Capitalism and the Very Long Term

1 Introduction

Capitalism is defined as the economic structure in which decisions over production are largely made by or on behalf of individuals in virtue of their private property ownership, subject to the incentives and constraints of market competition. Command economies are not capitalist, because production decisions are made in virtue of political power with no market competition; certain market socialist arrangements are also not capitalist, since production decisions are made by virtue of (for instance) individuals’ roles in worker cooperatives, even if they are subject to market constraints and incentives. Feudal economies are also not capitalist, since decision-making power largely depends on birthright rather than ownership of private property that can be exchanged in markets.

Note that in this definition, capitalism is a gradable concept: an economy can be more or less capitalistic depending on the extent to which production is determined by market-constrained private owners. ‘Pure’ capitalism, in which all production is determined by market-constrained private owners, is a fiction; in the real world other economic actors (such as states) are always significantly involved in production. However, the economic structure in which we find ourselves may reasonably be called ‘capitalism’, as ‘shorthand for something like “a hybrid economic structure within which capitalism is the predominant way of organising economic activity”’. (Wright 2010, 125) In this paper, I will argue that considerations of long-term welfare support anticapitalism in a weak sense (reducing the extent to which the economy is capitalistic) and perhaps support anticapitalism in a stronger sense (establishing an alternative economic structure in which capitalism is not predominant).

Why consider long-term welfare? I assume that the evaluation of economic structures should depend, at least to some extent, on their effects on people’s welfare. When will most of the effects of a present structure occur? A plausible answer is: very far into the future. There are currently fewer than 10¹⁰ human beings. Hillary Greaves and Will MacAskill estimate that there are in expectation 10¹⁴ people to come – 10 thousand times more – on their most conservative estimate,^[1] which ignores the possibilities of space settlement or the creation of human-like artificial intelligence (2021, 8–9). The lives of future people matter no less than our own. They will also feel pain and joy, lust and envy, hope and fear. So, when evaluating economic systems we should take into account their likely effects upon all of the lives we expect to come. Many more of these lives will be in the very far future than in the present or near future. Therefore, most of the welfare effects of our economic structures will be on the very far future.

So says a ‘longtermist’ approach to the evaluation of economic structures.^[2] Longtermism, a branch of effective altruism associated with philosophers such as Greaves and MacAskill,^[3] holds that the best way to benefit others is by taking actions that bear on the very long-term future: hundreds, thousands, and perhaps millions of years away. Longtermists should therefore be sensitive to the ways in which current economic structures promote or impede long-term welfare.

My arguments do not apply only to card-carrying longtermists, however. In two ways my starting points are weaker – harder to deny – than those typical of longtermists and effective altruists. Firstly, I focus on economic systems, not individual actions. It is more plausible that economic systems have significant effects in the very long term than it is that individual actions in the present do.^[4] The effects of the latter might ‘wash out’ (Greaves and MacAskill 2021, 9–10), whereas those of the former are likely to be greater in scale and more systematic in direction. The washing out of individual actions would undermine longtermism but not my conclusion that considerations of long-term welfare support anticapitalism. Moreover, effective altruism has been criticised for its emphasis on individual action (Collins 2019; Dietz 2019; Srinivasan 2015); those sympathetic to this criticism may also be persuaded towards anticapitalism by my arguments. Secondly, I only assume that we should take the very long-term effects of economic structures into account when evaluating them (perhaps amongst other properties such as their short-term effects or their justice), not that the very long-term effects of those structures are the most important effects they have. Reflection on the scale of the future might make the former claim seem plausible, even to those who reject the latter, and thus reject longtermism in its stronger forms.

As I have said, I will argue for the two following claims.

1. Considerations of long-term welfare support anticapitalism in a weak sense: reducing the extent to which the economy is capitalistic.

2. Considerations of long-term welfare support anticapitalism in a stronger sense: establishing an alternative economic structure in which capitalism is not predominant.

