1. Introduction
Life without choice would be grim. So it’s no surprise freedom of choice is an important value in liberal societies. And philosophers provide good arguments too, citing respect, autonomy, responsibility and wellbeing as reasons to value choice. However, even for the best things in life, more is not always better. For example, a quick internet search tells me that Amazon sells around 75 million products in the United States. Isn’t there a point at which we have enough or even too much choice?
So, how much choice should we aim for? One answer popular in political philosophy is:
The Sufficiency View: expanding freedom of choice is an important goal for institutional design, public policy, and for societies more generally, but only up to some level L, where L is not too far from the level enjoyed by well-off citizens in rich liberal countries today.
Several prominent theorists endorse the Sufficiency View or at least endorse views that seem to assume something like it (Dworkin Reference Dworkin1982; Raz Reference Raz1986; Frankfurt Reference Frankfurt1987: 42–43; Blake Reference Blake2001: 269; Haybron Reference Haybron2008). To make this a distinctive view, I have added ‘L is not too far from the level enjoyed by well-off citizens in rich liberal countries today’. The interesting debate is not whether there might be a hypothetical and absurdly high level at which additional choice lacked value but whether we should give importance to increasing choice at current or near-term future margins.Footnote 1
To support the Sufficiency View, theorists typically claim that additional choice above L neither improves wellbeing nor autonomy. More choice might even make us worse off. Such worries resonate not only with culturally conservative and economically left-wing critics of liberalism. Some hold it is backed up by empirical research. In his influential book The Paradox of Choice: Why More is Less, psychologist Barry Schwartz argues that when choice reaches high levels, ‘choice no longer liberates, but debilitates. It might even be said to tyrannize’ (Schwartz Reference Schwartz2009: 2). He suggests excessive choice might also explain why happiness has been flatlining for decades while depression has gone up.
Such claims are serious and, if true, would have major implications. However, in this article, I argue that we should reject the Sufficiency View along with the dramatic claims that we have too much choice. Instead, I defend:
Liberal Optimism: increasing people’s freedom of choice, even far beyond levels experienced by well-off citizens in rich liberal countries today, should be an important goal in institutional design, public policy, and for societies more generally.
I argue that none of the theoretical and empirical arguments succeed in supporting the Sufficiency View. Instead, traditional arguments from the liberal tradition along with empirical data support Liberal Optimism.
Whether we should adopt Liberal Optimism or the Sufficiency View has important implications. How strongly should we commit to the liberal project and expand freedom of choice? And how far should we trade off choice for other goods? It also has implications for applied ethics, political philosophy and public policy. For example, do autonomy and freedom give us reason to open borders to more immigration? Should rich countries slow down economic growth or even ‘de-grow’ in response to climate change? How important are markets and the seemingly endless options they generate? And how far should we balance freedom with the value of community?
I proceed as follows. In section 2, I outline what freedom of choice is, introduce the Sufficiency View and Liberal Optimism, and present reasons to value choice that together make for a presumptive case for Liberal Optimism. In section 3, I discuss arguments for the Sufficiency View that hold that too much choice reduces wellbeing. I argue that they fail. Neither appeals to paternalism, nor experimental research in psychology, nor do any macro-trends provide the needed support. If anything, broader trends and evidence support that freedom of choice improves wellbeing. In section 4, I argue that future expansions of choice, far beyond L, are still possible and valuable. In section 5, I present modified versions of my arguments to briefly respond to the view that high levels of choice are unhelpful or even counterproductive for autonomy. In section 6, I sketch some implications for discussions around immigration, economic growth, markets and the value of community. I conclude in section 7.
2. The Value of Choice
2.1. Freedom of choice
‘Freedom of choice’ means having the freedom to choose between different options. Some take ‘options’ to be social opportunities, and primarily focus on the absence of external social constraints (Pattanaik Reference Pattanaik2018). Others take options to be abilities (or capabilities), which require external social opportunities, the absence of external natural constraints, and the internal abilities to make use of opportunities (Parijs Reference Parijs1997; Sen Reference Sen1999; Kramer Reference Kramer2003; Schmidt Reference Schmidt2016a). My arguments could largely be run with either understanding, but I assume options are abilities, which is also intuitive: for my freedom of choice, it matters what I am actually able to do, and physical capabilities are often necessary for most other options (Venkatapuram Reference Venkatapuram2013).Footnote 2
On many theories, freedom of choice is an important but insufficient part of freedom. For example, liberal theories often emphasize not only the availability of options but also how many options are constrained by other people (Steiner Reference Steiner1983: 74). Republican theories emphasize that being a free person requires not just choice but also the absence of domination (Pettit Reference Pettit2014). We can ignore these niceties here, because I only focus on the value of choice rather than ‘all-things-considered freedom’.
Finally, freedom of choice comes in degrees. A rich literature in economics and philosophy offers different measures. While the details are beyond my scope, note three things.Footnote 3
First, when comparing option-sets, we should individuate options consistently across sets. For example, imagine César can choose from two bottles of wine today, one red and one white (option set Q ={red wine, white wine}). Tatiana has a mansion with two extensive wine cellars, each of which containing many whites and reds, and she can choose from one of the cellars today (option set S = {wine cellar 1, wine cellar 2}). While Q and S both contain two options, they clearly do not offer the same freedom of choice. A plausible measure individualizes consistently across option sets and judges that S offers more choice than Q.
Second, we have more choice when we can combine options. For example, Clara can choose between having a family and having a career. Sylvia can choose between having a family, having a career, and having a career and a family. Sylvia has more choice than Clara.
Third, is freedom of choice only a matter of how many options one has or also of how good those options are? Discussions on freedom measurement have grown so complex, they exceed my scope. But, for the purposes of this article, I assume the Hybrid View: all additional options, whether valuable or not, increase freedom of choice; but the more valuable those options are, the more they increase choice – which can be captured using evaluative weighting factors. To capture value, some focus on a person’s actual or idealized preferences or a range of ‘reasonable’ preferences. Other proposals eschew preferences for other values associated with freedom, such as how option-sets contribute to wellbeing, experimentation or agency.Footnote 4 I elsewhere defend a view broadly of the latter type (Schmidt Reference Schmidt2014). Here, I remain mostly neutral, though I dismiss simple measures, like revealed preference views, that don’t cohere with our diverse reasons to value choice (reasons I outline below).
Finally, it seems intuitive that more diverse option sets give us more choice. Some therefore propose measures where diversity is its own dimension. My argument is compatible with such measures. But for simplicity, I assume that the Hybrid View’s qualitative and quantitative dimensions already capture our intuitions around diversity: more diverse option sets typically contain more valuable options (Kramer Reference Kramer2003: 463–471) and/or more ‘fecund’ options that enable many future options (Carter Reference Carter1999: 198–204).
Now, according to the Hybrid View, adding consumer options increases choice even at high levels, like in the Amazon case. At the same time, it captures that some options increase choice far more than others, either because they are more valuable or more fecund. For example, at the margin, the 50th brand of bottled water seems to add neither value nor new follow-up options. Becoming affluent enough to afford a car, in contrast, is valuable and opens up new options going forward. (I revisit these issues in section 6.)
