Injecting market incentives into the solution of social problems: Social Policy Bonds - Social Policy Bonds

Go to content

Main menu:

Injecting market incentives into the solution of social problems: Social Policy Bonds


February 2017: the paper below was written about 20 years ago. I am currently working on updating it.

Despite decades of economic growth and efforts to make their public sectors more efficient, most western countries are plagued by seemingly intractable social problems. Deregulation and the freer operation of self-interest in parts of the private sector have made many individuals very wealthy indeed, but the less well off have gained little, and many social objectives remain as remote as ever. Most people agree on what needs to be done: unemployment must come down; housing, education and health services must all be improved, and the environment should be cleaned up. At the global level too, there is broad consensus over objectives: nuclear war must never be allowed to happen, and environmental catastrophe, whatever form it takes, must be avoided. All these challenges are urgent, and all are largely resistant to current efforts to address them.

I think this is because too often, instead of rewarding meaningful outcomes, we reward organizations, activities, inputs or outputs, in ways that do not encourage effectiveness or efficiency in achieving our social goals. I propose that a new financial instrument, Social Policy Bonds, be issued as a means of injecting market incentives into the achievement of our social and environmental outcomes, to the benefit of all. Importantly, these bonds need not be issued by government: they can be issued by anybody - public or private sector, non-governmental organizations, philanthropists, or members of the public - or indeed any combination of such bodies that collectively wants to solve any quantifiable social or environmental problem.

Social Policy Bonds

My proposal is to issue non interest-bearing bonds that are redeemable only when a targeted social objective has been achieved. Social Policy Bonds would be issued by government, or private sector groups, at whatever price they will fetch on the open market and thereafter could be bought and sold by any willing individual or institution at their free market price. Once the targeted objective had been achieved, they would each be redeemable, as are conventional bonds, for a fixed, predetermined sum. They therefore differ from conventional bonds in that they would have an uncertain redemption yield which, in combination with a fixed redemption value, implies an uncertain yield.

Take for example the goal of reducing unemployment. Assume that a fixed number of bonds is issued, redeemable for £10.00 only when unemployment is down to, say, three per cent. Those bonds are floated, nationally and overseas, on open tender as at an auction: those who bid the highest price for the limited number of bonds available will be successful. Importantly, all bonds would find a buyer: the price might be very low, but the bonds will all be sold. What factors will determine the price of these bonds? Most obviously, the market's assessment of how likely and when the objective will be achieved. Interest rates on alternative investments will be a factor. The bonds could go for as little as one penny if people thought there was virtually no chance of this particular social goal being achieved in their lifetime. People will of course differ in their assessment of the value of the bonds, and their views will change with time as events make achievement of the targeted objective a more or less likely prospect. The bonds, once issued, would be tradeable at any time; market prices would be publicly quoted just like those of ordinary bonds or shares.

Assume now that the bonds targeting unemployment have been issued and sold. The opening value of the bonds might have been, say, £2.50. People, or institutions, now hold bonds that can quadruple in value once unemployment is down to three per cent. The bond issuers have nothing more to do: the holders of the bonds now have a strong interest in seeing the value of their bonds increase as quickly as possible. If other people's interest is stronger they will bid more for the bonds than they are worth in the hands of current holders. So Social Policy Bonds will be owned by those with the strongest interest in seeing the objective attained. These owners will form a new type of organization: one whose structure, composition and activities are entirely subordinate to the targeted outcome.

Who would buy the bonds?

Passive investors hoping to make a capital gain

These include:

  • casual purchasers, who might buy bonds in the same way as they would a raffle ticket;

  • speculators, who believe that the likelihood of the targeted objective being achieved is greater than the rest of the market thinks it is; and

  • hedgers, who stand to lose if the particular objective is actually achieved. They might buy bonds as a form of insurance policy against this possibility.

All these passive investors would want to become 'free-riders', hoping to benefit from any increase in the bond price without actually participating in any objective-achieving projects. But the way markets work would limit the opportunities for these would-be free riders. The more bonds these passive investors own, the more remote the targeted objective becomes, and the more they would stand to lose as the value of their bonds falls. So passive investors, if their collective holdings became significant, would find themselves becoming poorer and poorer, until they sold them to - or became themeselves:

Active investors

These include:

  • organisations trying to achieve a targeted goal, whose bond-holding would directly enhance the profitability of any successful actions they take, or

  • investors or brokers who would use their own capital, or borrow on the strength of the redemption value of the bonds, in order to support projects aimed at helping achieve targeted objectives.

