Prevention can be worth far more than it costs to society and still be a bad deal for whoever is asked to fund it. Until the question of who pays is answered as carefully as the question of what prevention is worth, the gap between the two will keep deciding what gets done.
Two paradoxes with one name
In 1981 the epidemiologist Geoffrey Rose described what became known as the prevention paradox: a measure that brings large benefits to a population may offer little to each person who takes part in it. Most people who lower their blood pressure would never have had a stroke. The benefit is real, but it is visible only in aggregate.
Biological security has Rose’s paradox and a second, sharper one. The second is not about individuals and populations. It is about payers and beneficiaries. When a veterinary service funds surveillance that catches an exotic disease early, the benefit is spread across livestock keepers whose animals are not culled, processors whose supply is not interrupted, exporters whose markets stay open, rural businesses whose visitors keep coming and, for zoonotic disease, the health service. The veterinary service itself receives only part of the benefit — the response spending it no longer has to make. The rest flows to people who did not pay.
The flagship article in this issue identified this misaligned ledger as one of five structural reasons why biological security underinvests before the crisis. This piece examines it in detail: who benefits from prevention, why the mismatch persists, what compensation does to it, and how some countries have tried to close it.
A worked example of the gap
The methodology note for the OHS Prevention Ledger includes a fictional illustration that makes the problem concrete. A government veterinary service considers a five-year enhanced surveillance programme for an exotic livestock disease. In the central case, the programme costs £1.9 million in present-value terms and reduces the probability of an outbreak over five years from 30% to 15%. The expected value of the consequences it avoids is £3.25 million. For society, the Prevention Ratio is 1.7: the programme is worth 1.7 times what it costs.
But only £1.22 million of that avoided value is response spending the government would otherwise have incurred. The remaining £2.03 million is animal losses that would have fallen on producers. Judged by its own budget, the veterinary service would be spending £1.9 million to save £1.22 million — a ratio of 0.64. The programme is worth doing for society and not worth doing for the organisation that has to find the money.
The numbers are invented; the shape is not. It recurs whenever the payer is a single public body and the beneficiaries are many.
Who actually benefits from prevention?
The first step in taking the paradox seriously is to map the beneficiaries properly. For an exotic livestock disease in a country like the UK, the list is long.
Government. Avoided response costs: culling, disposal, cleansing and disinfection, veterinary and administrative staff, and — the largest single item in many outbreaks — compensation. Avoided indirect costs to the Exchequer through lost tax receipts and emergency support to affected sectors.
Farmers whose animals would have been infected or culled. Compensation rarely covers the full loss: business interruption, loss of breeding lines, restocking delays and the years it takes to rebuild a herd are generally not compensated, and nor is the psychological cost that After the Cull examined in Issue 001.
Farmers who are never infected. Movement restrictions, closed markets and falling prices affect producers far from any infected premises. So do trade bans: when a country loses disease-free status, every exporter in the sector is affected.
Processors, hauliers, markets and retailers. Supply chains lose throughput and margins, sometimes for months after an outbreak ends.
The wider rural economy. The 2001 foot-and-mouth epidemic demonstrated this most vividly. Estimates published the following year put losses to tourism at a scale comparable with the losses to agriculture and the food chain. Most of those tourism businesses had nothing to do with livestock and received no compensation.
The health system. For zoonotic disease — avian influenza, rabies, bovine tuberculosis, brucellosis — prevention in animals is prevention in people, and the savings fall on a different department’s budget.
Insurers. Where livestock or business-interruption cover exists, prevention reduces claims.
Households. Through food prices, through rural employment and, for zoonoses, through health.
Other countries. Disease controlled at source protects trading partners and neighbours. The protection a country gives others is part of what its prevention buys, even if it appears in no national account.
Two features of this list matter for policy. The first is that the payer — usually a single government department — is one line among many. The second is less obvious: not everyone benefits from prevention. When an outbreak closes export markets, domestic prices may fall; when it removes a competitor’s supply, prices elsewhere may rise. Some producers and some trading partners gain from outbreaks they did not suffer. That makes voluntary cost-sharing harder, because the potential contributors do not share a single interest.
Why the gap persists: three features of biological risk
Prevention is partly a public good
Economists describe a public good as one that is non-rival — one person’s benefit does not reduce another’s — and non-excludable — people cannot easily be prevented from benefiting. Disease freedom has both properties. When a country is free of foot-and-mouth disease, every livestock keeper benefits, and none can be excluded from the benefit because they declined to pay for it.
