Holden Karnofsky recently discussed the problem of vetocracy. This post suggests why that problem may be more tractable, without the need for partisan politics, than is widely realized. Vetocracy is a subset of broader coordination problems, which I hope to discuss in future posts.
Epistemic status: the comments about vetocracy are based on existing work in policy analysis, institutional economics and other fields. I work nearly full time on the NIMBY problem and have spent years reading around this topic to try to address that problem. I am reasonably confident based on examples in different countries that the mechanisms suggested here will have some effect in at least a limited set of cases. The broader conclusions are far more speculative.
- There are existing but not widely known techniques from institutional economics, policy analysis and policy studies that have been effective in overcoming veto obstacles in the past.
- It is possible to implement such techniques in relatively non-partisan ways. That may make it politically easier to get them implemented.
- Examples include opt-outs and opt-ins, policy bundling, alienability, exemptions (phased or otherwise), transition periods and compensation. There are also other forms of policy design that could help.
- Large deadweight losses (e.g. from existing zoning rules) necessarily imply large potential benefits from reforms. Reformers can use these benefits to build a winning coalition for change.
- Vetocracy may become even more damaging to welfare unless we do something about it. Mitigating the problem could potentially have large benefits for welfare.
- Vetocracy is a subset of broader coordination problems in governance. Fixing those may have even larger benefits.
Vetocracy: the problem
Holden is concerned that ‘the modern world sees a lot of instances of stakeholder strength growing to the point where making any kind of change to the status quo becomes prohibitively difficult.’ He points to problems of ‘vetocracy’, such as two San Francisco residents who have been solely responsible for blocking large numbers of beneficial city projects. I agree with this, and have written about how this has caused housing shortages, which have contributed to many other problems including pollution, inequality and obesity.
Holden states that there is ‘probably plenty of room to significantly mitigate these challenges via well-designed processes for considering - but not being totally beholden to - stakeholder input.’ I agree.
There is extensive literature in, for example, the fields of institutional economics, policy analysis and public policy on practical mechanisms to achieve policy change without necessarily being involved in partisan struggles. One work on ‘transitional gains traps’ reviews various other mechanisms to overcome political blockage for change: opt-in (or opt-out) mechanisms, phase-ins, phase-outs, exemptions, compensation, or making rights tradeable. There are also other means of policy adaptation.
Vetocracy arises where there are figures who have veto power through some kind of overriding control. In most building projects, except in rare cases where some particular owner of an essential parcel of land is holding out, it is the state legal rules controlling construction that give veto power to such stakeholders. If an overriding, top down constraint can be bypassed, those veto players will have less power and the system will have fewer single points of failure.
The fact that stakeholders often do not negotiate directly but have their views considered by higher officials is another reason why we have such a vetocracy today. The phenomenon of ‘blame avoidance’ by officials or politicians means that they generally prefer not to do controversial things. It is better in their eyes to quietly sacrifice another small piece of growth rather than lose more voters. That mechanism gives disproportionate veto power to small but highly vocal groups. Direct democracy can lead to much more radical outcomes than governing politicians might have chosen on their own.
Jean-Claude Juncker, former President of the European Commission, once said: ‘We all know what to do, but we don’t know how to get re-elected once we have done it.’ He may have been too confident about the first part, but the second is a constraint on politicians in many cases.
Currently vetocracy often operates through an authority considering stakeholder input in a discretionary fashion. The YIMBY movement in England has been trying to move past this by giving stakeholders the power to bypass veto players by consenting through direct democracy, even if the support is not unanimous. That mechanism is currently often missing in the ‘vetocracy’ that Holden identified. I believe that in a subset of cases it can be powerful.
Much of our legal system for granting permits, particularly in relation to what can get built, is profoundly top-down in nature, and should be reformed in order to benefit from decentralized systems of granting consent that can compete with the existing system.
In other fields such as the Internet and use of the blockchain, we have seen the potential power of decentralized mechanisms. I suggest that there is considerable potential to use them more to solve various current problems of policy and governance.
Why not just get rid of vetoes?
Sometimes you can't get rid of vetoes. And even if you can, it's not always a good idea to do so.
The instinctive answer to the problem of vetocracy is to say that officials and politicians should just ignore such veto players. But in some sectors and places it may be politically difficult or impossible to pass laws to strip away veto powers.
And there may be good economic reasons why it is hard to strip away such protection against spillover effects (which economists call ‘externalities’), as discussed below. Given that there was sufficient political force to demand a veto in the first place, groups are understandably unwilling to give them up, and can pressure politicians accordingly.