2.2. Why value choice?
Choice has intrinsic value, when it is valued for its own sake, and instrumental value, when valued as a means to attain something else of value. But note that even if it only has instrumental value, choice can still have non-specific value: ‘A phenomenon, x, has non-specific instrumental value iff x, without regard to the nature of its specific instances, is a means to some other valuable phenomenon y’ (Carter Reference Carter1999: 44). Consider money as an analogy. We might not value money intrinsically. Still, as Ian Carter writes, ‘we do not value money only as a means to buying the latest Mozart recordings or as a means to eating a bar of chocolate, but also as a means to satisfying whatever our future desires turn out to be’ (Carter Reference Carter1999: 34).
Here is my take on what I consider the best arguments for why freedom of choice has non-specific value.Footnote 5
First, individuals change over time and often change their preferences, values and priorities, which highlights the importance of preserving choice for the future (Carter Reference Carter1999: 50–54). Interestingly, people typically underestimate how much their current self will change (Quoidbach et al. Reference Quoidbach, Gilbert and Wilson2013). We have good reason – and better reason than we think – to keep future choices rather than irreversibly commit to things our future selves might not want, value or enjoy.
Second, before committing, choice lets us gather information about what is available. For example, you shouldn’t buy the first house you view, take the first job that sounds fun nor marry the first person interested in you. As computer scientists would say, before you can ‘exploit’ good options, you should first do some ‘exploring’ (Christian and Griffiths Reference Christian and Griffiths2016).
Third, choice also enables experiments in living whose benefits go beyond learning what is available. For example, they might help you develop your agency, tastes, values and likes, and teach you lessons about the world (Mill Reference Mill1859). Moreover, experimentation can be fun in itself: diverse and new experiences help us enjoy life more and make time pass more slowly, and learning new hobbies and skills increases reported happiness (Howell et al. Reference Howell, Chenot, Hill and Howell2011).
Fourth, individuality and self-expression require choices (Scanlon Reference Scanlon and McMurrin1988): when a Maoist regime achieves grey conformity – where we all wear the same grey clothes and read the same red book – it leaves little space to express ourselves through our choices.
Fifth, autonomy and agency require choices. The relationship can be ‘conceptual’. For example, Tom Hurka argues that agency requires that I can reject options (Hurka Reference Hurka1987). And without options, I cannot reject anything and thus cannot exercise agency. This will make choice intrinsically valuable if you think autonomy is intrinsically valuable. But the relationship can also be ‘empirical’. For example, one tradition in psychology suggests a causal relationship between learned helplessness and depression (Peterson and Seligman Reference Peterson and Seligman1984). Conversely, feelings of personal control might reduce or prevent depression. And we are unlikely to experience personal control and its psychological benefits, if we have no choices.
So, choice is non-specifically valuable for individuals. But choice becomes even more valuable, if we view it as a social value. The disagreement between the Sufficiency View and Liberal Optimism is not (only) about the value of choice for one individual but about its value for whole societies. And while reasons to value the former provide reasons to value the latter, other reasons pertain specifically to choice as a social value.
First, people have different personalities, likes and conceptions of the good. Particularly in diverse societies, giving people choices can facilitate that people get what they want, value and like.
Second, freedom of choice can have positive externalities. For example, rather than conducting my own experiments of living, I can also learn from other people’s experiments and conceptions of the good. A society committed to freedom of choice will generate more diversity and innovation in living and thereby better conditions for individuals to learn. Moreover, as Friedrich Hayek argued, freedom of choice in market economics has economic advantages, as it helps with economic and social discovery and progress (Hayek Reference Hayek1960).
Finally, choice as a social value is also required for respectful social relations: to view and treat each other as free and autonomous agents, we must, under normal conditions, grant each other some freedom of choice. A society that restricts individual choice too much is unlikely to generate the conditions under which we respect and view each other as free and equal.
2.3. But how much choice?
The arguments presented make a compelling case for choice having non-specific value, particularly as a social value. First, prima facie, most (or maybe even all) arguments appear to have no natural nearby point of sufficiency.Footnote 6 Second, freedom is valuable for several reasons. So, even if one reason satiates at some sufficiency level, that need not hold for the others. So, the first leg of my argument is that, given liberal arguments, and absent any arguments to the contrary, more choice is better than less at the societal level. This gives us a strong presumptive case for:
Liberal Optimism: increasing people’s freedom of choice, even far beyond levels experienced by well-off citizens in rich liberal countries today, should be an important goal in institutional design, public policy, and for societies more generally.Footnote 7
Note three clarifications about Liberal Optimism.
First, it does not imply that the state itself must always actively increase people’s options. Often the state needs to set favourable rules, expectations, laws or regulations or simply not interfere so that non-state actors and individuals can increase their own or other people’s options.
Second, more choice need not always be better all things considered and must sometimes be sacrificed because of externalities and competing values. At the same time, Liberal Optimism considers choice a weighty goal that shouldn’t be traded off for other goods too easily, as should become clear through my examples and discussions below.
Third, Liberal Optimism is not about one person’s choice but about producing good distributions of choice across persons. Various theorists implicitly endorse something like Liberal Optimism combined with a criterion for the interpersonal distribution of freedom. Philippe van Parijs argues we should maximize the freedom of the worst off (and then that of the second-worst off and so on) (Parijs Reference Parijs1997), and several other theorists suggest we should aim for maximal freedom or the greatest equal freedom (for examples, see Carter Reference Carter1999: 76–83). I here do not provide a criterion myself (see Carter Reference Carter1999: Ch. 3; Schmidt Reference Schmidt2014 for more on these issues). But I assume Liberal Optimism does not imply maximizing the unweighted sum of individual choice. A plausible criterion would likely include some egalitarian or prioritarian weighting.
I have argued that, absent strong countervailing reasons, liberal arguments support Liberal Optimism. However, others think strong countervailing reasons exist and endorse:
The Sufficiency View: expanding freedom of choice is an important goal for institutional design, public policy, and for societies more generally, but only up to some level L, where L is not too far from the level enjoyed by well-off citizens in rich liberal countries today.
Different versions of the Sufficiency View can take different stances on what happens above L:
Strong Sufficiency View: The Sufficiency View plus: above L, there is reason to reduce choice so that it goes down to L again (or thereabouts).
Medium Sufficiency View: … above L, there is no reason to increase choice any further.
Weak Sufficiency View: … above L, there is no reason, or at best only very weak reason, to increase choice any further whenever one must trade off choice with another good.
Note two clarifications.
First, the disagreement between Liberal Optimism and the Sufficiency View is not about the best interpersonal distribution of choice. For example, both positions would likely agree that we should prioritize those who currently have very little choice over those who have a lot (like the global rich). Instead, the Sufficiency View would hold – whereas Liberal Optimism would deny – that if everyone were at L, we would have no reason (or at best very weak reason) to increase choice further.
Second, the disagreement is not about the specific value of particular options above L. For example, imagine we are at L but discover a medical product that helps people with vertigo. The Sufficiency View would endorse adding this option because it has specific value. The disagreement between the Sufficiency View and Liberal Optimism concerns whether expansions above L add non-specific value.
