subject of the book, Sen claims that his aim is to give a definition of development and a discussion of how to achieve it. His first pass at a definition is that development is “a process of expanding the real freedoms that people enjoy.” This is meant to contrast with views of development more narrowly focused on economic and monetary questions—perhaps one of his beefs with the World Bank. But before he gets too far with this project, Sen decides that really, he first needs to define justice. Or, at least, describe how it might be defined.
Justice and Freedom
He starts with an original parable:
Annapurna wants someone to clear up the garden, which has suffered from past neglect, and three unemployed laborers—Dinu, Bishanno and Rogini—all very much want the job. She can hire any one of them, but the work is indivisible and she cannot distribute it among the three. Annapurna would get much the same work done for much the same payment from any of them, but being a reflective person, she wonders who would be the right person to employ.
She gathers that while all of them are poor, Dinu is the poorest of the three; everyone agrees on that fact. This makes Annapurna rather inclined to hire him (“What can be more important,” she asks herself, “than helping the poorest?”). However, she also gathers that Bishanno has recently been impoverished and is psychologically most depressed about his predicament. Dinu and Rogini, are, in contrast, experienced in being poor and are used to it. Everyone agrees that Bishanno is the unhappiest of the three and would certainly gain more in happiness than the other two. This makes Annapurna rather favorable to the idea of giving the job to Bishanno (“Surely removing unhappiness has to be” she tells herself, “the first priority”).
But Annapurna is also told that Rogini is debilitated from a chronic ailment—borne stoically—and could use the money to be earned to rid herself of that terrible disease. It is not denied that Rogini is less poor than the others (though certainly poor) and also not the unhappiest since she bears her deprivation rather cheerfully, used—as she has been—to being deprived all her life (coming from a poor family, and having been trained to reconcile herself to the general belief that, as a young woman, she must neither grumble nor entertain much ambition). Annapurna wonders whether, nevertheless, it might not be right to give the job to Rogini (“It would make the biggest difference,” she surmises, “to the quality of life and freedom from illness”).
The point is that theories of justice owe a lot to their “informational bases,” or the factors that they consider when judging the justness of a situation. A utilitarian (at least a simple-minded one) will take into account only the level of subjective happiness experienced by the people involved. An economics-first development theorist will zoom in on the amount of monetary wealth or economic resources controlled by each individual. A dogmatic libertarian will consider only whether individual rights have been violated. Sen classifies the three candidates for Annapurna’s largesse as follows: “Dinu’s income-egalitarian case focuses on income-poverty; Bishanno’s classical utilitarian case concentrates on the metric of pleasure and happiness; Rogini’s quality-of-life case centers on the kinds of life the three respectively can lead.” Sen argues for an expansive informational base that includes and prioritizes the third: we should focus on the quality and type of life that is available to individuals.
This means that he rejects the other cases, or at least finds their reasoning incomplete. Utilitarianism and libertarianism come under his particular scrutiny. His version of utilitarianism consists of three components: consequentialism, welfarism, and sum-ranking. Consequentialism is, of course, the judgment of choices purely by their outcomes or results, denying the validity of general normative principles not based on outcomes. Welfarism narrows the way these outcomes can be evaluated, by restricting attention to the utilities (however defined) of the individuals in each outcome. Sum-ranking then requires that these utilities be aggregated by summing. An unjust society, according to Sen’s interpretation of utilitarianism, is one where the sum total amount of happiness is much lower than it might be.
The acceptance of utilitarianism in certain spheres seems driven by a sense of obviousness: how could the Correct Solution to Ethics not be to simply add up all the numbers and see which one is bigger? Where this conflicts with our intuitions, those intuitions must just be wrong. The general principle prevails because it is simple and reduces everything to a single dimension. The almost compulsive appeal of this perspective seems to have little hold on Sen; if anything, he wants to complicate ethical and distributional calculations by making them consider ever more data.