Blocking mechanisms such as judicial review and the California Environmental Quality Act have tended to multiply and become stronger over time, partly because there has been political demand for such blocking mechanisms. If more people had more confidence that things would not go ahead that affected them adversely, the political demand for such mechanisms would be reduced.
Stakeholder views can be useful
There are five main reasons why stakeholder views can be important:
- Negative externalities can be important.
- It is good to force epistemic humility on top-down policy makers.
- Reforms that are closer to Pareto optimality can be politically easier to pass.
- Direct democracy can provide an alternative means of getting approval, in competition with existing processes.
- Stakeholders can help to overcome information problems.
Negative externalities can be important
Imagine a small town, owned by a single landowner, with all the residential and commercial properties rented out. The landowner is approached by a company wishing to rent a vacant factory in the center of town. Unfortunately the company’s industrial process would unavoidably pollute the local air, although the pollution would decay to being harmless before reaching any other towns. She calculates that the rent that the company will pay will not offset the reduction in other rents that she will get because the whole town will be a worse place to live. So she decides not to go ahead.
That is the economically efficient solution: GDP per capita is higher if she does not go ahead. Economists would say that she has ‘internalized’ the externality of pollution. But the same reasoning applies if the ownership of the town is fragmented into a thousand homeowners and businesses. A legal system which, in those circumstances, repeatedly allows such polluting projects that reduce overall value, over the objections of the residents, is likely to produce lower growth and overall wealth per head.
So in some circumstances protection against spillover effects – what economists call ‘externalities’ – is efficient, by preventing value-destroying projects. Considering in which circumstances it is efficient is a hard and complex question, beyond the scope of this piece. (The work of economist Ronald Coase may be a starting point for those interested.) But we cannot simply assume that protection against externalities should be overridden in all cases. We might be in the same position as the town with the potential polluting factory.
It is good to force epistemic humility on top-down policy makers
There are cases where forcing something through in the teeth of wide opposition has proven, decades later, to be the wrong decision. Razing fine-grained residential areas to build broad highways through city centers in the 1960s is now more often seen as a mistake. Perhaps some of the authorities who devastated those areas, crushing broad local resistance, should have had a little more epistemic humility. More powerful veto players in some other cities blocked such devastation. Similarly, direct democracy would generally never have allowed it.
Also, some forms of infrastructure blocked by veto players might not actually be the best investment in any case. Building another road must be paid for, generally by the government; on the other hand, urban densification allowing people to use more active means of travel like walking instead of using cars, or introducing road pricing on the existing roads, can both generate net revenue. All three of these options face political constraints. If we could solve the political constraints, it would be much cheaper and better for the environment to densify sprawling suburbs or stop the current practice of subsidizing cars by letting them use roads for free, than to use more concrete and steel and valuable land for more roads or bridges to try to reduce traffic congestion.
Staying closer to Pareto optimality can improve the chances of successful reform
Economists call an exchange Pareto superior if all parties to the exchange consider themselves better off. That may require money or some other benefit to change hands. Market transactions typically result in Pareto superior exchanges, though those who are not party to the exchange may find themselves worse off.
In the real world, economists often satisfy themselves with ‘Hicks-Kaldor’ superiority, which is where a step could make everyone consider themselves better off, if we redistributed some of the benefits—even if in reality they are not redistributed. In practice, with large and complex effects, even Hicks-Kaldor superiority may often be impossible: some people oppose all change, no matter what the benefits it offers. In YIMBY parlance they are sometimes known as ‘BANANAs’ - ‘Build Absolutely Nothing Anywhere Near Anyone’. If they are powerful politically, we might try to mitigate their opposition by at least ensuring that they are not made financially worse off by a proposal.
But the further we move away from the principle of Pareto superiority, the larger the political obstacles to reform.
Direct democracy can provide competition for top-down means of getting permits
Many pieces of infrastructure are good for growth. And many potential pieces of infrastructure would generate immense revenues that could pay to compensate everyone nearby for the adverse effects on them. In some cases, those negotiations happen and the infrastructure is built. But in many cases, the infrastructure is blocked by a political or court process, even where the benefits would make a large majority of locals happier and better off.
In many such cases, I suggest that if approval is granted via direct democracy (perhaps with a high qualifying majority), that can provide us with good grounds to believe that overriding veto players will boost growth. Note that I am not suggesting a direct democratic power to veto, but merely merely a direct democratic power to prevent things from being vetoed.
