Translation doesn't need grammar; policy solutions don't need root causes.

An excerpt from a letter published in the Financial Times:

Translation is one of the many positive revolutions brought about by artificial intelligence, but, surprisingly it has gone largely unnoticed. While machine translation has made constant but slow progress since the first Georgetown-IBM experiment in 1954 the shift by Google by 2016 from structural rule-based systems relying on grammatical structures to statistical models and neural machine translation was a major breakthrough. Automated translation now takes into account the context to solve the ambiguity of some words or phrases, and to grasp the tone as well as the style in order to find the right vocabulary and the most suited idiomatic forms. Being trained on vast datasets, it is also very unlikely that it will get stuck on a rare or technical expression that a human translator will in turn struggle to find in a paperback dictionary. How AI is overcoming Europe's language barriers, Philippe Huberdeau, 'Financial Times', 30 April 2025

To me, this has application in the world of policy. Our current policymaking system relies on being able to trace the causes of our social problems, to do so continously, and then to try to deal with them. This identification-based model works well when cause and effect can be reliably identified, and don't change much over time, nor differ much over space.

In today's complex societies, it isn't good enough. There are too many variables; and feedback loops mean that implementation of policies changes how people and systems respond. The current model cannot address such complexities, so we get fossilised policies that also tend to be one-size-fits-all. 

I don't propose that we try to use AI techniques to solve policymaking goals such as reducing crime, improving health or, at the global level, eliminating war. Trying to condense all of human history into the context according to which effective policies could be made is impossible. Reducing the human experience into terms that machines can understand can't be done. Partly because it's too great, partly because it's too subjective and partly because it keeps expanding. 

But what we can learn from the application of AI to language translation is that we don't need to fully understand everything about how a language works - its grammar - to translate it effectively. And this is where the Social Policy Bond concept, with its focus on outcomes, enters the picture. Investors in the bonds don't necessarily need to identify and deal with any supposed 'root causes' of crime, war or any of our other social or environmental pathologies in order to address them effectively. Trying to do so would be a Herculean task and so can readily be invoked as an excuse to do nothing. What we can do, and what a Social Policy Bond regime would do, is give incentives to investors to look for the best ways of solving our problems on a continuous bases. That might, in some cases, mean looking for root causes. But it might simply mean researching, trialling and refining many diverse potential solutions and adopting those that are most promising - constantly. Achieving our social goals requires diverse, adaptive approaches. The current policymaking system delivers neither. 

For more on this theme see here and here.

Why we should target environmental outcomes

'Don't overlook the many benefits of plastics', writes the Economist:

Plastic packaging prevents perishable foodstuffs from spoiling, making possible global trade in meat, fish, fruit and vegetables. It enables essentials like rice, cooking oil and powdered milk to be stored and distributed safely and cheaply. A one-litre plastic bottle weighs 5% as much as a glass one; plastic packaging thus reduces shipping costs and emissions. ... When properly managed and well monitored, [landfill] is far less environmentally ruinous than often portrayed, and can be simpler and more effective than poorly executed recycling. Don't overlook the many benefits of plastics, the 'Economist', dated 19 April 2025

The sort of life-cycle analysis required to establish the environmental benefits or otherwise of shifts in our behaviour are bedevilled by boundary issues, measurement difficulties and the difficulty of weighting one type of environmental impact against another. They are better than blandly assuming that rail is ‘better’ than air travel, that solar power is better than coal-fired power stations, but for the making of robust policy they would need to be continually reassessed in the light of our ever-expanding knowledge of the environment and our ever-changing environmental priorities. Government policy cannot be so responsive: if government did use life-cycle analysis with the aim of altering our behaviour, it would probably do so on the basis of a one-time, necessarily limited, and (probably) subjective assessment of environmental costs and benefits. It’s not good enough, but even worse would be what we largely have now: government environmental policy based on corporate interests, media stories and the launching of visually appealing initiatives that look good but otherwise achieve nothing.