Sen’s objections are nothing particularly new, and indeed he had been expressing them in various forms for decades by this point. Sum-ranking is an easy target, since it inevitably leads to arguments over hypotheticals like “if you don’t sacrifice this child to Moloch, I will flip this coin one googolplex times, and if all the results are heads, then Graham’s number of beings will get the song ‘Thunder’ by Imagine Dragons stuck in their heads for a day.” More relevantly for Sen’s purposes, it means near-complete insensitivity to concerns about equality.
His other criticism is that it fails to take into account rights and freedoms except through their impact on utility. This is a situation where one person’s modus ponens might be another’s modus tollens. Sen takes the intrinsic desirability of freedom as axiomatic and therefore rejects a system that would judge “happy slaves or delirious vassals” as well-off. And if utility is defined on a psychological basis, it’s too malleable to serve as the only basis for judging circumstances. If the desperately poor can learn to be happy with their lot, where is the injustice in leaving them there? The core of the difficulty here really seems to be that it is difficult to define utility in a way that captures everything we want it to while also boiling down to a single number that can be aggregated by simple addition.
These arguments may not seem fair to many utilitarians; I found them a bit underdeveloped myself. (This is an issue throughout the book—it could have been twice as long without coming close to exhausting all the topics it tries to cover.) But the most compelling uses of utilitarian frameworks I have seen at least implicitly try to take these criticisms into account. Sum-ranking is far from universal among self-identified utilitarians, and there is a lot of work trying to make utility estimates robust to problems of misidentification. And there’s no shortage of utilitarians arguing that even though, yes, rights are in principle not inviolate, they are sufficiently good at producing good outcomes that we may as well pretend they are.
Sen’s criticisms of libertarianism are dual to his criticisms of utilitarianism. If utilitarianism is insufficiently sensitive to rights and processes, libertarianism is totally insensitive to outcomes. “Horrors of any degree of seriousness”, he argues, “can be shown to be consistent with a system in which no one’s libertarian rights are violated,” making pure libertarianism insufficient to detect obvious injustices. He combines this with a critique of the weaker libertarianism of John Rawls’ Theory of Justice, but his heart isn’t really in it. While Sen doesn’t think that rights should be inviolable, he does think they should have a representation independent of their consequences.
The alternative to all of these (and other bizarre ideas like just using per capita GDP that only make sense in the context of development) is a maximalist evaluative framework: we have to be able to take into account everything that people have reason to value. This book is written for a general audience, but the economist in Sen can’t resist slipping in a few hints of mathematical formalism: Every individual has a “capability set” from which they are able to choose their actual “functioning vector.” These functioning vectors can contain information about any collection of things an individual might be capable of—ranging from “have adequate nutrition” to “participate in the community” to “take a week-long vacation in the Caribbean.” We evaluate the state of the world by aggregating capability sets (or sometimes just functioning vectors) in some way.
“Wait a second,” cry the utilitarians. “Isn’t this just utilitarianism with extra steps? Once you choose a formula to evaluate and rank these functioning vectors, haven’t you just created utility under another name?” Maybe. But there are a few ways that the capabilities approach can differ in principle.
First, the formula used to evaluate the capability sets does not need to be constructed by assigning a single number to each individual and then aggregating those numbers. It doesn’t even have to output a single number that purports to capture the goodness or badness of a state of affairs. One could imagine an algorithm with less ambitious aims, like comparing only certain pairs of world states (a “partial ordering”) or delineating sets of clearly just or unjust situations. (Sen really doesn’t like one-dimensional evaluation: “To insist on the mechanical comfort of having just one homogeneous ‘good thing’ would be to deny our humanity as reasoning creatures.”) Even if you do end up converting capabilities into a one-dimensional output, that output is unlikely to coincide with classical notions of utility. In particular, it will probably be influenced by considerations of the process taken to get to the outcome, the distribution of outcomes, and the ability for individuals to choose.
It’s a bit hard to tell exactly how to classify Sen’s theory of justice. It looks almost like he’s just decided to use every ethical framework simultaneously, without any clear rules about when to use what. Is it utilitarian? Deontological? Is justice based on opportunities or outcomes? Well, all of the above. Maybe this is the point: if your theory of justice is capable of explicitly taking anything into account, you have to be explicit about what you’re actually using. You can’t smuggle extra assumptions in without anyone else noticing.