I am more confident in the weaker claim than in the stronger one, for reasons to be discussed in Sections 3 and 4. Even my arguments for the weaker claim fall short of a conclusive proof of the long-term inefficiency of capitalism (inevitably, given the complexity of the issue). I hope that my arguments convince some readers of these two claims, but even if they do not, this paper makes two important contributions. Firstly, it should reorient longtermists, who typically, like other effective altruists, have been apologists for (or silent about) capitalism. By making the best case I can that capitalism is ineffective at producing long-term welfare, I hope to convince longtermists to at least take anticapitalist thought and politics more seriously and to inspire further discussion (which will include, I am sure, defences of capitalism). Secondly, I hope to reorient anticapitalists. Capitalism is often criticised on the grounds that it is exploitative, alienating or inegalitarian. In the past, it was often criticised for being inefficient, in contrast to more ‘rational’ systems such as socialism. These criticisms are less common today. Indeed, one often hears a defence of capitalism along the following lines: ‘It might be exploitative, alienating and inegalitarian, but at least it generates welfare.’ Such defences often point to falling absolute poverty and rising life expectancy over the history of capitalism, and the failures of alternative systems in the twentieth century. Anticapitalists respond that efficiency with respect to welfare is an insufficient defence of an unjust system. I do not think, however, that anticapitalists need to concede the efficiency of capitalism so readily, given the importance of long-term welfare. Or so I will argue.

In Section 2, I will give two economic arguments for the claim that capitalism will be systematically inefficient at promoting long-term welfare and respond to the objection that its tendency to fast growth outweighs such concerns. Correcting these inefficiencies requires reducing the extent to which our economy is capitalistic – and thus considerations of long-term welfare support anticapitalism in this weaker sense. In Section 3, I suggest that whilst capitalism remains predominant such corrections will be difficult to impose and maintain. In Section 4, I give reasons to think that different economic structures could outperform capitalism. Together, these sections make a case on longtermist grounds for a stronger form of anticapitalism: the replacement of current economic structures with others in which capitalist decision-making is not predominant.

2 The Economics of the Long-Term Under Capitalism

In this section I provide two reasons to think that capitalism will fail to efficiently promote long-term welfare. Firstly, far future people’s interests are an externality, so are not reflected in market prices. Secondly, capitalist firms will employ higher than optimal discount rates. Both suggest that capitalism will be inefficient at delivering long-term welfare. I then respond to an objection that these inefficiencies are compensated for by capitalism’s tendency to produce fast economic growth.

3 Can Capitalism Accommodate Longtermist Reforms?

It is common knowledge that markets can fail and that capitalism requires a degree of intervention from noncapitalist forces, such as the state, to be stable and efficient. Many such interventions could be made perfectly consistent with capitalism maintaining its predominance in the overall economic structure. For instance, states can outlaw or tax the production of negative externalities and mandate or subsidise the production of positive ones. Such policies could in theory incentivise capitalists to invest as if future people’s interests mattered to them. This might involve subsidising research into catastrophic risk mitigation. It is harder to see how regulation alone could prevent over-discounting, although there are policies which could lengthen firms’ time horizons somewhat, such as imposing fiduciary duties to secure long-term goals or incentivising investors to hold their shares for longer periods of time.^[9]

These interventions may be said to reduce the extent to which the economy is capitalistic, since they involve the state taking decisions that influence production. They do so at one remove: capitalists would still control direct production decisions, subject to the market, but the market conditions would be deliberately shaped by the state to favour some kinds of production over others. This already involves the state stepping beyond a ‘nightwatchman’ role of simply enforcing the rules needed for capitalism’s markets to function. States – or other noncapitalist actors – can also involve themselves directly in production to solve some of the problems raised above. For instance, the state could itself produce positive externalities and direct a catastrophic risk mitigation programme, rather than providing subsidies and regulations to shape a market to do so. The state could nationalise, or place under noncapitalist control, industries that are particular drivers of catastrophic risk: arms, biotechnology, the nuclear sector, artificial intelligence. And the state, or other noncapitalist actors, could invest in long term projects that capitalists refuse to, due to their discount rates. We can imagine a ‘mixed economy’, with capitalist production in some areas, but a reasonably large noncapitalist sector looking after the longer term.