There are strong economic reasons to think that this would be a powerful mechanism in some circumstances. The crucial insight is that where a particular proposal is sufficiently profitable, there may be enough profits to spread round to make that proposal nearly win-win, and hence attractive to most voters. If the number of stakeholders directly affected by a proposal is small, and the economic benefits of the proposal are large, it can in some cases be possible to share enough of those benefits with stakeholders that they become enthusiastic supporters. Such development is very much not a zero-sum game. Conversely, unprofitable projects will be unable to ‘share the gains’ in this way and hence should not be as attractive to voters. Direct democratic approvals, particularly requiring high supermajorities, can also be more politically sustainable because by definition they happen when a broad majority of people support the proposal, which will reduce political demand for additional legal blocking mechanisms being introduced into the system. That is in marked distinction to many current permitting processes.
If the number of stakeholders directly affected by a proposal is small, and the economic benefits of the proposal are large, it can in some cases be possible to share enough of those benefits with stakeholders that they become enthusiastic supporters. The TAMA 38 regime in Israel allowed apartment owners to vote to redevelop and share a large proportion of the benefits. It accounted for approximately 35% of the new homes built in Tel Aviv in 2020. Another direct voting regime in Seoul accounted for a large fraction of the condos built in the mid-1990s. And in Vancouver, the small Squamish Nation voted by an 87% majority for the development on their own land of a new district with 11 towers and 6,000 homes. Incentives matter. No doubt some other people were unhappy, but they had no legal power to stop it. There was no ‘consideration’ of their concerns and no veto player.
The problem of information
Another reason why such direct democracy can be a powerful mechanism relates to the economic calculation problem. Proposals for infrastructure and other projects often occupy possibility spaces with a very high number of dimensions. And the people who may be affected often have widely different preferences, which are not known in detail by the central decision maker. So even with perfect information about the state of the world – which would be impossibly expensive to get – the decision maker could not make the welfare-maximising decision, because it would not know the exact preferences of the people involved.
But if there is a direct power for smaller groups of the most affected people to bypass local veto players, they can help to design proposals that they will strongly favour. Such decentralized systems can make it easier to consider and selecting incremental trade-offs rather than making all-or-nothing decisions. That in turn provides a strong core of political supporters who can help to protect the proposal. Of course, there will often be challenges in deciding which sets of people get to approve things by direct vote: how to draw boundaries, and what to do about compensation and mitigation for those beyond those boundaries.
Often, a given proposal is not welfare enhancing, but with sufficient modification it could be made so. That is one reason why decision makers listen to stakeholders and press for proposals to be changed. But again, blame avoidance will push decision makers to less controversial decisions. Direct negotiation and direct democracy may enable wildly more ambitious steps in the decision space that make everyone financially better off, and a large majority happy for the project to proceed.
So empowerment - in the sense of protection against externalities – need not lead to vetocracy.
Vetocracy seems likely to get worse without action
I expect the vetocracy problem to become worse unless we do something about it. Holden notes that ‘[a]s people become wealthier, more educated, and more informed, they become louder and more opinionated, active stakeholders.’ The prolific economist Harold Demsetz wrote a seminal paper on this point (available without paywall here). As one author paraphrased it, ‘Demsetz hypothesized that property rights emerge when some change in the relative value of resources occurs that makes it cost-effective to internalize costs that previously were experienced as externalities’.
As technology advances and the world becomes wealthier, the externalities that one person can inflict on many others become correspondingly more important to those others. Hunter-gatherers could not drop nuclear weapons on the other side of the planet; medieval London was too poor to address or even understand its immense hygiene problems.
Overall, we should expect that technology driven growth will mean that political demand will increase over time for legal protection against such externalities. Indeed, we see that has happened in many fields, from pollution regulations to control of nuclear materials. The only question is whether we can provide such protection with legal rules that are sufficiently flexible to allow broadly win-win ways forward to be found and negotiated in infinitely dimensioned decision spaces. And thus far the pace of policy and governance innovation seems to be far slower than the pace of technological innovation. This will become even more of a problem if the pace of technological change accelerates.
I think it is important for the effective altruism community to learn more about means to address such problems of vetocracy, and other governance problems. Not only will that help us to boost growth today, but it may also help us to address risks in the future.
The broader context: coordination problems
Vetocracy and other examples of bad governance are generally coordination problems. Almost no-one wants the whole world to be poorer if it means they are poorer too. Similarly, almost no-one wants the human race to be wiped out by some avoidable catastrophe. If we are, it will be because we have failed to coordinate ourselves well enough.
The problem is not just that we do not have institutions that are good at understanding, foreseeing and preventing long term problems. The problem is much bigger: we have very few institutions that are good at understanding, foreseeing and preventing any kind of problem.