On the other hand, the open-endedness of the capabilities approach means that it can be used to praise or condemn almost anything, as long as you find the right capability to talk about. One could imagine equally well arguments on either side of an issue justifying themselves as supporting some substantive capability. (Imagine the argument between “excluding people with penises from women’s bathrooms supports the capability to be safe from sexual assault” and “allowing people to use the bathroom they prefer supports the capability to perform necessary bodily functions in a space one is comfortable in.”) Capabilities are susceptible to a kind of conceptual gerrymandering, where the result of your analysis depends heavily on precisely how you divide up the world.
Sen doesn’t seem to worry about this; he’s rather loose with what gets to count as a capability. Here is a partial list of capabilities or substantive freedoms he describes:
They’re all over the place, broad and ill defined. The boundaries are a bit arbitrary: why is premature mortality an imposition on freedom while, uh, mature mortality is not? And how exactly do we judge someone’s ability to take part in the life of the community? Even worse, what happens when capabilities conflict?
Most development analysis based on the capabilities approach solves these problems by being extremely basic and therefore looking quite a bit like utilitarianism. The most famous use of the framework is the UN’s Human Development Index. It evaluates countries based on functioning vectors with just three elements: life expectancy at birth, mean years of schooling, and per capita gross national income. The HDI is the geometric mean of normalized versions of these variables. This is certainly more nuanced than just ranking countries by GDP, but it’s far from the expansive multidimensional evaluation Sen promised. (It doesn’t even care about distribution, although they did recently come out with a version that uses “inequality-adjusted” per-capita income.) The HDI doesn’t appear fundamentally different from a reasonably responsible utilitarian construct like the disability-adjusted life year.
As far as producing objective metrics for development, it doesn’t seem like the capabilities approach offers a whole lot more than any other approach used with caution. Sure, we get more than just economic output to go off of, but no one seriously argues that GDP is the only thing that matters. Maybe this is the Seinfeld is Unfunny problem: at the time Sen was promoting this approach, the emphasis on single-minded monetary evaluation of development may have been stronger. Today, a multifaceted approach is table stakes, and so an argument that we need a whole new approach to address these questions seems overblown.
But Sen isn’t writing this book to develop indices and metrics. He’s trying to show how the capabilities framework can be used to think through concrete issues. A good chunk of the book consists of discussions of development-related questions. Sen’s perspective tends to be at once moderate and radical. He’s ready to acknowledge all kinds of difficulties and mitigating factors while adamantly insisting that things can be much better than they are. It reminds me of the best of the effective altruism movement: taking the responsibility to make the world better seriously and trying to figure out how exactly to do that. (Unlike a lot of EA work, though, it tends to end up sounding very normal. Maybe this is because the capabilities framework provides lots of ways to avoid weird-sounding conclusions.)
The Object Level
So what does this mean for thinking about concrete issues like poverty? It will probably not be surprising that Sen conceives of poverty much more expansively than just a lack of money. He defines poverty as deprivation of basic capabilities. While monetary resources have an important influence on an individual’s capabilities, they are far from the only determiner. And the influence of money on capabilities depends a lot on context. Age, gender, social roles, family status, location, disease, and lots of other things affect both starting capabilities and the way money can be used to accomplish things. A person with a disability may be able to convert income into functionings at a lower marginal rate than someone with a standard complement of abilities. This is distinct from whatever effect their disability has on the ability to earn income itself. Looking only at income can be misleading if you want to understand the functionings someone can obtain.
Even people with relatively high incomes can be deprived of important capabilities:
For example, a person with high income but no opportunity of political participation is not “poor” in the usual sense, but is clearly poor in terms of an important freedom. Someone who is richer than most others but suffers from an ailment that is very expensive to treat is obviously deprived in an important way, even though she would not be classed as poor in the usual statistics of income distribution. A person who is denied the opportunity of employment but given a handout from the state as an “unemployment benefit” may look a lot less deprived in the space of incomes than in terms of the valuable—and valued—opportunity of having a fulfilling occupation.