Considerations of long-term welfare support at least some such interventions, since they will correct the inefficiencies of capitalism identified in the previous section. Few of these measures are currently being taken: many negative externalities, especially long-term ones, such as greenhouse gas emissions, are undertaxed; many positive ones are undersubsidised; firms are increasingly short-termist (see Haldane 2015 and sources cited therein), and the development of technology with potentially large long-term impacts, such as artificial intelligence, is currently largely left to capitalists. Since imposing these interventions would make the economy less capitalistic, this provides a case for anticapitalism in a weak sense: a reason, based in long-term welfare, to reduce the extent to which the economy is capitalistic.

However, this would be consistent with capitalism remaining predominant in our overall economic structure. I now want to argue that considerations of long-term welfare support anticapitalism in a stronger sense, favouring alternative economic structures in which capitalism is not predominant. The first part of this argument, which I will make in this section, is that as long as capitalism is predominant, the kinds of intervention mentioned above will be difficult to impose and maintain. The second part, which I will address in the next section, is that we can imagine alternative economic structures doing better with respect to the long term.

The following arguments assume that capitalism produces a class society in which capitalists form a powerful elite. This goes beyond the purely economic definition of capitalism presented at the start of the paper. It may be conceivable that a system could be capitalistic in the sense of that definition without granting capitalists class power: for instance, Rawls’s suggestion of a ‘property-owning democracy’ (Rawls 2001; for discussion, see O’Neill and Williamson 2012). The possibility of such structures presents a limit to my argument here; however, I do not think it is a serious one. As a matter of historical fact, capitalism has always produced class societies. Moreover, if property-owning democracy is possible, then the arguments in this section may be read as arguments for either anticapitalism in its stronger form or property-owning democracy (and the arguments for my weaker anticapitalist claim would still hold).^[10]

Given such a class society, then, why will capitalism’s predominance make the necessary regulatory and mixed-economy solutions difficult to impose and maintain? In short, because one can expect them to be resisted by capitalists. Compared with the average voter, capitalists tend to favour less regulation and less taxation (the latter a problem both for externalities taxes and for raising funds for public investment in longtermist projects).^[11] Since they are wealthy, they also tend to favour policies that will increase inequality, which many suggest is negatively correlated with society’s capacity to produce public goods, including those such as existential risk mitigation that are sought by longtermists (Chong and Gradstein 2007; Savoia, Easaw, and McKay 2010; Schmidt and Juijn 2024).

Whilst capitalism is predominant in the economy, capitalists, with these anti-longtermist policy preferences, will wield disproportionate political influence. One reason is that capitalists, in virtue of their economic power, can cause major disruption and hardship through actions varying from capital strikes to funding rebellions and coups. Another is that capitalists have disproportionate influence over the opinions of the population; in the Marxian formula (Marx and Engels [1846] 2000, 192), their control of the means of material production gives them control of the means of mental production (they directly own much mass media and hold significant leverage, via their economic power, over academic and educational institutions). Thirdly, capitalists’ resources give them the capacity to influence particular political actors through lobbying, bribery and campaign funding.^[12] For all these reasons, we can expect that longtermist policies will be frustrated so long as capitalists predominate.

Of course, there have been times when capitalist interests have lost out in politics, and states under capitalism have imposed regulations which capitalists opposed. The grip on power of the capitalist class is not total. As even Marxists say, ‘the capitalist class rules but does not govern’ (Kautsky 1903, 13; for discussion, see Elster 1985, 398–428; and Miliband 1969). The work of government is not usually done by business owners nor explicitly on their behalf: it is done by professional politicians who have their own interests and are subject not to market constraints but political (often electoral) ones. This provides states with a degree of autonomy. Thus they have introduced child labour laws, environmental regulation, state control of certain industries and so on. But their autonomy is not absolute: when states threaten capitalist interests too much, capitalists have significant means to resist them, and often put these means to use.