Social Policy Bonds would take a different approach. They would subordinate environmental policy to society's  desired environmental outcomes. Say we wish to reduce our use of plastics. A Social Policy Bond issue that rewarded achievement of such a reduction would generate incentives for bondholders to bring it about at least cost. They might well carry out life-cycle analyses in their attempt to do so. But there is an important difference between the way do they would conduct their research and the way government would do so: bondholders have incentives to achieve their goal efficiently. This is likely to mean responding to - and stimulating - increased knowledge of scientific relationships and technical advances.

A single environmental goal, such as reduction in use of plastics will necessarily require diverse, adaptive responses. These are precisely the sort of responses that government does very badly. Government can and should articulate society’s environmental goals, and can help pay for their achievement: in the democratic countries it performs these functions quite well. But most of our environmental goals require complete and responsive understanding of complex relationships, and actually achieving such goals requires continuous, well-informed and impartial decisions to be made about the allocation of scarce resources. For that purpose, Social Policy Bonds, with their incentives to achieve targeted outcomes efficiently would, I believe, be far better than the current ways in which environmental policy is formulated.

For more about Social Policy Bonds please see here. For applying the Social Policy Bond idea to environmental problems, see here. 

Climate change: I don't need to know

I don't know whether the climate is changing. Most of the essays, books, articles, I read do make me believe that anthropogenic greenhouse gas emissions are causing the climate to break down. But then I see papers like this, and I realise I just don't know:

The anthropogenic CO₂-Global Warming hypothesis, as articulated by the Intergovernmental Panel on Climate Change (IPCC) and supported by researchers such as Mann, Schmidt, and Hausfather, lacks robust empirical support when subjected to rigorous scrutiny. This analysis integrates unadjusted observational data and recent peer-reviewed studies to demonstrate that the assertion of human CO₂ emissions as the primary driver of climate variability since 1750 is not substantiated. Instead, natural processes—including temperature feedbacks, solar variability, and oceanic dynamics—provide a more consistent explanation for observed trends. A Critical Reassessment of the Anthropogenic CO₂-Global Warming Hypothesis, Grok 3 beta1, Jonathan Cohler et al, 'Science of Climate Change, vol 5.1 (2025)

Judging by the absence of any meaningful attempts to reduce greenhouse gas emissions, it seems that governments don't know either, or they regard climate change as a low priority for themselves or their citizens. What, then, is the best approach when confronted with what might turn out to be a hugely important and urgent problem - but might not? Our current policymaking systems require that we have a good idea as to the likely impact of a problem, and that we then take measures to reduce that impact. But climate change isn't like that. We just don't know enough to take actual (as against performative) measures to deal with it. Too many of us are unconvinced of the need to do anything and, for many of us, it's against our interests to believe anything to the contrary. 

One of the reasons I advocate applying the Social Policy Bond concept to the climate is that we can confront this lack of certainty by, in effect, contracting out the risk to those prepared to take it on. Let's assume that governments collectively decide to back Climate Stability Bonds; that is, to put up funds for redemption of the bonds once our climate goals have been achieved and sustained for a period, of, say 40 years. If the consensus of the market for the bonds is that the climate is unlikely to change very much, then the bonds would sell for a quite high price. The governments would not lose much by redeeming the bonds, as the bonds would not appreciate very much. But if the market believes climate change is happening and that therefore climate stability will be difficult to achieve, investors will attach a low value to the bonds when they are issued. Bond purchasers would stand to make large sums if they help bring about climate stability.

The crucial point is that under a Climate Stability Bond regime it would not be up to governments, the United Nations, or any panel of experts to make a one-time only assessment of the seriousness of climate change. Under a bond regime it would be the market that would be highly motivated to inform itself about all aspects of climate change, because it stands to gain most if they get it right. And investors in the bonds would be so motivated on a continuous basis, as the market for the bonds would be constantly generating opportunities for gain to successful gatherers and interpreters of the flow of data about climate change.

A Climate Stability Bond regime would thereby bring onside the skeptics, or those who are just reluctant to pay large upfront costs for an uncertain gain. It would contract out not only the achievement of climate stability, but also the assessment of how serious a problem it is. The costs of a poor assessment would be borne by investors in the bonds, rather than taxpayers. Action to bring about climate stability – significant action rather than what we've seen so far – would, I believe, therefore be more forthcoming.