This doesn’t mean unemployment benefits are bad, but Sen does suggest that they might obscure some important forms of deprivation.
This heterogeneity in the ability to convert money into capabilities is one reason I am hesitant about a universal basic income as a replacement for most welfare programs. Because there are significant differences in the ability of different people to convert money into things that matter, a UBI could leave some people—often the most vulnerable—in precarious positions. I think Sen would be much more interested in “universal basic capabilities.” Maybe this is actually the orthodox view on welfare and social services: things like food stamps and disability benefits are meant to support specific functionings rather than income.
Another illustration of the independence of money from meaningful capabilities is the fact that health outcomes often change independently from income. In Britain during the first half of the 20th century, life expectancy increased fastest at the times when per-capita income was growing the slowest (or even shrinking). Or consider that, at least when the book was written, African Americans in the US had significantly lower life expectancy than residents of various Asian states like China or Kerala, India, despite having per-capita income many times higher. One might quibble by asking whether people of African descent would actually do better in Kerala than in Alabama, hinting at, say, genetic or cultural influences. But we’re concerned with capabilities here, not necessarily their origins. Deprivation is deprivation, regardless of its cause.
As an economist, Sen naturally recognizes that there are tradeoffs involved in addressing questions like this. The aggregate total of resources and the distribution of resources depend on each other in complicated ways. And actually solving a problem like depressed life expectancy requires understanding its causes. But knowing that there is a problem does not.
Discussions of poverty come closely coupled with talk of inequality. Sen asks “inequality of what?” and I’m sure you already know the answer. Thinking about inequality in terms of capabilities means that there are many interacting axes on which inequality can happen. We shouldn’t think only about income or wealth inequality, but also inequalities in the ability to live a healthy life, or to obtain an education, or to participate in the political process.
Sen sees inequality itself as a potential component of poverty, because it impinges on what Adam Smith called “the ability to appear in public without shame,” and what Sen describes as “tak[ing] part in the life of the community.” The idea is that to be able to participate in society often requires access to resources that are determined by the modal level of wealth. Those significantly below this level, even though their wealth might be envied by those in poorer societies, are still excluded from an important functioning. For Adam Smith in 1770s England, the necessary items were a linen shirt and a pair of leather shoes; for Amartya Sen in 1990s America, they were a telephone, TV, and a car. Today there is a huge range of social and economic opportunities that someone without a smartphone or internet access can’t access. There’s still the question of whether it’s better to secure this capability by making sure everyone can get a smartphone or by changing society so that it’s easier to function without one. But the poor aren’t simply greedy or profligate for wanting smartphones.
Is all this intersectionality? I don’t know, but Sen’s work aligns in important ways with movements for social justice. Deprivation or inequality of any important human functioning can be seen as an injustice in the capabilities framework. And framing all kinds of issues as a reified type of justice is just the way we do things now. Despite introducing this expansive evaluative framework to define justice, Sen is hesitant to use it without very good reason:
The recognition of evident injustice in preventable deprivation, such as widespread hunger, unnecessary morbidity, premature mortality, grinding poverty, neglect of female children, subjugation of women, and phenomena of that kind does not have to await the derivation of some complete ordering over choices that involve finer differences and puny infelicities. Indeed, the overuse of the concept of justice reduces the force of the idea when applied to the terrible deprivations and inequities that characterize the world in which we live. Justice is like a cannon, and it need not be fired (as an old Bengali proverb puts it) to kill a mosquito.
See what I mean that this book isn’t really about development? To be fair, I’ve elided a bunch of stuff about famines, food production, fertility, and economic growth that’s more what you would expect from a development book. In my developed-world chauvinism I’ve focused on the bits most relevant to wealthy nations.