Moreover, such weakly anticapitalist interventions tend to come because of pressure on politicians from countervailing forces, in whose interests those interventions are – social movements, trade unions, coalitions of voters. Their influence, particularly in democracies, can sometimes outweigh that of capitalists. But future people do not have such capacity for influence or pressure. Future people cannot join demonstrations, or strike, or vote. So whilst we can expect capitalists to wield influence over politicians, we cannot expect future people to do so. In a struggle between their interests, then, future people will tend to lose out.^[13]

A possible rejoinder here is that capitalists need not be selfish – in fact, longtermists might add, some capitalists currently supply a large amount of philanthropic funding to effective altruist and longtermist causes. Therefore, it might be premature to assume that they would oppose regulatory interventions that are in the interests of future people; moreover, perhaps such interventions are less crucial, since capitalist decisionmakers might make ethical choices to reduce negative externalities, increase positive ones, and discount less without incentive or state intervention.

There are at least two good reasons to be sceptical about this appeal to ethical capitalists. The first is that such people do not tend to be as reliable in their altruism as they claim. The two most prominent capitalists to embrace longtermism are Sam Bankman-Fried (recently convicted of fraud) and Elon Musk (currently destroying my favourite social media platform). It is difficult to believe that these people have successfully maximised long-term welfare. There are structural reasons, relating to capitalism, for their failure. For one thing, capitalists in capitalism are a minority, alienated from the majority of people who are their instruments, as workers or consumers. This insulates them from criticism and the benefits of diverse input into their decisions, making them more prone to error. For another, in both of these cases, supposedly ethical capitalists have put their business interests ahead of ethical concerns. Bankman-Fried funded political candidates running against democratic socialists (Heer 2022). Musk publicly supported a coup against a Bolivian government that might have increased his supply costs (Bejarano and Prashad 2020) and allegedly held the Ukrainian army to ransom during their current war with Russia (Farrow 2023).

The second reason to doubt the appeal to ethical capitalists is that even if there were capitalists with purely altruistic, even longtermist, motivations, capitalism structurally limits the possibilities for altruism to be practised. Return to our fictionalised version of Alphabet, and their technology that imposes such existential risk that it would be better for long-term welfare if they did not release it. Perhaps those running Alphabet are longtermist altruists, and therefore are minded not to release it. But of course, as a capitalist firm, they need to make a profit. Shareholders are demanding a return on their investment in this new technology. Unwilling to deploy the dangerous technology, Alphabet fails to profit and finds it hard to raise more capital – which floods towards less ethical firms who are willing to produce negative externalities to make profits. The CEO is fired. A new CEO realises that if Alphabet doesn’t deploy, the firm’s slide will continue, and some less scrupulous competitor will replace them, and probably deploy a similarly dangerous technology anyway. By continuing to hold back the technology, she risks being outcompeted and fired for nothing. Therefore – even though she is a moral person who cares a lot about long-term welfare – she decides to deploy, with potentially catastrophic consequences. This is how competitive markets suppress the exercise of altruism. We can observe this dynamic in the current AI sector, where many developers have grave fears about what they are developing, but are forced by competitive pressure to continue development.^[14]

To sum up this section, certain interventions in capitalism could help to correct the inefficiencies I identified in Section 2. These interventions reduce the extent to which the economy is capitalistic, and therefore support anticapitalism in the weak sense. But we can expect such interventions to be resisted by capitalists, and so as long as capitalism is predominant, they will be difficult to impose and maintain. Supposedly ethical capitalists are unlikely to resolve this problem. Therefore, considerations of long-term welfare give us some reason to support anticapitalism in the stronger sense: ending capitalism’s predominance in our economic structure.