I've written in depth about Climate Stability Bonds on this blog and on my main site there there are links to other essays here. The same reasoning applies to other potentially devastating events to which we just cannot assign a probability; nuclear war for example, and I propose a similar application of the Social Policy Bond idea to address that and conflict in general: links are here.

Going beyond root causes

One strength of the Social Policy Bond idea is that it doesn't try to, nor need to identify relationships between cause and effect, between policy and outcome. It's not always efficient necessary to look for root causes. Indeed the perceived need to look for them can be an excuse to delay or ignore the social problem. Take war: it can have multiple causes, ranging from the childhood experiences of political leaders to ethnic rivalry, or the real or imagined need of a country for more resources, the insecurities of military commanders....

Rather than seek to identify, weight and address any of its myriad possible causes, a Social Policy Bond regime would instead target war itself. Holders of World Peace Bonds could choose to look for root causes, but only if they think that approach is worthwhile compared to alternatives. When the goal is peace sustained for, say, four decades, then alternative approaches could include: making the educational materials of schoolchildren less belligerent; encouraging student exchanges; influencing media outlets to tone down inflammatory propaganda; distracting, deposing or otherwise undermining leaders who foment violence and hatred. 

Policymakers, if they choose to, can do a good job of solving simple social and environmental problems, whose causes are obvious. But our complex society, with its multitude of variables, feedback loops and time lags makes it difficult for any single, conventional organisation - even ones as powerful as national governments - to deal with problems such as war, violence, crime, air pollution or poverty. Even if the causes of these problems were unchanging and invariate over geographic region, such a task would be difficult and contentious. But they aren't static nor uniform: they change over time and vary from one locality to the next. We need diverse, adaptive approaches, that no single conventional organisation, however rich and powerful, can possibly address: such organisations and the people working for them have their own agendas which often deviate from, or even conflict with, their ostensible goals.

Social Policy Bonds would lead to the creation of a new sort of organisation: ones whose sole remit is to solve targeted problems as efficiently as possible, and whose every activity would be in the service of society's targeted goal. Having a protean structure and composition, it would be motivated to respond rapidly to changing circumstances. Importantly too, it would not subscribe to any ideology that would limit the range of approaches it could explore.

The good people at antiwar.com took just a few hours to reject this brief essay, so I'm free to post it here. For followers of my work there's little new. Readers who wish to know more about applying the Social Policy Bond concept to war should please see this page and the links thereon.

Incentives for peace

We are, understandably, not quite rational about war. We regard its opposite, peace, as an ideal: as unattainable as it is desirable; something to aspire to from afar - something that will never actually happen. War appears to many of us, as it did to the ancient Greeks, to be part of the natural order of things. I think we can do better. What I am suggesting is that we put in place incentives for people to end war; to bring about permanent world peace through diverse, adaptive and efficient approaches.

We have to focus on our ultimate goal: sustained world peace. ‘Sustained’, because ceasefires, truces and other short-term measures often just postpone conflict. Our goal is peace that lasts for generations. We need to reward the successful achievement of this goal, rather than activities that are supposedly aimed at achieving it. There are too many people with a vested interest in keeping conflict going. While almost everyone would like to see a permanent end to war, there are too many in positions of power or influence who are half-hearted about peace, who feel threatened by it or who, for whatever reason, actively promote violence.

Ideally, then, we need a way of promoting peace that can modify or circumvent these people's uncooperative or obstructive behaviour. We need to mobilise the interests of the vast majority of people who want peace. We need to find ways of converting, bypassing, distracting, or undermining those opposed to our goal.

Ideally too, we would use market forces. In economic theory, and on all the evidence, markets are the most efficient means yet discovered of allocating society's scarce resources. But markets have been undermined, corrupted and abused such that they are nowadays mainly invoked only to defend extremes of wealth and poverty, or to further degrade the environment. So it is important to remind ourselves that competitive markets can serve society, rather than billionaires and large corporations.