Active holders, in the unemployment example, could be expected to increase employment by using part of the present value of their expected above-normal yield from early redemption of the bonds to finance their own, or others', labour recruitment drives. Note that active investors in the bonds would co-operate or collude with each other to help achieve the targeted objective.

Prospective holders of the bonds have an incentive - and given free capital markets, the means - to buy them from current holders if they think they can do a better job of achieving the targeted objective. Thus, the provision of the means by which the objective is attained is not in the hands of entrenched interests.

Many of the initiatives that would be stimulated by the unemployment-reducing bonds, for example, are taken by governments nowadays, but the critical difference is that, under a bond regime, the initiatives are stimulated by the self-interest of the bondholders and are not operated by a bureaucracy that, however well-intentioned, is not rewarded in ways that correlate to its success in achieving objectives. Social Policy Bonds provide a strong motivation for bondholders to seek out those ways of reducing unemployment that will give them the best return for their outlay. The bonds direct self-interest into those processes necessary for objective achievement that will respond most readily. The bond issuers do not have to plan this: it is the self-interest of bondholders that ensures it.

Current efforts by government generally focus on the most obvious symptom of a social problem - not on the problem as a whole. So inefficient industries on the verge of bankruptcy might receive vast amounts of taxpayers' money at the expense of cheaper job-creation initiatives. The bonds improve on ad hoc arrangements that are not only inefficient, but also expose decision-makers to bribery or corruption. Another significant advantage over conventional policy is that whoever backs the bonds, be they taxpayers or other bodies, pay only when their targeted objective has been achieved.

Note that agencies that are currently fully funded by government could also benefit from holding the bonds. In the transition to a bond regime, their funding from tax revenues would fall but, if they were efficient, the value of their bonds would rise - or they would receive funds from bondholders acting as investors in efficient objective-achieving projects.


Key criteria for policy areas within which Social Policy Bonds would show the most marked improvement over current programmes are:

  • existing policies have objectives that are unstated, uncosted, obscure or conflicting;

  • financial rewards to those involved in achieving objectives are uncorrelated to their effectiveness in doing so;

  • a wide array of diverse, adaptive approaches may be necessary;

  • our knowledge of the problems, their causes and solutions, is improving all the time.

Many national and global social and environmental problems satisfy these criteria, especially those that are currently the responsibility of governments or are not targeted at all. So, for instance, health care or crime prevention are largely dependent on government-financed bodies, whose employees have little or no incentive actually to be efficient. Global problems, such as civil war, or war between countries, are hardly targeted for reduction in any explicit, effective way - certainly not in ways that reward effectiveness or that allocate resources to their prevention in proportion to the suffering they cause.

Social Policy Bonds could allow the targeting of such problems, each of which probably requires diverse, adaptive approaches, of the sort that cannot be organised by any government or, indeed, by any of our current organizations. Long-term goals, such as world peace, could be targeted effectively because the bonds do not stipulate how such a remote objective shall be achieved, but instead reward those who actually achieve it once they have done so, rather than those who are simply charged with that goal and are paid for activities that may or may not do accomplish anything positive. The bonds reward only those ways of achieving policy goals that are efficient: these can be direct or indirect, and they can be completely different from any currently carried out. Importantly also, even if the bonds were to stimulate inefficient projects, the costs of these inefficiencies would be borne by the private sector, not taxpayers and the bond mechanism would see them halted before they became too much of a drain on resources.

Social Policy Bonds and social goals

Crime prevention

Currently touted methods of combating crime are longer sentences for convicted criminals, and more money for the police force. Social Policy Bonds targeting crime would home in directly on what society actually wants to achieve: they would become redeemable only when crime and the fear of crime had fallen to low levels.


Social Policy Bonds targeting unemployment could replace a wide range of measures including protectionist barriers to imports of labour-intensive manufactures, which are aimed at maintaining employment in certain industries. Here the efficiency gains from bonds that target unemployment directly could be dramatic.