4 Could Another System Do Better?

In Section 2 I pointed out systematic features of capitalism that will make it inefficient with respect to very long-term welfare. In Section 3 I argued that correcting for those features will involve reducing the extent to which the economy is capitalistic, and that capitalism’s predominance will frustrate efforts to make such corrections. But it may be countered that all economic systems are in the same boat in this respect. Perhaps, as Winston Churchill is supposed to have said of democracy,^[15] capitalism is the worst system, apart from all of the others. If so, there would be no reason to think that replacing capitalism with an alternative economic structure would be better for the far future.

It is beyond the scope of this paper to design an alternative economic structure. But in this section I will give some reasons to think that a noncapitalist structure would be better at providing the public goods that are crucial to long-term welfare, invest at a lower discount rate, and give greater influence to longtermist altruistic motivations than would a capitalist one. Putting this together with the arguments above, we have a case that considerations of long-term welfare support the replacement of capitalism with an alternative economic structure in which capitalist decision-making is not predominant.

The noncapitalist provision of public goods is already a significant feature of our predominantly capitalistic economy. With respect to education, for instance, it is the leading solution to the externalities problem. The benefits of an education are not only enjoyed by the person who receives it, but also by their employers, colleagues, family, fellow citizens and so on. A well-educated population is good for everyone. Education therefore has positive externalities, and will be underproduced by capitalists subject to the market. The most common solution is for states to provide education themselves up to a point. Other noncapitalist actors are active in education, such as religious organisations and families. The world’s largest encyclopedia, Wikipedia, is produced by a global network of volunteers and offered free to all.

Education, and other examples such as roads, stand as a proof of concept: noncapitalist provision of public goods can be successful. Given that many items on the longtermist wishlist are public goods, longtermists should want to expand their provision, and at least in some cases noncapitalist structures are an effective way to do so. Often, longtermists advocate for particular state interventions to provide public goods: asteroid detection by NASA, for instance (Greaves and MacAskill 2021, 11–12). But there is reason for longtermists to want to expand the capacity of society to provide public goods in general. As Schmidt and Juijn (2024) put it, recommending social structures under which public goods provision is better and collective action more possible:

Targeted actions like washing your hands regularly or getting a flu shot can reduce your risk of dying from an infection. But you will also do well investing in a strong immune system, as that is an ‘all-purpose good’ in lowering your risk of dying from any bacterium or virus. Investing in good institutions might similarly be an all-purpose good: rather than tackling individual sources of existential risk directly, we improve conditions for tackling whatever existential risks may come our way. (2024, 80)

According to this logic, whilst supporting the provision of particular longtermist public goods under capitalism would be good for long-term welfare, altering the economic system to one which tended to provide superior public goods in general would also be good, perhaps better. Since even under capitalism many public goods are best provided through noncapitalist means, and since noncapitalists are generally more amenable to funding public goods than are capitalists, it seems likely that a noncapitalist alternative structure would be generally better at public goods provision. This is not a necessary inference, and different structures will differ in their results. But it is a prima facie case that there is some noncapitalist structure that will outperform capitalism with respect to the externalities problem identified in Section 2.1.

A noncapitalist structure, then, could probably do better than capitalism with respect to longtermist public goods.^[16] Could it also do better with respect to overdiscounting? Again, we have reason to think so from both theory and from the current economic structures. I have already made a high-level theoretical case in Section 2.2. Capitalists tend to discount at the real interest rate prevalent in the market. The optimal social planner, discounting only for catastrophic risk, will tend to employ a lower discount rate. Therefore, we should expect that subordinating economic decisions to society (rather than, as under capitalism, to capitalists) will lead to greater investment in long-term welfare. We might support this theory, as with the case of public goods provision, by observation of noncapitalist parts of capitalist economies: states often provide the ‘patient capital’ invested in research and development projects that generate large but not immediate returns (and so are unattractive to private investors) (Mazzucato 2013, 2015).