But one of Sen’s favorite things seems equally relevant for everyone: democracy. He finds its appeals so obvious that he spends most of the time responding to various gritty realist criticisms of the system rather than building a positive case. Some of these critiques, like the idea that humans are just too selfish to cooperate outside of markets, read like the sort of thing a left-wing caricature of an economist would say, but others have more pull today. A big one is the idea that poor countries aren’t ready for things like democracy, liberalism, and human rights, that their economies need to grow under a stricter regime until they reach a certain level of prosperity. Maybe after that they can transition to more pluralistic rule. Advocates for this view point to the experiences of (mostly Asian) countries like South Korea, Singapore, and China, which have experienced dramatic economic development under regimes that seem oppressive. Sen bills this the “Lee thesis,” after Singapore’s Lee Kwan Yew. And let’s be real: the choice between living in Singapore on $68k a year or Jamaica on $7k a year seems pretty compelling in favor of Singapore.
Sen is having none of this. First, he denies the presumed trade off, noting that the evidence that authoritarianism is better for economic growth than democracy is pretty equivocal. For every fast-growing autocratic Singapore, there’s an equally fast-growing democratic Botswana. More recent research has found a mild positive effect of democracy on economic growth. I wouldn’t take these results as definitive, but they do show that the effect is far from clear-cut. At worst, democracy has no average effect on economic growth. In Sen’s view, the reason for this positive effect is that democracy aligns the interests of rulers and the general population. Yes, by chance you might get a well-meaning dictator who happens to know all the right policies to institute better than any democratically elected leader would, but the incentive structure doesn’t point that way. We don’t want to pin our hopes on always having perfect people in charge; that’s not how you produce consistently good results in the long run.
And of course, even if democracy were merely neutral (or even slightly negative) for economic growth, it would still expand the set of freedoms that people enjoy—if nothing else, their political freedoms. The willingness to trade liberties for money only goes so far. Replace Singapore and Jamaica with, say, Saudi Arabia and Costa Rica and the case for preferring the less-free state feels much weaker, even though there is still a significant economic gap. In Sen’s framework, democracy and human rights are constitutive components of development, not independent competing priorities. This is of course the starry-eyed end-of-history view—it does seem a bit convenient that the constitutive components of development never conflict with each other. But the idea of a strict tradeoff between different desirable features is perhaps equally naive. Are we really arrogant enough to believe we’re anywhere near the Pareto frontier?
But wait, you say, who’s to say that the Western ideals of liberty, democracy, and human rights are even valuable to other cultures? Isn’t it presumptuous and colonialist to think that we get to determine what people actually value in other societies?
Sen will also have none of this, and spends most of a chapter gleefully dismantling the idea that “Asian” values (because the example is invariably Asia) are universally shared across Asia or have bearing on the sort of government that is suitable for Asians. His remark that “these justifications of authoritarianism have typically come not from independent historians but from the authorities themselves” is a good start. The main thrust, though, is that in essentially every cultural tradition it is possible to find threads supporting freedom and equality as well as threads emphasizing stability and authority. It’s important not to interpret any of these threads as defining a uniform, unchangeable cultural pattern. After all, historical “Western” thinkers were often highly selective about who exactly deserved personal freedom, and someone like St. Augustine could be just as “authoritarian” as Confucius.
Figures from Asian history expressing ideas of freedom and tolerance include the Indian emperor Ashoka, who after converting to Buddhism had inscriptions erected across the country with messages like “a man must not do reverence to his own sect or disparage that of another man without reason.” Also the Moghul emperor Akbar, who decreed that “no man should be interfered with on account of religion, and anyone [is] to be allowed to go over to a religion he pleased.” (Sen wryly notes that at the time of Akbar’s reign, western Europe was undergoing the Inquisition.) Even Confucius was not as uniformly deferential to authority as some of his modern interpreters would like us to think. He says that one should “tell [a prince] the truth even if it offends him,” and suggests opposition to a corrupt government: “When the [good] way prevails in the state, speak boldly and act boldly. When the state has lost the way, act boldly and speak softly.”
Toleration and equality did not (and do not) come for free in the West either. If liberal democracy is a constitutive component of development, maybe we think of it as a Western value because the West developed first. That is, democracy, like sugary drinks and penicillin, is part of universal culture. It’s adaptive, or at least sticky, and it just so happens that the West got it first.