Priorities for health services are strongly influenced by groups of medical specialists with little incentive or capacity to see improvements in the general health of the nation as an objective. So funding of medical specialities depends on the strength of their lobby groups. And what is arguably the most efficient way of spending the taxpayer's health dollar - preventive medicine - receives derisory funding, because it has no powerful lobbyists.

Targeting general indicators of well-being - life expectancy, infant mortality, disability, or Quality Adjusted Life Years - would ensure that scarce resources are allocated in ways that would directly and impartially achieve society's health objectives. Health Bonds would divert funds into those areas of the health service that would most efficiently use them to achieve the targeted objectives, but they would also encourage exploration and implementation of ways of improving health that are currently outside the purview of healthcare bodies.


Bonds could target results achieved in basic literacy and numeracy tests taken by schoolchildren.


Bonds could target nationally averaged levels of water or air pollution, or the effects of such pollution on human, animal and plant life.

In all these examples there would be difficulties in the specification of the objective to be attained. 'Approved housing units' for instance, or 'reported crimes' could be subject to varying interpretation, or to deliberate attempts to falsify the information required to monitor achievement of the objective in question. But these difficulties are not insuperable, so long as the following three processes are soundly carried out:


The objective must be capable of being quantified, or there must be a strongly correlated proxy for the objective, whose targeting would inevitably result in the objective being achieved.


Careful thought will have to be given to the definition of each objective targeted by the bonds. Consider the unemployment example. It would be unsatisfactory to redeem the bonds when unemployment was down to a certain level for a short time only. The objective is a sustained level of low unemployment, and this is how it would have to be defined when the bond is issued. The Social Policy Bond principle is versatile in that it can target combinations of targets, all of which shall have to be reached and sustained before the bonds are redeemed.


All bond issues will require reliable and accurate monitoring of the targeted problem so that progress toward its solution can be reliably and unambiguously assessed. This surveillance must also be seen to be independent of the government or interest groups, both of which could benefit unfairly from dubious data collection. The nature of the monitoring - whether it would carried out at local, regional or national level, for example, or the level of aggregation at which independent organisations are involved - will depend on the objective being targeted and, to some extent, on the amount of funding at stake.

The market for Social Policy Bonds

For the bond mechanism to work it is essential that active investors purchase the bonds and help to solve social problems. But there is no need artificially to boost investor interest in the bonds: the anticipated supernormal profit arising from early redemption of the bonds generates the required self-interest and so supplies the motivation for achieving the government's objective provided there is a buoyant market for the bonds.

Social Policy Bonds, once issued and sold, must be readily tradeable at any time until redemption. (See Why the bonds must be tradeable.) This is critical to the operation of the bond mechanism. Many bond purchasers will want, or need, to sell their bonds before redemption - which may be a long time in the future. If the bonds were not tradeable (as with Social Impact Bonds, for instance) , these holders would not be able to realise any capital appreciation experienced by the bonds. This would remove much of the incentive to purchase the bonds in the first place.

But there is another important reason for requiring the bonds to be tradeable: active investors may be able to speed up only one, or a few, of the processes necessary for the targeted objective to be achieved. Once these investors have done their bit, and seen the capital value of their bonds in line with the increased probability of the bonds' early redemption, they may have no wish to speculate on the speed at which the remaining processes will be carried out. Other groups of active investors, who will have greater expertise in performing these later processes, must be given an incentive to use their expertise to accelerate attainment of the targeted objective. The possible capital appreciation of bonds bought from previous owners and sold at a still higher price [or redeemed] provides this incentive. The new owners will, if they are successful in these later stages, realise this capital appreciation.

Bonds therefore would, as from would-be free riders, tend to flow outwards from those who had helped as much as they can to achieve the targeted objective. In fact, though, it is not necessary for there to be any actual flow of bonds. What would flow could be bonds, but it could also be finance from bondholders. The important point is that the bond mechanism ensures that the people who allocate the finance have just as much incentive to allocate efficiently as those who receive it have to solve the targeted social problem efficiently. At the limit we can conceive of just one single buyer of all the bonds. If this buyer were determined to hold on to the bonds until redemption, then the bonds would function as a sort of performance-related contract, with the government paying only when the objective has been achieved. The buyer could contract out most, or all, of the work required to achieve the objective, with the incentives given by the bonds for speedy accomplishment cascading down from the bondholder to those subcontracted to do the work.