For example, we as a species failed to coordinate ourselves enough to prepare well for a pandemic; we failed to disseminate accurate information on the efficacy or inefficacy of various measures such as masks and handwashing; we failed to pay in advance for the development of broader-spectrum boosters; and we have failed to regulate gain-of-function research at a global level. A failure to regulate can have many causes, including veto players or simple failure of a range of players to act.
This whole discussion is part of the broader question of how to improve governance and policy. Founders Pledge recently wrote a report on longtermist institutional reform. 80,000 Hours currently lists Improving Institutional Decision Making as one of its second-highest priority areas. I understood IIDM to be generally taken to exclude policy itself, although the 80,000 Hours report does mention changing institutional incentives, which is an aspect of policy. And it may be possible to develop or enhance meta-techniques for improving institutional incentives that can be applied across a range of situations or sectors. It may be easier to fix decision making if we fix the incentives that lead to crony beliefs.
We can improve policy without partisan politics
In contrast to longer-term objects of current EA focus, there are large economic sectors that currently suffer from substantial economic deadweight losses. Those seem promising avenues for working out how to improve governance by focusing on theories of change that avoid partisanship.
Many who do not specialize in policy, including some in the effective altruism community, gain most of their knowledge of ‘policy’ from reporting of the partisan struggles that dominate the media. Those perceptions are skewed by selection bias. Matt Yglesias coined the term ‘Secret Congress’ to describe the progress on policy that can be made in a non-partisan way, out of the public eye. Caleb Watney and Alec Stapp of the Institute for Progress argue cogently that that is often the easiest way to improve policy.
The mechanism of direct democracy suggested above can be an example of the ‘pulling the rope sideways’ that Hanson and Yglesias describe, i.e. proposing a reform that is orthogonal to partisan divides. It can be highly successful, as in the examples discussed here.
The quality of policy can often be worse than it should be, not least because of the rent-seeking and other reasons noted by Mancur Olson. But the beauty of focusing attempts at nonpartisan policy reform on areas of huge existing inefficiency or deadweight loss is that fixing those deadweight losses provides huge benefits that can be shared out to form a winning coalition for change. Focusing on such areas of large deadweight loss therefore stacks the odds in favour of those who design reforms to fix them.
The larger the inefficiencies resulting from current bad governance, the more potential benefits of reform that can be shared around to build a winning coalition for change. The key insight is realizing that policy space and policy preferences are not one dimensional. In a multi-dimensional policy space, the median voter theorem does not apply: you can assemble a winning coalition just by clever bundling of policies. The simplistic notion of ‘political will’ is a far from complete model in thinking about how to engineer policy change in an iterated game of near infinitely-dimensional policy choices and preferences.
In principle, such areas of deadweight loss might be fixed by profit-seeking entrepreneurs. But regulatory change is often painfully hard and slow. It may be impossible for entrepreneurs to capture a large enough share of the enormous societal benefits to provide sufficient incentive for the immense and risky investment of the required time and resources.
Improving policy and governance has large potential impact
With high uncertainty, I think that focusing a small amount of resources on improving broader coordination techniques for reducing such large deadweight losses in various areas could be a highly impactful, tractable and neglected area of research. Many of the techniques mentioned above, particularly policy bundling, can be very helpful in addressing broader problems than vetocracy. As a species, we are better at planning and coordination among ourselves than most other species, but our natural biological techniques for such coordination evolved in a world where each individual interacted with fewer others, and we are still learning the optimal social techniques.
I believe that we could substantially raise global GDP per capita, without further damage to the environment and without increasing inequality, by developing better coordination techniques to fix some of the areas of grossly inefficient regulation in sectors including housing and other construction, transport, energy, finance, education, healthcare, and law – without partisan politics. That increase would potentially have profound benefits for human welfare and increase the funds available for direct giving and other high impact interventions.
If it is indeed possible at low cost to improve policy substantially in those areas, that would be a high impact use of funds. Success in such areas might also improve recruitment to effective altruism by delivering visible improvements that are salient to a wider range of people. In addition, learning coordination techniques to achieve such improvements may prove useful in helping to address various catastrophic risks.
Many thanks in particular for their helpful comments to George McGowan, Larks, John Kroencke, Alasdair Slacks, Maxim Harper, Jonathan Kitson and Lawrence Newport. All errors are my own.