Finally, the disagreement is not about whether, theoretically, there can ever be a satiation point. Liberal Optimism can allow that. For example, if people at any given moment had near-infinite choices to do and be nearly anything imaginable, Liberal Optimism might not care about adding another 100 options. Our reasons to expect positive marginal value from more choice might have run their course. Rather, the interesting disagreement is about whether we should place much weight on increasing choice at current or near-term future margins. Those, after all, are the practical concerns in the ‘too much choice’ discussion.Footnote 8 Accordingly, I assume L is either around the level had by better-off people in liberal countries today (say the top 4% income percentile worldwide) or at some level achievable soon.
Considerations in section 2.2 make a good presumptive case for Liberal Optimism that holds absent any strong arguments to the contrary. I now discuss the strongest and most common argument raised against Liberal Optimism and in favour of the Sufficiency View: above L, more choice does not contribute to wellbeing; if anything, it makes us worse off.
3. Choice and Wellbeing
I now discuss:
The Negative Claim: above L, adding more choice can reduce people’s wellbeing to an extent that additional choice either has negative marginal value above L or that any positive marginal value is largely outweighed by its disvalue (‘nullified’).
The Negative Claim supports the Sufficiency View if we believe – as I think we should – that wellbeing provides weighty reasons for action in institutional design, public policy, and for societies more generally.
Something like the Negative Claim might be gleaned from Gerald Dworkin’s famous article Is More Choice Better than Less? (Dworkin Reference Dworkin1982). But psychologists have also added much research that seemingly supports the Negative Claim. Before assessing such arguments, I first need to show that two theoretical responses to disarm the Negative Claim will not do.
3.1. What about intrinsic and non-specific value?
The first response goes like this: because arguments for the Negative Claim challenge only the instrumental but not the intrinsic value of choice above L, liberal optimists can ignore those arguments if they believe choice has intrinsic value. However, this response is unsound: if arguments for the Negative Claim hold up, instrumental disvalue above L might nullify any gain in intrinsic value. Here are two reasons.
First, even if choice has intrinsic value, its marginal value might still decline with growing choice. One popular argument for its intrinsic value is this: first, autonomy is intrinsically valuable; second, some choice is a necessary constitutive component of autonomy; therefore, some choice is intrinsically valuable. However, the marginal intrinsic value of choice might still decline, because its contribution to autonomy might decline. This might happen, for example, when choice goes up without the other ingredients necessary for autonomy going up.Footnote 9 Now, if the Negative Claim holds up, the marginal gain in value above L might then be outweighed or nullified more easily by a loss in instrumental value.
Second, even if the intrinsic marginal value is constant, it might still be nullified if the marginal instrumental disvalue above L is big enough. So, either way, even if choice has intrinsic value, liberal optimists better address the Negative Claim.
The second ‘philosophical’ response to the Negative Claim would point out that choice has non-specific value which remains intact even if additional choices can sometimes reduce wellbeing. However, this response does not disarm the Negative Claim either.
First, as we will see below, several arguments for the Negative Claim concern the non-specific disvalue of more choice. So, against the Negative Claim, liberal optimists need to show that non-specific disvalue does not outweigh non-specific value.
Second, we are here not only concerned with whether choice has pro tanto value no matter how weak that pro tanto value might be. After all, the Weak Sufficiency View can allow that choice has very weak positive marginal value above L. Instead, liberal optimists must show that the marginal value of choice above L is weighty enough to justify a firm institutional commitment and trading off other goods. Therefore, Liberal Optimism needs to address the Negative Claim head-on.
3.2. Paternalism
The first argument for the Negative Claim holds that, sometimes, paternalistically reducing people’s options can increase their wellbeing (Dworkin Reference Dworkin1982: 76–78). For example, when designed well, tobacco control reduces choice but increases health and wellbeing.
However, paternalism alone does not justify the Negative Claim. The first response is that the tension between paternalism and freedom of choice is often exaggerated. Here are two reasons.
First, not all forms of paternalism restrict people’s choices. For example, so-called nudge interventions only change the way options are presented, for example by changing the default (Thaler and Sunstein Reference Thaler and Sunstein2008). Moreover, financial incentives or subsidies – for healthy food for example – do not restrict freedom of choice either. And other paternalist interventions only replace options. For example, until recently, many products contained ‘trans fat’ until they had to be replaced with healthier fat types.
Second, some paternalist interventions – even some quite intrusive forms – actually increase freedom of choice rather than reduce it (Schmidt Reference Schmidt2017). Freedom of choice is not just a ‘momentary’ good but one that extends across time (Carter Reference Carter2013; Schmidt Reference Schmidt2017). Beyond momentary choice, Liberal Optimism should promote people’s choice throughout their lifetimes. For example, at time t, the option to smoke cigarettes expands my freedom of choice at t. However, once addicted, my expected future options would shrink drastically, as tobacco kills up to one in two smokers and reduces life expectancy by around 10–15 years. Under some conditions, curtailing the freedom to smoke cigarettes can increase people’s expected lifetime freedom (Schmidt Reference Schmidt2022b). Similar arguments can be made for options that carry significant financial or health risk, such as high-stakes gambling, driving without a seat belt, using heroin, and others.
The second response holds that even if some options reduce wellbeing but not (intertemporal) choice, this would not undermine the non-specific value of choice. For the disvalue here is mostly specific rather than non-specific: it concerns not the disvalue of choice as such but the disvalue of what happens after choosing a particular option. Accordingly, the proper response – where possible – is not to limit choice overall but to target those specific options. For example, assume that while money has non-specific value, it also allows you to buy cigarettes that make you worse off. It would be misguided if I said ‘well, let’s reduce everyone’s income, as being able to buy cigarettes has such disvalue!’. It makes more sense to preserve the non-specific value of income but regulate particular consumer options instead.
3.3. Choice overload
Philosophers who endorse something like the Sufficiency View often cite Dworkin’s influential article as support. However, Dworkin’s intention is limited to providing examples ‘in which more choice is not necessarily to be desired’ (Dworkin Reference Dworkin1982: 65). As such, his examples alone do not entail the Negative Claim.
First, a firm commitment to a value that guides social practices or institutions – such as autonomy, democracy or the rule of law – can be justified even if there are individual instances where it would (theoretically) be better to have less of it. Moreover, such instances are compatible with more of this social value being better overall. Imagine, for example, you work for a police force in a country that is progressing towards the rule of law but still has some way to go. Imagine you encounter instances in your job where ‘more rule of law’ is not necessarily to be desired and you wish you could pause it momentarily. Such instances are compatible with more rule of law being overall better for the country. Similarly, individual instances ‘in which more choice is not necessarily to be desired’ do not imply that increasing choice at the societal level would reduce wellbeing.
Second, and relatedly, the disagreement between the Sufficiency View and Liberal Optimism is about choice as a value across individual lives (or longer stretches of time) and for whole societies. Dworkin’s examples, however, are limited to one person being temporarily worse off. As such, they might not justify strong conclusions about the non-specific value of choice across longer times and for whole societies.