World Peace Bonds

World Peace Bonds are a new way of channelling the market's incentives and efficiencies into what must be our highest priority: the permanent end of violent political conflict.

My suggestion is that a combination of NGOs, philanthropists, governments and ordinary citizens put up initial funding for a new type of financial instrument: World Peace Bonds. Funds for the redemption of the bonds could be further swelled by non-governmental bodies, and the wider public. The bonds would be floated by auction and redeemed for a fixed sum only when the number of people killed by violent political conflict fell to, say, 50 000 a year, for a sustained period. Importantly, the bonds would make no assumptions as to how to bring about peace, nor who would do so: these decisions would be made by bondholders. Unlike normal bonds, World Peace Bonds would not bear interest and their redemption date would be uncertain. Bondholders would gain most by ensuring that peace is achieved quickly.

Some years ago I envisaged a tradeable version of the now widely deployed Social Impact Bonds, which are increasingly accepted as a way of stimulating suppliers of social services to do better. World Peace Bonds, because they would be tradeable on the open market, would be closer to my original idea. People would buy bonds only if they expect to make a profit on them. If they're tradeable, they wouldn’t have to hold them to redemption to make a profit. The bond issuers could therefore target very long-term goals, such as our world peace goal, sustained for several decades.

Buyers of the bonds would work together, tacitly or otherwise, to improve the prospects for peace in the long term. As they did so, the market price of their bonds would rise, and they could realise a gain in value of their assets. This is the key: as the level of violence falls, so the bond price would rise. Bondholders would have incentives to do what they could to achieve peace, then to sell their bonds at a higher price to those who will take the next steps to our ultimate goal.

The bond’s backers, who supply the bonds’ redemption funds, need decide only on the definition of peace to be targeted - not on how to achieve it. That would be left up to investors in the bonds, who would have every incentive to maximise their, and the backers’, reduction in violence per unit outlay. So a sufficiently-funded World Peace Bond regime would stimulate research into, and implementation of, ever more cost-effective ways of achieving peace.

Bondholders would be in a better position than governments to undertake a range of peace-building initiatives. They could lobby or work with governments to, say, change and enforce laws that make wars at home or overseas a less likely prospect. They could finance sports matches between potential protagonists, promote anti-war programmes on TV, or set up exchange schemes for students and schoolchildren. They could try to cajole the financial supporters of conflict into redirecting their funds along more edifying lines. They could offer poor countries innovative forms of aid, including education and scientific aid, and measures aimed at enlightening populations. They might even subsidise intermarriage between members of different ethnic or religious groups. The crucial point is that bondholders have more freedom and incentive to explore and carry out such diverse, adaptive and long-term initiatives than governments or other international bodies.

In today's emotional climate decision-making is too often reactive. It is too easily swayed by those with a propensity for violence or those who benefit from it, whether financially or emotionally. There are enlightened, hard-working, people and organisations currently working for peace, but their ability to deploy resources effectively is constrained. The funding of the United Nations and other publicly-financed conflict-reduction bodies is conditional on their carrying out a range of activities limited by the bureaucracy and insecurities of their sponsoring governments. Private peace-building bodies work in admirable and diverse ways, but their efforts are small-scale and uncoordinated. For neither type of organisation are the financial rewards from building peace correlated with their effectiveness in actually doing so. Some of their approaches will be more effective than others but in the current environment nobody has incentives to find out which. World Peace Bonds, in contrast, would explicitly reward movement toward a sustained peace outcome, however it is done, and whoever does it. They would focus on an identifiable outcome and channel market efficiencies into exploring ways of achieving it. They could be the most effective means of achieving the peace that people all over the world yearn for and deserve.

Buy-in: just as important as efficiency

In my efforts to promulgate Social Policy Bonds I’ve usually emphasised their efficiency, which arises from a number of sources, including their harnessing of market forces, their encouragement of diverse, long-term approaches, and their capacity to adapt to changing circumstances. All these attributes are, in my view, essential if policymakers are going put in place systems that solve our urgent social and environmental problems.