Anther source of efficiency is that the market for Social Policy Bonds will generate extremely valuable information for policymakers. They will do so even as the bonds are issued: the price they fetch will be an important indicator of how remote the market believes is the targeted objective. Thereafter bond prices, and the way in which they change, will supply continuously updated information on which policy programmes and events are, in the market's view, likely to be most effective at achieving the targeted goal.

Government, while it may profit from appreciation of the bonds it purchases, will also be interested in the cost of its social policies. The Social Policy Bond principle is superior to existing budgetary mechanisms in that the cost of each scheme is not only inexorably linked to attainment of its objective, but its maximum cost can be decided in advance. The number of bonds is limited, and the most the scheme could cost the government would be the cost of redeeming the bonds very soon after they are issued (assuming a negligible issue price] plus all the administrative costs. Even then, though, the objective will have been achieved before any cost is incurred.

The efficiency of the bonds could be tested by allocating the same sums of money as are currently allocated for a particular social objective to the redemption of bonds targeting the same objective. The maximum cost to the government of the issue could then be set so as not to exceed the expenditure that would anyway have been incurred in pursuit of the same objective.

We should note too that Social Policy Bonds allow for the complexity of social problems. No single approach will solve them, so a wide variety of approaches to their solution is essential. Social Policy Bonds will encourage and reward the most efficient of these approaches. This occurs because of the nature of the bond mechanism, and requires no selection or supervision by government [or government agency] of the most efficient policy. Only the objective, not the policy, is dictated by government. This feature tends to stabilise the political environment. Obviously the objectives will have to be carefully defined, but there is near-unanimity over most social goals. A government is unlikely to repudiate such universally desired objectives, even if the associated bonds had been issued by ruling parties with a different political outlook. The risk that it might, and so become the first government openly to support higher unemployment, worse standards of health care, etc, would not be much greater than that of a government refusing to redeem fixed interest stock issued by any of its predecessors. This risk, always present, in now way impedes the operation of bond markets.


  • The main advantage of Social Policy Bonds is that they make the achievement of social objectives more efficient by injecting self-interest into the exploration, implementation and refinement of projects aimed at achieving social and environmental goals. Additional gains accrue for other reasons:

  • The bonds guarantee stability of policy objectives. Policy instability is an important reason why people do not undertake projects or activities that could benefit society. Objectives with a necessarily long lead time - to halt climate change or bring about world peace - could be targeted by a bond issue, and holders of the bonds would not be deterred from taking measures to achieve them by fears that the goal itself, rather than the means of achieving it, had become unpopular or made obsolete by new knowledge.

  • The bonds make policy objectives more transparent. Apologists for current policies often point to benefits that can result only haphazardly - if at all - from their implementation. Social Policy Bonds would ensure that objectives are explicitly identified, and that indirect methods of achieving them would be encouraged only if they were efficient. Policymaking often takes the form of arcane discussion about funding arrangements of institutional structures. A bond regime would focus on goals that are meaningful to ordinary people. Crucially, this means that the public can participate in policymaking debates; this can be an end in itself, but it also means more buy-in to agreed social and environnmental targets, and so help close the ever-widening gap between politicians (and their sponsoring corporations, and bureaucrats) and the people they are supposed to represent.  Also, note that explicit targeting of objectives is likely to lead to explicit calculation of the value of their achievement - a useful discipline, but one rarely followed by today's politicians.

  • A less obvious distributional benefit would arise from the existence of a means of acquiring wealth with which private gain is strongly correlated with public benefit. Many bondholders would be rich and, if their bonds were redeemed early, they would become richer. But this would be a socially acceptable way of acquiring wealth. And the existence of such a way of accumulating wealth would allow other, less socially beneficial ways to be taxed more heavily.