So, for the Negative Claim, we require systematic evidence against freedom of choice as a social value, which makes the psychology literature on ‘choice overload’ a good candidate. First, unlike the paternalism argument, it concerns the non-specific (dis)value of more choice and, second, provides empirical research rather than individual cases. (Interestingly though, many of the studied effects revolve exactly around the cases anticipated by Dworkin.) Here are some of the studied effects.
First, with more options, decisions can take more effort and time (Dworkin Reference Dworkin1982: 66–67). And with more diverse options, weeding out undesired options requires greater energy (see Scheibehenne et al. Reference Scheibehenne, Greifeneder and Todd2010: 411). Such effort can feel pointless when the chosen option turns out not much better than the one you would have chosen from a smaller or more homogeneous set. Schwartz, for example, describes buying jeans in a store with a vast selection of shapes and varieties. He writes ‘The jeans I chose turned out just fine, but it occurred to me that day that buying a pair of pants should not be a daylong project’ (Schwartz Reference Schwartz2009: 2).
Second, with larger, better and more diverse option-sets, we are more likely to feel bad about missed opportunities. We are more likely to remember rejected good options as missed opportunities and to regret even good decisions. As a result, we experience more regret and counterfactual thinking, both of which tend to reduce happiness (Schwartz Reference Schwartz2009).
Third, with more options, we must also take more responsibility, which can be a psychological burden (Dworkin Reference Dworkin1982: 67–68). The more choices people have, the more they might view bad outcomes as their own fault. And individuals who use causal attribution models where they primarily blame themselves for bad things happening in their lives are more likely to experience depression. So, as Schwartz argues, expanding the space for such attribution and self-blame might reduce happiness and increase depression (Schwartz Reference Schwartz2000, Reference Schwartz2009).
Fourth, with increased choice, our hedonic expectations increase. Choosing from a diverse option-set, we might have high hopes for the option we end up with. Yet, as so often happens, the chosen option is just alright. So, with greater choice comes greater disappointment.
Finally, the above effects might sometimes compound and cause choice paralysis such that individuals end up making no choice at all (Scheibehenne et al. Reference Scheibehenne, Greifeneder and Todd2010). Sylvia Plath poetically describes such perils (Plath Reference Plath1971: Ch. 7):
I saw my life branching out before me like the green fig tree in the story. From the tip of every branch, like a fat purple fig, a wonderful future beckoned and winked. One fig was a husband and a happy home and children, and another fig was a famous poet and another fig was a brilliant professor, and another fig was Ee Gee, the amazing editor, and another fig was Europe and Africa and South America, and another fig was Constantin and Socrates and Attila and a pack of other lovers with queer names and offbeat professions, and another fig was an Olympic lady crew champion, and beyond and above these figs were many more figs I couldn’t quite make out. I saw myself sitting in the crotch of this fig tree, starving to death, just because I couldn’t make up my mind which of the figs I would choose. I wanted each and every one of them, but choosing one meant losing all the rest, and, as I sat there, unable to decide, the figs began to wrinkle and go black, and, one by one, they plopped to the ground at my feet.
The above effects all suggest that sometimes, and in some domains, too much choice can make us worse off. But do these effects compound in a way that justifies the Negative Claim? I here investigate this question in two ways. First, I analyse whether the choice overload evidence is sufficiently robust to justify the Negative Claim. Second, I discuss whether there is direct aggregate or society-wide evidence.
3.4. How strong is the evidence?
For choice overload to support the Negative Claim, the empirical effects should be sufficiently robust and large. Let us look at two meta-reviews.
Scheibehenne et al. analysed 50 studies and found that the mean effect was virtually zero, with some studies showing no effect, others a positive effect (such that more choice improved outcomes), and others a negative effect (Scheibehenne et al. Reference Scheibehenne, Greifeneder and Todd2010). While the variance displayed was likely not random, the authors did not find sufficient conditions to explain it. In a later meta-review, Chernev et al. propose the following such conditions: decision task difficulty, i.e. structural characteristics such as time constraints, number of attributes per option, and presentation format; choice set complexity, i.e. features that relate to the chooser’s values and preferences, such as overall attractiveness of the options and their complementarity and the presence of a dominant option; preference uncertainty, i.e. the extent to which individuals already have preferences over existing options; and decision goal, i.e. what goal consumers have when looking at an option-set, for example, do they merely want to learn about option-sets, do some browsing, or do they need to actually buy something (Chernev et al. Reference Chernev, Böckenholt and Goodman2015: 336).
Overall, the meta-reviews suggest that choice overload is not strong and robust enough to establish the Negative Claim. Remember that the effect would have to be strong enough to outweigh the positive marginal value of more choice, including its non-specific value across time and in social contexts. Moreover, beyond very specific circumstances, the effect would have to be robust across choice contexts, which is not the case.
There are additional reasons why the choice overload literature does not warrant the Negative Claim.
First, besides robustness across choice domains and studies, there is also ‘inter-temporal robustness’: would greater choice lower wellbeing across longer time periods? Most studies observe behavioural or subjective effects in one instance and measure subjective effects short-term. However, temporarily struggling with more choice does not imply being worse off long-term. For some decisions, you can invest time and effort upfront but then reap benefits across many iterations. For example, imagine that, like Schwartz, you go to a shop that offers various styles of jeans, such as relaxed, casual, skinny, straight, and so on. After trying on some of them, you find one that is slightly better than the jeans you would have chosen had they had fewer options. Luckily, you can now keep buying the same style in the future such that those small benefits compound and outweigh the higher start-up costs.Footnote 10 Other decisions are not iterated but long-term. For example, when you have many study and career options, choosing well can be hard work and stressful. And sometimes the chosen career might only be slightly better than the one you would have chosen with less thinking or from fewer options. However, given how much time we spend working, the benefits of a (slightly) better career compound long-term and typically justify the start-up cost of considering more options.
Second, as the list of moderators for choice overload shows, it matters how choice-sets are presented and structured (also see Loewenstein Reference Loewenstein1999). Let me make one conceptual and one empirical point on this.
Informational access to one’s options – being able to find out what those are and how to access them – increases one’s freedom of choice (see Kramer Reference Kramer2003: 80, 421–25; Bruin Reference Bruin2022). Imagine, for example, the 75 million options on Amazon were entirely uncategorized and you had to click through items one by one. You would clearly have much less freedom of choice, considering the time, effort and opportunity costs involved in finding anything. Structured choice-sets facilitate better information and access to options and thereby give you more choice. So, arguments against larger choice-sets often just imply that we can enhance freedom of choice by structuring choice-sets better.
Empirically, companies in market economies often also have incentives to structure choices for consumers. Amazon would not stay in business very long if it did not have product categories and a search function and only offered uncategorized products (see Sugden Reference Sugden2018: 146 for this argument). So, markets provide pressure to make choices easier and convenient. (Of course, they also provide incentives to go against consumer interests, which I come back to in section 3.6.).
Finally, the choice overload literature is mostly (but not entirely) about consumption decisions. And even here, choice overload does not occur robustly. We should thus be hesitant to extrapolate choice overload to non-consumption choices. For example, in many countries, sexual norms today are much more liberal than they used to be. And dating apps further increase romantic options. For the debate between the Sufficiency View and Liberal Optimism, non-consumption choice domains matter too, yet the choice overload literature alone does not warrant strong judgements about them.