Less obvious, but just as important, would be a bond regime's transparency: the bonds would target outcomes that are meaningful to ordinary people. So: if we are targeting national levels of crime, instead of arguing over funding, structures and operations of police forces, or surveillance cameras, or street lighting, etc, a bond regime would target crime, as experienced by the country's citizens. Yes, there would be arguments about how to weight different crimes, and how to define crimes in terms that could be robustly quantified, but the important point is that ordinary people understand the outcome and can (if we want) participate in the targeting process. As well, there's more consensus over outcomes than the supposed means of achieving them. 

Even if we take no real interest in the targeting process, our having the opportunity to make a contribution gives us something critical - but sadly missing - in today's political environment: buy-in.

This would apply even if our views are over-ridden by others: at least, we'd have been consulted. If people have the chance of participating in such discussion, we shall come to understand the limitations and trade-offs that are intrinsic to public policymaking. This means quite a few things, but to my mind buy-in is the most important. It could reconnect citizens with our policymakers; it would entail the sharing of responsibility and concern for policy initiatives.

This matters hugely when government has to do things that hurt people's narrow, short- or medium-term interests: improving our air and water quality, for instance. The current system discourages buy-in because it's difficult and tedious to follow. As such, it's easily influenced by the wealthy or powerful, be they in the private- or public sector, who can afford to pay people to follow the process. The lack of transparency does much to widen the gap between politicians and the people they are supposed to represent. Social Policy Bonds, because of their focus on outcomes, would help close that gap.

Pdfs of past blog posts

Posts on this blog from October 2024 to yesterday, 21 March 2025 can now be downloaded as a pdf from here.

All posts on this blog from December 2004 to yesterday, 21 March 2025 can now be downloaded as a - large - pdf file from here. 

Whipping up a tempest

From the current Economist: 

Mr Trump is expanding his threats, promising to hold Iran to account for the Houthis' attackes and warning them of 'dire' consequences. Yet his approach is only hardening the mullahs' hearts. They may seize the chance to rally their embittered people against a common enemy and go for confrontation and a nuclear momb. Israel might then join the fray. The Houthis would come to their patron's [Iran's] aid and fire again at Gulf cities and oil terminals. It is all too easy to imagine the worst. America’s strikes on the Houthis could whip up a regional tempest, the 'Economist', 20 March 2025

Indeed. I keep returning to the possibility of a nuclear exchange, because it would be a catastrophe in its own right, as well as a terrible portent. We could rely on some combination of world governments to make nuclear conflict less likely but, in today's political environment, I think we should look for a complementary solution. 

My suggestion is that we issue our own (pdf) Nuclear Peace Bonds. All it would take would be for some interested philanthropist to put up the funds, and let the market for the bonds do the rest. Of course, once the ball got rolling, contribution from other bodies and members of the public could be solicited, which would swell the total redemption rewards. Even governments, if they could put aside their short-term interests for a moment, could add to the pot. 

The goal of sustained nuclear peace would actually make an ideal target for the Social Policy Bond idea. One, because it's a complex, long-term goal that will require diverse, adaptive solutions. Two, it's an easy goal to verify. And lastly, it's a goal that, on all the evidence, including that of the above excerpt from the Economist, is unlikely to be reached under current policy. The idea would be to issue bonds that reward a sustained period of nuclear peace. This could be defined, as, say the non-detonation of a nuclear device that kills more than 100 people for 40 years - the long time period is necessary so that systems are put in place that work in the long term. With sufficient backing the bonds would help offset and (one hopes) outweigh the incentives currently on offer to the military-industrial complex and to ideological and religious fanatics.

Those billions of us who would benefit from nuclear peace are presumably a massive numerical majority, but we currently have few means of channel our wishes effectively. The tendency is to assume that governments will do what's necessary, with the support of hard-working, well-intentioned people in the private sector. But the rewards to all these people are not linked to their success. This is unhelpful in itself but, more importantly, it discourages investors who, seeing little opportunity to benefit from working to reduce nuclear conflict, will focus instead of less edifying enterprises. Most important of all is that our current strategy is just not working.