Potential problems

The biggest potential problem of Social Policy Bonds is probably the incentive they will give bondholders to achieve specified objectives at the expense of other societal goals. For instance, assume that the concentration of atmospheric lead is targeted in a bond issue. It might be that targeting lead in this way would cause people to increase their use of substitutes - which could be more dangerous than the original levels of lead. One way of anticipating this problem could be to aim initially at unambitious reductions in the lead level. Depending on the effects of this reduction on the use of offending substitutes, other bonds could then be issued targeting these substitutes, or further targeting the level of lead. A better approach though would be to target, more comprehensively, atmospheric pollution. This could be expressed, perhaps, as an index of atmospheric pollutants weighted according to their lethality and other factors. In general, objectives that are complementary and that, if not pursued jointly, could conflict, should be targeted by a single bond issue.

Another safeguard against legal, but negative activities undertaken in pursuit of a targeted objective, could be provisos on the bonds specifying indicators of social welfare which, while not explicitly targeted by the bond issue, must be satisfied for the bonds to be redeemed. Thus Social Policy Bonds targeting unemployment could embody provisos to the effect that the bonds would not be redeemed if the inflation rate exceeded a certain limit.

Illegal activities could, of course, be dealt with by existing laws, possibly backed by a system of bondholder registration, which would identify those with the biggest incentive to commit them.

The question of how well Social Policy Bonds would achieve societal goals needs to be considered alongside current policymaking methods. In today's environment policymakers can escape or deflect censure because the adverse results of their policies are difficult to relate to their cause. If the bonds were to lead to negative-but-legal effects, the relationship between these effects and their cause would be identifiable, and the filtering out of negative effects would be a simple matter compared to the methods available to today's policymakers.

Of course, a Social Policy Bond regime would need some getting used to. There would need to be a transition period during which existing bodies, whose funding is largely unrelated to their effectiveness, would see a fall in their resources, matched by the possibility of increased funding from bondholders dependent on how effective and efficient they are.


Resources are always going to be limited and Social Policy Bonds will not change that. Priorities and choices will always have to be made: under a bond regime the choice of problems to be targeted, and the funds allocated to their solution, will all be limited. The difference is that the bonds would allow and encourage public participation in the forming of policy goals and their relative priority.

Today's democratic governments are actually quite good at representing and articulating their people's wishes; they can be effective and even necessary when a single, relatively clear, policy approach is the beest solution to a social or environmental problem. They fall down, though, on complex and longer-term policy issues, where a diverse array of adaptive approaches will be the only route toward an effective solution. Government or, indeed, any single conventional organization, finds it difficult to assimilate and act on new scientific knowledge, either about problems or potential solutions. (For more on this see A new type of organization.) As society grows more complex, these deficiencies become more accute.

A Social Policy Bond regime would see government, in effect, contracting out the achievement of social goals to the private sector, though importantly government would still be articulating our social goals and could play an essential role in raising the revenue for their achievement. Even so, the surrendering of responsibilities to the private sector could be politically difficult, and must be a gradual process. But the potential benefits should not be ignored. Western governments' spending on education, health, the justice system, the environment and other items continues to grow, despite all efforts to restrain it. Even relatively small gains in efficiency in this spending could greatly benefit those who are most in need.

In economic theory, and on all the evidence, markets are the best way of allocating scarce resources to achieve prescribed ends. Social Policy Bonds allow governments, individually or collectively, to do what they are best at - prescribing ends - and markets to do what they are best at - allocating resources to meet these ends. In so doing, the bonds can achieve a society's social goals more efficiently and less randomly than the current combination of ad hoc policies and trickle down. But even more, they can encourage the targeting of goals that can be solved only in the long run and only in ways that cannot be anticipated by any number of actors in today's world: these goals, thought to be too lofty or idealistic with today's policymaking technology, could become realistic and, given our boundless ingenuity, easily achievable if they were targeted by Social Policy Bonds. These goals, desired by the vast majority of human beings on this planet, are worth the relatively small transitional costs that would be incurred by a change from the current system. Goals like universal literacy, the minimising of the risks and consequences of natural disasters, or world peace.

For more about the operational aspects of Social Policy Bonds, click on The book. For examples of how the Social Policy Bond concept can address health, education, environnmental and other concerns, click on Applications.

free web counter
Back to content | Back to main menu