Overall, the current evidence on choice overload does not support the Sufficiency View.
3.5. What about aggregate effects?
Instead of analysing experimental data, supporters of the Sufficiency View might try and produce aggregate and society-wide data to show that too much choice either reduces wellbeing or at least fails to increase it. Here are four arguments sometimes given to this effect – along with responses as to why they don’t succeed.
First, Schwartz argues that increasing choice might explain increasing depression rates, what he calls a ‘depression epidemic’ (Schwartz Reference Schwartz2000, Reference Schwartz2009: Ch. 10). However, it is unclear whether there is an epidemic. In countries other than the United States, pre-pandemicFootnote 11 depression rates have stayed mostly stable, whereas the reported increase for the USA might be due to measurement issues (Baxter et al. Reference Baxter, Scott, Ferrari, Norman, Vos and Whiteford2014).
Second, various authors, including Schwartz, suggest there is a trade-off between freedom of choice on the one hand and community and social relations on the other. Arguably, community and involved relationships make demands on our personalities, behaviour, commitment and time. In this sense, they restrict how far we can just ‘do what we want’. If there is a trade-off between choice and community, Liberal Optimism would have a problem, as social relationships contribute to life satisfaction and happiness.
Unfortunately, I have not seen good data or arguments on this trade-off. Along with Schwartz, we could of course make an intuitive case. Greater freedom of choice loosens communal ties because many such ties require constraints on choice. For example, fewer people attend religious ceremonies in Western democracies. But even this intuitive case is not clear-cut. Freedom of choice also affords one greater freedom to choose which social relations and communal ties to enter, build up or sustain, which in turn begets new social relations and communities and might improve the quality of our relations.
Of course, numerous media articles claim we are in the midst of a loneliness epidemic. However, the evidence for this epidemic is shaky at best. There is a rise in people living alone and spending time alone. However, aloneness and loneliness are not the same. The share of people who report being lonely seems not to have gone up and, in part, even declined (see Ortiz-Ospina Reference Ortiz-Ospina2019 for more).
Finally, Schwartz also argues that for all this expanding choice over the last decades, happiness seems to have remained constant (see Schwartz Reference Schwartz2009: Ch. 5 and references therein). Our choices have skyrocketed but our happiness stayed the same.
However, the data here again are less obvious than they seem. The claim is mostly made about the United States. Most other countries – liberal countries in particular – have reported rising subjective wellbeing over the last decades (Inglehart et al. Reference Inglehart, Foa, Peterson and Welzel2008; Ortiz-Ospina and Roser Reference Ortiz-Ospina and Roser2013).
But a proponent of the Sufficiency View might point specifically to the relationship between income and happiness. As more income means more choice, we should expect to see more happiness as a result of growing incomes (Schwartz Reference Schwartz2009).Footnote 12 Yet the well-known Easterlin Paradox suggests that while average economic income improves wellbeing up to a point – a point long surpassed in rich countries – any further income does not increase happiness (Easterlin Reference Easterlin1974).
However, recent research – which also goes beyond US data – finds that GDP and life satisfaction likely do correlate, although they do so logarithmically (Stevenson and Wolfers Reference Stevenson and Wolfers2008, Reference Stevenson and Wolfers2013; Sacks et al. Reference Sacks, Stevenson and Wolfers2012).Footnote 13 Across countries, any doubling of income – even in rich countries – seems to increase happiness equally. And while logarithmic growth seems slow, sustained economic growth is exponential such that doubling a country’s GDP is achievable within decades (Cowen Reference Cowen2018).Footnote 14
I have argued that there is no society-wide evidence that supports the Negative Claim. Discussions around income even suggested that, if anything, rising incomes are more likely to increase wellbeing. But, in any case, freedom of choice is more than income. Beyond average GDP, choice increases wellbeing in other ways. And there is indeed good positive evidence that more freedom of choice produces more wellbeing. Here are two examples.
First, Inglehart et al. (Reference Inglehart, Foa, Peterson and Welzel2008: 264) perform regression analyses on time series data on 56 countries and conclude that ‘the extent to which a society allows free choice has a major impact on happiness.’ The dependent variable was a composite of both life satisfaction and subjective happiness and the independent variable was ‘perceived freedom of choice’ (i.e. surveying to what extent people feel they have free choice and control over their lives). Perceived freedom of choice reflects not just economic conditions but strongly correlates with how democratic and socially liberal a country is (for example, how far it accepts liberal norms around gender and same-sex relations).
Second, in a book-length study, Bavetta et al. (Reference Bavetta, Navarra, Maimone and Patti2014) examine how far freedom – in its different conceptualizations – is conducive to wellbeing. Using several different measures of freedom of choice, they conclude that it robustly correlates with wellbeing.Footnote 15
So, we should conclude that neither choice overload research nor any aggregate societal trends justify the Negative Claim. Given general arguments to value choice (from section 2) along with evidence that choice and wellbeing move up together, we should accept Liberal Optimism instead.
3.6. Can Liberal Optimism ignore choice overload?
The literature on choice overload still contains lessons for Liberal Optimism. Liberals should think about how to best increase choice and supplement it with other policies, norms and interventions that help people get even more wellbeing out of the choices they have.
First, how options are categorized and information displayed can make a big difference. And markets often provide incentives to categorize options better and reduce search costs for consumers. However, sometimes companies have incentives to do the opposite and design choice-sets that are predictably bad for consumers (see (Akerlof and Shiller Reference Akerlof and Shiller2015) for many examples). And sometimes bad choice architecture is just unintentional.
Where effective – and otherwise justifiable – nudge interventions keep what is valuable about larger choice-sets but also nudge people towards better decisions.Footnote 16 For example, employees are often less likely to sign up for a retirement plan if they must choose from a greater number. A classic nudge switches the default such that employees are signed up unless they opt out. Another range of interventions tries to simplify choices by testing how information is displayed or by sending individuals reminders.
Second, we should sometimes design choice environments so they include the freedom not to choose. For example, when they know little about different options – in medical contexts for example – people sometimes prefer that someone more knowledgeable decides for them. Some interpret this as a preference for ‘less choice’. But the preference can actually be met by expanding their choice-set: if I have the freedom to choose myself and the freedom to have an expert choose for me, then I have more freedom than if I only have the freedom to choose myself. More generally, valuing choice is compatible with giving people the option to ‘choose not to choose’ (Sunstein Reference Sunstein2015). For the same reasons, I would add ‘choosing to choose from less’. Imagine you hire an interior decorator to give you ten great options for your living room. Being able to hire a consultant increases rather than reduces freedom of choice. Or imagine I am faced with hundreds of healthcare plan options, many of which are good, others mediocre, and others terrible. My freedom of choice is increased, if I can choose to have an expert provide me with a much smaller number of pre-selected really good options. Conversely, if receiving good healthcare is conditional on countless hours weeding out bad healthcare plans, my freedom of choice is lower.