We need to reward those who achieve nuclear peace at least as much as those working to undermine it. We don't know exactly how to reduce the chances of a nuclear exchange, nor who will be best placed to do so, over the long period during which our goal is to be achieved, but we have no excuse for not encouraging people to find out. Nuclear Peace Bonds would apply the Social Policy Bond principle to this goal. Investors in the bonds would form a protean coalition of people dedicated to achieving it as efficiently as possible. Their goal would be exactly the same as society's. Human ingenuity knows no limits. Currently, too much of it is devoted to relatively unimportant or socially questionable goals. Nuclear Peace Bonds would channel our ingenuity, and stimulate more of it, into minimising the risk of a global catastrophe. 

My short piece on Nuclear Peace Bonds is here. The links in the right-hand column of that page point to papers on similar themes: Conflict Reduction, Disaster Prevention, and Middle East Peace

Nobody takes a panoptic view

Peter Gøtzsche writes:

In 2024, PubMed indexed 1,728,666 articles. Compare this with the little progress there is in healthcare from year to year and consider also that most research results are unreliable or outright false. Ridiculous names for predatory journals (clicking the link will download a docx file), Peter C Gøtzsche, Institute for Scientific Freedom, 7 March 2025

This is exactly the sort of overview we need to take in health and other policy areas: a comparison of the resources devoted to a goal and the actual outcomes achieved. Unfortunately, few people have the incentive or capacity to make such comparisons. Professor Gøtzsche himself has been vilified for questioning the role that the pharmaceutical industry plays in psychiatry. Yet taking a panoptic view makes it clear that there's something very wrong with the world's healthcare. It's not just healthcare:

From 2004 to 2014, aid spending increased by 75%. “There was a real feeling,” says Stefan Dercon of the University of Oxford, “that if there was a time things were going to get going, this was it.” Things did not get going. From 2014 to 2024, the world’s 78 poorest economies grew more slowly than in the decade to 1970, when aid was first emerging. This is perhaps unsurprising, given earlier studies. In 2004 William Easterly of New York University and co-authors found that, from 1970 to 1997, aid was just as likely to shrink the world’s poorest economies as to help them grow. Aid cannot make poor countries rich, the 'Economist', 6 March 2025

Again, how many people take this sort of overview, and what influence do they have over policy? Very few, and negligible, I'd say. Arguably, the same failings occur in education in some of the rich countries. Certainly they apply to climate change if we take greenhouse gas emissions as an indicator of policy success:

Why are we so hopeless at making effective and efficient policies? One answer is that we rarely evaluate their effectiveness:  

[G]overnment bureaucracies non-self-evaluate. At a minimum, agencies with evaluative responsibilities are not invited to evaluate - they are kept out of the loop, their opinions unsought. At a maximum, government agencies actively suppress their own internal evaluative units and are discouraged from evaluating the beliefs and policies of other agencies. Why States Believe Foolish Ideas: Non-Self-Evaluation By States And Societies (pdf), Stephen Van Evera, Massachusetts Institute of Technology Political Science Department and Security Studies Program, 2002

Another answer is that our policies do not adapt to changing circumstances, nor are national policies sufficiently adapted to different regions. From the same article in the Economist: 

[D]isillusioned economists have turned to the work of Esther Duflo, a Nobel laureate at the Massachusetts Institute of Technology, who uses randomised controlled trials to study interventions. Yet she has come to a dispiriting conclusion: there is no reason why what works in one neighbourhood will do so in the rest of a district, let alone on another continent. In one Indian village, for instance, giving women pensions made their granddaughters (if not their grandsons) healthier; in another, handouts failed to improve health or even raise household consumption. Ms Duflo’s findings chime with other research....

Especially for long-term goals, we need policies that are diverse and adaptive. Social Policy Bonds, as well as injecting market incentives into the solution of our social and environmental problems, would encourage investors to explore different approaches, to refine those that are most promising and, importantly, to terminate failing approaches. 

For more about how the Social Policy Bond principle could be applied to health, see here. For how it could be applied to development see here.