This observation notwithstanding, there is a rejoinder available to my theoretical argument. Anticapitalist alternatives would not necessarily subordinate the economy to ‘society’ modelled as a rational agent, or (equivalently) to an optimal social planner. History has shown that those who claim to represent society’s best interests should be treated with suspicion. The best way we have yet found to treat such people is to hold them accountable to the population through democratic means. However, as mentioned in Section 3, future people cannot engage in standard democratic processes: they cannot join demonstrations, strike or vote. Politicians, then, accountable to presently existing voters, may be expected to neglect future people’s interests, just as would managers of private enterprises, who are accountable to capitalists with high discount rates.

One thing to note here is that there may be reforms to a democratic system that would incentivise longer-term decision-making. But there may be similar reforms available to reduce short-termism in business, as discussed above. The best case, I believe, for elected politicians being more sensitive to long-term interests than capitalists or their managers (or at least, having greater potential to be sensitive to long-term interests) has to do with the way in which capitalism tends to suppress altruism.

Caring about far future generations is necessarily caring altruistically. Voters and workers may be as myopic and selfish as capitalists, and tend to discount future people’s interests in similar ways. However, at least some are, to at least some degree, altruistic and motivated by future people’s welfare. Some capitalists are too, of course. But under capitalism, even an altruistic capitalist must put profit ahead of doing good, on pain of being outcompeted in the market – Marx’s ‘coercive laws of competition’ ([1867] 1990, 433). In systems of less ruthless competition (or, in Mill’s phrase, of ‘contest, who can do most for the common good’, see Mill [1848] 1965, Section 2.1.3) altruistic motives could influence decision-making far more. In democratic elections, at least some voters vote at least somewhat altruistically, whereas market actors are forced to seek profit (in some capitalist ideologies this is a moral duty, see Friedman 1970). This makes more democratic economic structures likely to outperform capitalism with respect to future welfare, other things equal. This effect will be especially significant if one believes – as many effective altruists do – in the likelihood of moral progress, conceived of as an ‘expanding moral circle’ (Singer 2011), so that people will become more altruistic – and more likely to care about far future beings – over time. In that case, the capitalist system, which suppresses altruism, would become a fetter on producing welfare relative to a system which gave greater scope for altruistic motives to affect economic decisions.

In this section we have seen that not only does capitalism inefficiently promote long-term welfare and potentially frustrate attempts to correct such inefficiencies, but also that alternative structures involving more public goods provision, more democratic control and more room for altruism can reasonably be expected to outperform capitalism with respect to the far future. The claims made in this section are necessarily more speculative than those of previous sections, and hence I am less confident in the stronger anticapitalist claim that considerations of long-term welfare support the replacement of capitalism with an alternative economic structure.

Of course, there are better and worse ways of filling in the details of such structures, and some will face other problems of their own. But the seeds of such structures are already present in our present hybrid capitalism, and over the long term we have a chance of developing them into a new, more optimal system. Moreover, those chances will increase if more people embrace anticapitalism, and thus more brains attend to the problem.

5 Conclusions

I have given reasons to think that capitalism will be systematically inefficient with respect to long-term welfare. These inefficiencies are not likely to be compensated for by capitalism’s relationship with economic growth, and so require correction through intervention by noncapitalist actors such as states. This means that considerations of long-term welfare support anticapitalism in a weak sense: that of reducing the extent to which the economy is capitalistic.

Because these corrections will be difficult to impose and maintain whilst capitalists retain a dominant economic position, and because we can imagine, drawing from current noncapitalist parts of capitalist economies, an alternative system doing better with respect to long-term welfare, this paper offers some support for a stronger form of anticapitalism: replacing capitalism with some other form as the dominant part of our economic structure. Support for this stronger claim, though, relies on a greater deal of speculation, and thus is less robust.

Should longtermists, then, be anticapitalists? I have given only partial arguments based on some systematic aspects of capitalism. There may be countervailing considerations. Perhaps noncapitalist structures have other flaws which make them bad bets from a longtermist perspective. Perhaps capitalism’s dominance is so great that replacing it is an impossible task – or at least, is so difficult that longtermists would do better to focus on more tractable, more moderate policy changes. But I hope to have convinced those worried about the far future that our choice of economic structures is important, and that anticapitalism is at least worth serious consideration.