Third, the literature on choice overload might also contain lessons on how we should make decisions. Decision-making with ever-expanding choice-sets requires skills and adjusting one’s expectations (Schwartz Reference Schwartz2009). There is also much work on how to make decisions in complex environments with many choices (Gigerenzer et al. Reference Gigerenzer, Todd and Research Group2000; Christian and Griffiths Reference Christian and Griffiths2016). For example, in complex environments, simple heuristics can outperform more complex forms of decision-making. And given expanding choices, it is often important we learn how and when to satisfice rather than maximize in our choices. If we get better at making decisions, we can squeeze more wellbeing out of our greater freedom. This also has a collective and political dimension. Better choice architecture might help but so might better education that teaches individuals how to make better choices in complex environments.Footnote 17
Let me end this section with two clarifications.
First, some authors writing on the ‘tyranny of choice’ oscillate between individuals having more choice and what we might call the ideology of choice. The latter includes tropes such as ‘consumerism is the road to happiness’, ‘if you fail you only have yourself to blame’, ‘you must work hard so you can buy more stuff’, and so on.Footnote 18 Liberal Optimism is committed to expanding choices but not to this ‘ideology’. Moreover, freedom of choice includes but goes far beyond consumption options. Strong social pressure to lead your life a certain way – such as work all day so you can buy more stuff – would limit choice by narrowing the possible lives individuals can lead. As such, liberal optimists will want to reduce such ideological pressures.
Second, Liberal Optimism views expanding choice as a weighty pro tanto goal. It does not imply that expanding choice is the only or always the most effective way to increase overall wellbeing. Governments and societies should also pursue more targeted actions, such as better diagnosis and treatment of mental illness.
4. The Future of Freedom: How Much More Choice is Possible and Desirable?
In section 2.2, I argued that choice has non-specific value, particularly as a social value. Those arguments, I suggested, make a strong presumptive case for Liberal Optimism. In section 3, I responded to the worry that too much choice reduces our wellbeing. It turned out that, if anything, the evidence suggests the opposite: more choice is good for wellbeing.
However, one might challenge my arguments so far: that more choice is valuable within our current range does not guarantee its value beyond this range. The Sufficiency View might still be right: level L might simply be the upper bound of choice within our current range (or close to it). If so, our current data do not show we should expand choice much beyond L.
However, I now present four reasons to think further expansions are both possible and valuable (in expectation). Again, it is not just that some additional options will have specific value. The Sufficiency View can agree with that. Instead, further choice extensions are possible and non-specifically valuable, that is, valuable because they expand choice.
First, one reason to value freedom of choice is that it allows us to find new ‘ends’, conceptions of the good, and inspirations for living. Moreover, without being challenged by people who live differently, inherited norms and traditions can sometimes appear necessary or even ‘natural’ rather than as something whose value can be challenged. Now, given that choice helps us discover new sources of value, it is hard to say, ex ante, which choice expansion will or will not have value. Conversely, we should be sceptical, at least prima facie, of anyone professing to know when we have reached the sufficiency level.
Call the second argument The Inductive Argument: imagine people in the past made claims that their level of choice back then was sufficient and that any further increase would lack non-specific value. With hindsight, such claims seem implausible now. For example, Dworkin wrote his article in 1982. Take a typical Northern European citizen in 1982. Their life expectancy was shorter than it is today, GDP per capita was 40–45% less, and social norms were far less open to diverse forms of living, particularly considering LGBTQ issues and gender. Moreover, globalization and the internet have expanded cultural options, such as food, music and films. And not only can citizens travel much more easily, within the European Union, they can work and live wherever they want. Now, a sufficiency claim indexed to choice levels in 1982 might have seemed plausible then, but it does not seem plausible now. For inductive reasons, we should therefore be hesitant about sufficiency claims indexed to our current or near-term future choice level.
The Inductive Argument achieves its full effect, however, if we take a longer-term perspective and realize what a strange time of human history we are in, as shown by Figure 1.
World GDP over the last two millennia.
Imagine someone made the sufficiency claim indexed to the level of choice available a few centuries ago. Such a level seems absurdly modest today. Similarly, considering how fast the world is changing, future generations might find ‘sufficiency levels’ indexed to our time weirdly modest too.
The third argument I call The Skipping Argument. In a recent study, respondents (mostly US-American) were asked at random times: ‘If you could, and it had no negative consequences, would you jump forward in time to the end of what you’re currently doing?’ (Killingsworth et al. Reference Killingsworth, Stewart and Greene2020). On average, respondents wanted to skip around 40% of their day. When prompted, people wanted to skip work 69% of the time, commuting and travelling around 60% and housework around 50% of the time. Playing, reading a book, watching television and going for a walk were far more popular, and ‘making love’ came out on top (1.5%). On average, respondents also said that when leading a good life, one would not want to skip more than 15%, far lower than the 40% people actually wanted to skip. Now, imagine future choice expansions would allow us to skip many of those things we have to do but rather wouldn’t. Maybe such progress would massively cut work hours or reduce hours spent on tedious things such as housework, commuting or ‘life admin’. There are also some incentives for those innovations to happen: given that (enough) people wish to skip those experiences, there is demand for innovations to ‘liberate’ us from some of them. This choice extension would be very valuable in expectation, as it would free up more time for activities we actually want to be awake for.
The fourth argument simply lists candidate areas where further extensions seem possible and, if successful, non-specifically valuable. Here are some.
First, medical progress can likely further expand our choices by increasing human capabilities and lowering mortality and morbidity. Maybe interventions to slow down ageing will be successful and maybe biotechnology, DNA, RNA and mRNA technology will reduce illness and extend our capabilities. While often subject to ethical debate, pre-natal interventions and human enhancement might also expand our choices. Moreover, in absolute numbers, more time and resources, computing capabilities, and researchers and developers are dedicated to medical progress and public health. So, even beyond any individual technology, we should expect further progress in medicine and public health just from increased capabilities alone.
Second, we might see continued economic growth, technological progress and greater productivity. Artificial intelligence might also boost productivity. Besides potential for higher incomes, growing productivity – and a good distribution of the spoils – might allow individuals the freedom to cut down work hours. This would free up time that people can freely decide how to use. Having time is an important contributor to freedom, as time is required for any activity one pursues (Goodin et al. Reference Goodin, Mahmud Rice, Eriksson and Parpo2008).
Third, even liberal countries have probably not reached the pinnacle of social liberalization. Many social norms might still prevent people from experiments in living and unduly restrict their freedom to be the person they want to be. For example, gender roles still impose unnecessarily specific expectations about what it means to be a woman or man. Particularly for women the process of expanding choices seems far from over. Moreover, many argue that moving beyond binary gender expectations will give us even greater freedom to carve out our own identity and lives. Our freedom of choice in romantic and personal relationships might also further expand. For example, some argue society should be more accepting of non-monogamous and polyamorous relationships: even in today’s liberal societies, heteronormative and monogamous expectations still restrict what kind of romantic and sexual relationships people can enjoy. Similarly, there are pervasive pressures towards the nuclear family model. More options around family living might be possible and valuable. Finally, we might expand people’s choices by allowing for more international mobility and by weakening the rigidity that comes from current identities built around traditional cultural groups and nation states. Already, many in the international elite enjoy greater freedom to carve out their own personality and to decide in which country to study, travel and live, along with a greater freedom not to have one’s identity be defined by one’s cultural, ethnic or national group. Reducing the normative pressures that come from cultural and national identity might expand such freedoms to more individuals. Of course, these are just quick speculative comments. But at least some such areas of social liberalization are likely to offer valuable avenues for greater choice.