Mushtaq Khan also commented on political constraints, in the same interview with Rob Wiblin: ‘If there were no political constraints, it’s always beneficial for the powerful to drive growth, right? [...] They get more power. So, if they’re not doing it, it’s always useful to ask, “Why are they not doing this?” And there are only two possible answers. One is that they’re really abominably stupid, right? Or the other is there is some constraint politically, or in terms of organizational capabilities, or something that you need to identify. And I think there are very, very few leaders in the world — and certainly not Nehru or Ayub Khan, or the others who were leading these Asian countries in the 1960s — who were that stupid. They had political constraints which prevented them from taking on these people who were capturing rents.’
It may be that most economists underemphasize such decentralized, horizontal arrangements. In a different context, development economist Mushtaq Khan said: ‘if you look at the history of how development happened in advanced countries and how a rule of law emerged, it was always through these incremental processes where people at different levels of society in their own interests started to follow rules and impose rules on their peers. And that horizontal process is for some strange reason not adequately examined by economists and institutional economists. We are often much more concerned about these vertical enforcement strategies, which then typically fail.’ And see also the comment from Rob Wiblin below.
If the factory does not go ahead, the national government can make everyone better off than if it had, by paying money to the people who would have gained from the factory, and raising the money by imposing non-distortive taxes such as taxes on land value. The benefits of the factory not being occupied are great enough that the national government could make everyone better off than they would be in the world where the factory had been built, by compensating those who would benefit from the factory with funds raised by non-distortive taxes. Of course, the national government often does not do that redistribution; but that is a separate question. If we can redistribute some of the wealth, we will get higher overall welfare if the factory does not happen. In an iterated game, an economic set of rules which consistently leads to outcomes with lower overall wealth will reduce welfare in the long run.
Or at least, far below the marginal cost of such use, when costs such as congestion are taken into account.
Law professor Michael Heller defined the concept of an anticommons, where overlapping veto rights make it impossible for anything to get done. That is effectively what we have in many sectors today. It is the opposite of the tragedy of the commons, where no-one has the power to stop anyone overusing the resource. Nobel laureate Elinor Ostrom showed that in fact communities can often negotiate in a repeated game to avoid the tragedy of the commons. Similarly, an anticommons can be addressed by allowing direct democracy to override veto players.
Another way to express this point is with another key element in the economics of property rights – the question of alienability, or whether the right can be transferred, in particular for some kind of benefit.
And one major reason for our current vetocracy is not that we are giving legal protection against externalities, but that we are making these rights inalienable, even by a sufficient majority of people. It is perfectly possible to give people legal protection against something but also give them the right to agree to give up that protection in return for some other benefit. That happens with local and state governments, and within companies and partnerships.
On this topic see also William Riker's work on ‘heresthetic’. I aim to expand on this in a future post.
‘I’m very excited about this framework, and getting myself and listeners to use it a little bit more. I studied economics in undergrad, and we spent lots of time just thinking about what policies — in theory, or maybe even as a result of studies — would be optimal in principle, if you could get them enforced. But we spent almost no time on this institutional economics framework, where you think about what is actually going to happen in practice, given the interests and the capabilities of the organizations that exist in society.
And I think that’s one reason why the policy advice that economists and people like me often offer comes across as extremely naive. It’s good in theory, but it won’t work in practice. It either gets rejected, or it’s implemented and then peters out and doesn’t achieve its desired goals, because we don’t use this framework of… And I guess it’s understandable, because it’s very hard to do. It’s a lot harder than just doing like economics 101 theory to figure out what policies to have. But we don’t often think about how the political process is going to play out, and how the implementation will actually play out.
So this institutional economics framework is useful with anti-corruption, but it also potentially helps a lot with thinking about development policy as a whole, and with thinking even about what policies are actually likely to be implemented properly, or even get up out of the political process in rich countries as well. This process is going on in the U.K. and the U.S. and Australia and other places. And very often I think it can explain why particular policies get undermined, ultimately, and don’t achieve their goals, or why they’re just not going to get through the political process. Because they’re too contrary to the interests of the people who are most concerned.’
Rob Wiblin: ‘I’d like to get to a world where people have the traditional economy lens, but also a political economy lens. Then these things get integrated, and people go back and forth looking at the different angles on these problems. It seems like there’s a big barrier to getting the political economy lens applied systematically and regularly to all of these questions, because it’s hard to simplify as a set of questions that you ask one after another, or tools that you apply, like supply and demand curves, to all of these questions. I wonder whether there could be a lot of value in trying to boil it down to an actual toolkit that a policymaker… What if I’m a bureaucrat, and I’m trying to suggest something. I used to work in the Australian government, and we would suggest things about the electricity industry. But I wouldn’t know where to begin if I was trying to provide a political economy/institutional economics lens on things, because there’s so many different things you could do that it feels somewhat arbitrary which threads you decide to pull on.’