Despite his contempt for authoritarians who try to use culture to justify their authority, Sen recognizes the problems that arise for native cultures in the face of the developmental advantages of globalization. This is again to be addressed by democracy, not elites deciding what’s best for the culture. The society needs to determine how much it is willing to sacrifice to maintain traditional ways of life. And in order to do that, the people need to talk about it. Culture is about all the people who participate in it, not just the elites. And in order to do that, they need to be well informed and able to communicate. And in order to that, they need to have a number of other capabilities. Everything is capabilities, and capabilities are all connected.
Social Choice and Democracy
These questions of culture and society highlight a difficulty that I and Sen have both glossed over up to this point. It arises with the shift from using the capabilities approach to evaluate how we rank a situation as individuals to using it to make decisions that affect other people. Who actually gets to do the evaluation then? It’s fine for me to collect sets of functioning vectors and decide how I feel about the state of the world based on that information. But when measures like the HDI get used to make decisions about the allocation of real resources, the actual choices made in the analysis start to really matter. Who gets to decide what to value?
Of course, Sen’s answer is democracy. This is not a trivial claim for him to make. A lot of his academic work is in the field of social choice theory, which tries to study how societies can and should make decisions consistent with the preferences of their members. But the field’s most famous results are ones showing that this aim is impossible. The Condorcet paradox, Arrow’s impossibility theorem, and the Gibbard–Satterthwaite theorem all seem to indicate the infeasibility of the whole project. Social choice looks like a field dedicated to proving its own futility. Sen himself introduced a variation on Arrow’s theorem that he termed the liberal paradox. So why is he so enthusiastic about democracy, especially when he seems to want it to reach definitive conclusions about abstract things like capability sets?
The answer, Sen would have us believe, is again that the informational base assumed by these impossibility theorems is too small. Arrow tells us that producing a consistent society-wide preference ordering based only on the collection of all individual preference orderings is impossible, so we must need to use more information. Why shouldn’t the social decision function be based on more than just preferences? Maybe it should incorporate information about inequality, or individual rights, or maybe even…capability sets? Wait a second.
I was terribly confused when Sen first suggested using more information as the solution to the problem of social choice. He wants to use social choice to solve the problem of how to weight all the information involved in evaluating capability sets. But if we need to do basically the same thing to construct a social choice function, we haven’t solved the problem at all, just shuffled it around. The parts of the book where he talks about these two topics are somewhat separated, so maybe he just thought he could sneak that claim past most readers?
I don’t think Sen was trying to pull a fast one on us, but it is a bit difficult to piece together how this is actually supposed to work. There’s a hint of the solution when he talks about the problem of three people sharing a cake. If the partition is determined by majority vote, there’s no stable solution, since two people can always collude to increase their shares at the expense of the third. This doesn’t mean it’s impossible to equitably share the cake, though; the process just needs some extra shared context, and a combined desire to escape the compulsive logic of the zero-sum game.
What I think Sen ends up describing is a kind of democratic bootstrapping. Theorems about social choice tend to assume a static collection of individuals with fixed preferences and a single social decision to make. That’s not how people or democracy work. Preferences are often fuzzy, and they respond to argument and discussion. In these discussions, people share values, and (implicitly) discuss different ways of weighing them to make decisions. The Overton window shifts. Certain types of preferences (like about your neighbor’s religion) may not disappear, but cease to be seen as legitimate grounds for society-wide decisions. A treaty of toleration starts to take shape. This doesn’t mean agreement on everything, but there is a broad sense that certain things are acceptable and certain things are not, and that there is an established way to change things when it’s necessary. Egregious, widespread injustices are usually addressed, at least as long as most people recognize them. Before you know it, your society has made its way to the liberal consensus. It’s not perfect, but it’s also not severely unjust. And that’s all Sen hopes for from a society-wide theory of justice anyway.