So, more choice is still possible and, in expectation, valuable. Note that I have not argued that more choice is inevitable or even likely. Humans might mess it up. Liberal Optimism does not imply an empirical belief that humans will preserve and increase choice – it only says they should.
5. Choice and Autonomy
I have defused the worry that excessive choice reduces wellbeing and have presented a positive case for choice, both at current levels but also at levels far higher than that. Those arguments now also allow me to give a brief response to another argument sometimes hinted at for the Sufficiency View:
(A) Respecting and promoting individual autonomy is a central duty for institutional design, public policy, and for societies more generally.
(B) There is no reason besides autonomy to value freedom of choice, or there is no relevant duty in institutional design, public policy, or for societies more generally, besides the duty to respect and promote individual autonomy.
(C) The duty to respect and promote individual autonomy implies a central duty to ensure that individuals have an adequate range of freedom of choice. But it does not imply a duty to promote freedom of choice beyond that range, because additional choice above that range does not improve autonomy – if anything, it is more likely to reduce it.
(D) Therefore, there is a central duty in institutional design, public policy, and for societies more generally to ensure that individuals have an adequate range of freedom of choice but no duty to promote freedom of choice beyond that range.
Joseph Raz and Michael Blake endorse an argument along these lines (Raz Reference Raz1986; Blake Reference Blake2001).Footnote 19 How convincing is it?
First, we should probably reject premise B. As argued above, freedom of choice is an important social value for reasons other than autonomy. Moreover, while many would agree that autonomy should be an important societal and institutional value, few will think it should be the only one.Footnote 20 And once we include other likely candidate goals – such as preference-satisfaction, hedonic wellbeing or simply uncertainty about the good – freedom of choice seems back on the table.
However, beyond this external critique, let us see how well the argument fares ‘internally’. In particular, how plausible is premise C?
C’s plausibility in part depends on what we mean ‘by autonomy’. Distinguish first a narrow and a broad use. In a narrow sense, autonomy is about whether a person’s pro-attitudes – such as their desires – are truly their own or, at least, not externally imposed. Call this pro-attitude autonomy. For example, a heroin addict will experience urges that undermine her autonomy or a religious cult might use manipulation and social pressure to impose desires on her. What I call broad autonomy is about the overall conditions for individuals to lead autonomous lives and this is what Raz and Blake are after. Blake holds that for agents to live as autonomous agents, they must be ‘capable of selecting and pursuing plans of life in accordance with individual conceptions of the good’ (Blake Reference Blake2001: 271). Pro-attitude autonomy is necessary but insufficient for broad autonomy, one also requires external choice and the absence of coercion (these days, many might write ‘absence of coercion and domination’; see Schmidt and Engelen Reference Schmidt and Engelen2020: sec. 4.1).
Note first that even when expanding choice would leave pro-attitude autonomy unchanged, it can still increase broad autonomy. For example, imagine my coherent set of autonomous desires revolves around my passion for dancing and is not due to irrationality, manipulation or domination.Footnote 21 Assume that adding external options would not make my goals any more autonomous than they currently are.Footnote 22 However, my overall autonomy could still be improved. To make a living as a dancer, one must be both very good and very lucky. Imagine I now win the lottery and no longer require a wage: my options to spend my life dancing – and thus pursuing my conception of the good – have vastly improved. Arguments for C require evidence that more choice would fail to improve broad autonomy rather than just pro-attitude autonomy. Moreover, we require evidence not just that it fails to do so in individual instances but does so systematically.Footnote 23 Our best bet is to return to the choice overload literature, specifically the link between choice and decision deferral: as several studies suggest, individuals sometimes put off making a decision when options increase. However, one problem is that such studies judge short-term decision deferral as a negative outcome or even equate it to ‘choice paralysis’. That would make sense if decision deferral becomes systematic and vitiates pursuing one’s conception of the good, as it did in the Sylvia Plath quotation above. However, sometimes – particularly for important decisions – browsing, gathering information and postponing decisions can be perfectly rational and strengthen our broad autonomy. A fine-grained normative analysis of broad autonomy should not view all non-decisions as negative (also see Anderson Reference Anderson2003). Add to this that choice deferral only holds with specific moderators present and that the literature’s general robustness is somewhat low, and we are unlikely to find strong support for premise C here.
What is more, any forthcoming arguments for C must meet a high bar and show that the negative marginal effect on broad autonomy outweigh positive marginal effects. For I now argue that, mutatis mutandis, my earlier arguments suggest that more choice improves conditions for broad autonomy.
First, I argued that choice has non-specific value, which involved arguments from exploration, experiments in living, uncertainty and disagreement about the good, people changing in what they value and like across time, and others. Absent strong arguments to the contrary, those apply to broad autonomy too. Moreover, non-specific value justifies some general scepticism about confident pronouncements about when a sufficiency level has been reached.
Second, I argued that further choice expansion is possible and in expectation valuable. The case extends to autonomy too. First, past judgements about how much choice was sufficient seemed implausible later on. Claims that nearby levels have reached a sufficiency level for broad autonomy should be met with some healthy scepticism too (the Inductive Argument). Second, future choice expansions might allow us to skip things we must do today but rather wouldn’t. This would leave more time for activities that further our conceptions of the good (the Skipping Argument). Finally, in several candidate areas, further choice expansion seemed valuable and, prima facie, those seem to improve conditions for autonomy too: liberation from work, further social liberalization, higher incomes, further reductions in morbidity along with greater human capabilities and longer life spans all seem to improve people’s conditions to lead autonomous lives. For example, think about how many people’s life projects and dreams are thwarted by illness and death. Or consider how much time is taken up by work people find neither fulfilling nor meaningful. With more healthy life years and more freedom over their waking hours, people would have greater opportunities to pursue their conceptions of the good.
6. Some (More) Applications: Community, Growth, Markets and Migration
I have defended Liberal Optimism: it should be an important pro tanto goal for our societies to increase people’s freedom of choice even beyond already high levels. I also offered ‘applied’ examples where future expansions seem possible and, in expectation, valuable. I now give three examples of how my conclusions – in broad brushstrokes rather than worked-out detail – also matter for more near-term applied discussions.
6.1. Freedom of choice and community
There is often talk of a trade-off between freedom of choice and the value of community, the latter of which should get priority seeing that we already have enough or even ‘too much’ choice. If I am right, we should be cautious of this narrative: it is neither obvious there is this systematic let alone inescapable trade-off nor that we currently have enough choice.Footnote 24 We should be sceptical of quick ‘community-based’ arguments, particularly when used to defend traditional or hierarchical community norms that restrict choice by imposing narrow cultural and gender roles.
6.2. Growth and markets
First, should rich countries continue to grow their economies? There is growing scepticism about whether all this growth is doing us any good. With a climate catastrophe looming, some even call for de-growing the economy. Liberal Optimism creates a strong pro tanto argument to continue economic progress. Of course, such an argument is not decisive, and it must factor in environmental externalities of economic growth as well as uncertainties about how far we can de-carbonize growth. Still, Liberal Optimism gives us pro tanto reason against giving up growth too easily: not extending our choices, let alone reducing them, is a serious cost.Footnote 25
Second, a classic argument for markets is that they increase consumer choice. Critics respond that we should interfere or even get rid of markets, because the ever-expanding choices do not make us better off, the contrary if anything (also see Sugden Reference Sugden2018 for a discussion). Given Liberal Optimism, we should be sceptical of this response. That functioning markets expand choice, even beyond high levels, weighs in their favour. Of course, such an argument is only pro tanto and empirically contingent. Moreover, Liberal Optimism by itself does not imply free market capitalism. While liberal optimists should reject the ‘too much choice objection to markets’, they must still consider the many other worries that critics of markets have raised.
Let me now add some nuance to both arguments. ‘Growth’ is not a value-neutral concept. Moreover, GDP (per capita) is – if anything – an imperfect proxy for whatever type of economic progress we should pursue after weighing all relevant normative reasons. Even if we focus on increasing GDP, deep normative questions remain about how to do so. I lack the space to detail the specific type of economic progress that Liberal Optimism would favour. Nonetheless, I argue that it likely sidesteps the primary concerns often raised about growth and consumer choice.
One central worry is that growth has negative externalities. For example, expanding choice in cheap meat is a key driver of climate change and responsible for most excess land use and degradation (plant-based diets require much less land). Moreover, it has severe opportunity costs: if densely populated countries like the UK or the Netherlands use so much of their land or greenhouse gas budget for animal husbandry, they forego opportunities to use that land and budget for other options, such as nature reserves, housing and innovative industries. However, Liberal Optimism captures these concerns by focusing on distributions of choice: options with heavy negative externalities reduce other people’s choice, either now or in the future;Footnote 26 and severe opportunity costs are captured by ranking overall distributions too.
The second worry is inequality. Since the mid-1970s, real income gains have disproportionally gone to higher income brackets – something that GDP does not capture. Liberal Optimism, however, considers how choice is distributed. And most plausible distributive criteria would, other things equal, prefer a more equal distribution of choice and, thus, income gains.Footnote 27
A third worry is that so many additional consumer options seem worthless: they simply don’t do much to make us better off, particularly compared with other important goods. For example, while access to disposable electronics and fast fashion has gone up, freedom of choice in housing, health care or education has not kept pace and, partly, even declined in some rich countries. Housing costs, in particular, have skyrocketed. Is this the growth we should want?
The Hybrid Measure implies that expanding consumer choice does increase freedom of choice (other things being equal). I think this is plausible. Still, it also implies that some options contribute far more to freedom of choice than others. Remember my earlier example: at the margin, the 50th brand of bottled water seems to add neither value nor new follow-up options. This is different for increasing options in affordable and high-quality housing: not only is housing a basic need, it also opens up many other (valuable) options, like access to good jobs, a stable home, better social life and space to raise children. Similarly, education increases one’s valuable options down the line, health is a prerequisite for most other future options, and travel and mobility expand the number and quality of cultural, leisure and economic options. So, Liberal Optimism captures that – compared with ‘trivial’ consumer options – fecund and valuable options matter more.
The final worry is that markets generate ‘bad’ options. Should liberals really welcome access to opioids and payday loans? However, when discussing paternalism, we saw we can both non-specifically value choice expansions whilst also restricting specific options. More importantly, bad options often reduce expected lifetime freedom. Opioids or cigarettes typically reduce my future options, as do exploitative payday loans that burden my future self with spiralling debt. So, typically, Liberal Optimism gives us good reason to constrain options that reduce lifetime freedom.
Liberal Optimism gives us a pro tanto reason to broadly favour growth and expanding options in markets. At the same time, it mitigates central worries around the ‘wrong type’ of economic progress.
6.3. Immigration
In discussions around immigration, some philosophers argue that freedom of movement does not support a right to immigrate. People only have a right to enough freedom of movement – or a right to ‘sufficient’ or ‘adequate’ options more generally – which is fulfilled when people are free to move within a country (Miller Reference Miller2007: 207; Pevnick Reference Pevnick2011: 83–85). Some even argue that immigration restrictions can support individual autonomy when they preserve a national culture that provides a ‘context of choice’ (see Patten Reference Patten2004 for discussion). Liberal Optimism, in contrast, provides a pro tanto argument to reduce immigration restrictions: appeals to ‘sufficient’ or ‘adequate’ options are not enough to dismiss such freedom-based arguments (also see Hidalgo Reference Hidalgo2014). Of course, this argument in no way settles whether states have a right to control immigration or individuals a right to immigrate. But Liberal Optimism does suggest that the value of choice generates pressure towards greater freedom of movement across borders.
7. Conclusions
In this article, I have defended Liberal Optimism: it should be an important pro tanto goal for our societies to increase people’s freedom of choice even beyond already high levels. I presented reasons why choice has non-specific value and why this makes for a presumptive case for Liberal Optimism (in section 2), responded to ‘choice overload arguments’ and argued that, if anything, more choice increases rather than reduces wellbeing (section 3), and argued that even in today’s rich liberal countries, further significant increases in choice are both possible and valuable in expectation (section 4).
Therefore, we should reject the Sufficiency View according to which increasing choice matters only up to some existing or nearby sufficiency level. Neither paternalistic arguments, nor research on choice overload, nor any aggregate effects support the Sufficiency View. While I responded mostly to arguments revolving around wellbeing, I suggested the case extends to autonomy too.
Accordingly, we should also push back against cultural critics who see all that new choice as either pointless or even harmful. Moreover, my arguments imply we should set greater store by freedom of choice in applied discussions. I outlined several areas where I expect future expansions of choice will be possible and, in expectation, valuable. I also argued that freedom of choice generates pro tanto arguments for continued economic progress – of a type that improves the overall distribution of freedom – and fewer restrictions to migration.
Acknowledgements
I would like to thank Brian Berkey, Laetitia Mulder, Frank Hindriks, Lisa Herzog, and two referees for very helpful written comments. I should also thank audiences for their many helpful comments and questions at the panel on freedom at the annual ECPR conference, Ethics and Public Policy Working Group of the American PPE Society, the Groningen Wharton PPE workshop in Groningen, and the OZSW Workshop on Behavioural Paternalism, Rational Choice, and Public Policy in Amsterdam.
Competing interests
I declare that there are no conflicts of interest and that the submitted article complies with all of the publisher’s ethical standards.
Research ethics
Not applicable, as the article does not produce original empirical research.
Andreas T. Schmidt is Professor of Moral and Political Philosophy at the Faculty of Philosophy and the Centre for PPE in Groningen. He works in political theory, normative and applied ethics, and the philosophy of public policy. He is particularly interested in longtermist political philosophy, freedom, consequentialism, inequality, liberalism, behavioural policies, public health and biomedical ethics.
Open access