So liberal democracy itself is maybe the best instantiation of the capabilities approach, more so than any computed metric like the HDI. It fails terribly at producing a consistent or precise ranking of alternative possible worlds. But it does tend to get very important things right. (Sen’s favorite factlet is that famines basically never happen in functioning democracies, no matter how poor.) And specifically liberal democracies, with checks and balances and constitutionally enumerated rights, in principle take into account more than just aggregated preferences in making decisions.
There probably still needs to be a kernel of agreement with the liberal idea to start from for this process to work. But as Sen pointed out while skewering authoritarianism, such kernels are found in most cultural traditions. The resulting democracy will probably look different depending on the culture, but as long as it enables political participation and supports other freedoms and capabilities, that’s fine.
Although Sen never quite puts it this way, the general sense is that a broad liberalism is really the only option. Claims about the legitimacy of other political structures based on culture are incoherent because people have and always will disagree about what exactly their culture means. Someone like Lee Kwan Yew or Viktor Orbán might have an internally consistent alternate framework, but these epistemically closed systems can’t justify themselves externally, or even to everyone that is supposedly inside them. Liberalism is the ur-epistemically-open framework, and is basically the only way to incorporate multiple frameworks without totally collapsing under contradictions.
This is a highly stylized and relatively optimistic description of how democracy develops and works. It’s worth asking whether this actually happens in practice. Do societies actually feel their way toward liberal democracy when given the chance? Or is it just a cultural value that Westerners tend to give far too much weight? And perhaps more importantly, does having a liberal democratic government actually improve people’s material condition? Are there important things that democracy tends to get wrong? The case for democracy looked a lot more optimistic in 1998 than it does now. Russia was groping its way toward an open society, or at least hadn’t invaded anyone recently. The West was full of hope that China’s economic liberalization and resultant growth would result in political changes. India seemed to be on a more liberal path than it is now. And nominally democratic strongmen didn’t usually attract admiration from Western politicians.
Besides the obvious setbacks liberal democracy has suffered, there seems to also be an increasing weariness toward the general liberal idea. Perhaps this is just an artifact of the increasing visibility of neoreactionary and Marxist ideas, or the growing sense across the political spectrum that our institutions are broken. Sen’s 90s sensibility doesn’t seem to feel any real need to address this neoliberal malaise, or argue that democracy is still relevant today, beyond his discussion of issues like the Lee thesis. And he doesn’t address at all what happens when the people elect a dictator, or the sorts of political failures that can make this seem like the only viable option.
I wish I could be this optimistic about democracy. I would chalk Sen’s optimism up to it being the 90s, but I don’t think his views have changed in the last two decades, and I think they were pretty similar before that, too. But maybe my pessimism can also be attributed to a particular temporal and cultural setting. It seems obvious from today’s perspective that liberal democracy is neither inevitable nor sufficient for development. And to be fair, I don’t think Sen believes that either. He spends most of the book arguing that development consists of a network of mutually reinforcing processes, and it’s only when you have them all running that development fully occurs. But he does seem to put quite a lot of responsibility on democracy specifically, in a way he doesn’t for, say, equitable access to food. And I don’t know if democracy is capable of carrying that burden without some external cultural support. The network of dependencies is denser than Sen acknowledges. Maybe this doesn’t matter so much for the broad goals of development. The fact that democracy is a system that can recognize and rectify extreme injustice might be enough, even if it can’t quite accomplish every task Sen assigns it.
In the end, Sen’s project is to build a society that values freedom, broadly construed, and tries to make it happen. He calls this “individual freedom as a social commitment,” and ends the book with his optimistic case that we can make the world a better place through reason and liberty. The challenges of poverty, authoritarianism, and social choice are real, but it’s possible to overcome them and move forward. Despite today’s cynicism toward democracy, markets, and institutions, Sen’s enthusiasm is inspiring and more than a little infectious. This is another way he fits in with r/neoliberal. Even when I’m not fully on board with them, that community’s earnestness, vision, and willingness to think things through makes it a pleasant place to be. The same is true for Development as Freedom. Amartya Sen makes an almost painfully earnest case for the power of humanity to make things better by working together, and I can’t help but want him to be right.