Public goods are classically underprovided by voluntary contribution, because each potential contributor can benefit from others’ contributions without making their own. That is the argument for public funding. But public funding brings the payer’s paradox back: the department that funds the good receives only part of its value, and competes for budget against services whose benefits are more visible and more attributable.
Much of prevention’s value is an externality
An externality is a cost or benefit that falls on someone who is not party to the decision that caused it. A farmer who invests in biosecurity protects not only their own animals but their neighbours’. A farmer who delays reporting a suspicious lesion puts every farm within reach of the virus at risk. In both cases, the farmer bears the full cost of the decision but captures only part of its consequences.
The result is predictable. Private biosecurity investment will tend to be lower than is socially optimal, because the farmer is paying for protection that others receive free. And the private incentive to report will tend to be weaker than the social value of reporting, because the farmer bears the cost of the cull while the benefit of early detection is shared across the industry.
Surveillance is a weakest-link good
Some public goods depend on the sum of everyone’s contributions. Others depend on the smallest. The economist Jack Hirshleifer called the second kind weakest-link goods: a dyke is only as high as its lowest section. Disease surveillance is close to this model. A national surveillance system with excellent coverage in most regions and a gap in one is not 90% effective; the pathogen that enters through the gap spreads regardless. Internationally the same logic applies: a region is protected against a transboundary disease only as well as its least capable member can detect and report it.
Weakest-link goods have an unusual implication. The most valuable investment is often not in the strongest part of the system but in the weakest — frequently someone else’s. That is the economic logic behind international funding for veterinary services, which WOAH’s Performance of Veterinary Services Pathway and the Pandemic Fund both reflect. It is also the logic behind the argument, made in the flagship, that surveillance weaknesses in one jurisdiction are vulnerabilities for all.
Compensation: the hinge of private incentives
If the externality problem explains why private incentives are weak, compensation is the main tool governments use to strengthen them. It is also where the design choices are most consequential, and most often made for reasons other than incentives.
Compensation for animals culled in disease control does two things at once. It shifts part of the cost of an outbreak from the farmer to the taxpayer, which is a question of fairness. And it changes what farmers do, which is a question of behaviour. Issue 001’s Paying for Cooperation examined the Danish and British models from this second perspective; the economic literature helps explain why the design matters so much.
The core tension was set out by Gramig, Horan and Wolf in 2009. Compensation that is too low discourages reporting: a farmer who expects to lose everything if the disease is confirmed has a reason to hope it is something else, and to wait. Compensation that is too generous creates moral hazard: a farmer who expects to be made whole has less reason to invest in costly biosecurity, because the downside of infection has been insured by the state. Earlier work concluded that no single compensation level could solve both problems.
Rob Fraser, writing in Environmental and Resource Economics in 2018, argued that the dilemma is less stark than it appeared. Modelling farmers who are risk-averse, he found a range of compensation — broadly between 40% and 60% of market value in his analysis — within which farmers have incentives both to invest in biosecurity and to comply with reporting requirements. More risk-averse farmers widen that range. His conclusion about the British arrangement, which compensates at full market value, was blunt: it is “almost certainly destructive” of the incentive for voluntary biosecurity investment.
That finding should be read with care. It is a model, not an observation, and the right range will vary with the disease, the sector and how farmers actually perceive risk. It also leaves out much of what After the Cull described — the losses that compensation does not cover, and the trust that determines whether farmers believe the rules will be applied fairly. But it establishes something important for the prevention paradox: compensation is not only a transfer from taxpayers to farmers. It is a price that the state sets on the farmer’s decisions, and it can be set to reward prevention or to undermine it.
Compensation rules also shape incentives in quieter ways. In Great Britain, compensation for avian influenza is paid for birds that were healthy when culled, not for birds that were already infected or dead. The logic is to reward early reporting, since a keeper who reports promptly will have more healthy birds. But the same rule means a keeper who reports late loses more — and the effect on behaviour depends on whether keepers understand and trust the rule. Timing matters too: in 2022 the UK changed the scheme so that compensation could begin at the start of planned culling rather than at the end, because delayed payment was itself a cost to cooperation.
Closing the gap: three models of cost-sharing
If the benefits of prevention fall on many parties, one answer is to ask more of them to pay. Several countries have built standing arrangements to do this. Their experience is instructive, not least because it shows that cost-sharing changes governance as well as funding.
Australia: a deed agreed before the emergency
Australia’s Emergency Animal Disease Response Agreement (EADRA), signed in 2002, is a legally binding deed between the Australian Government, all state and territory governments and the national livestock industry bodies. It sets out in advance how the costs of responding to an emergency animal disease will be shared, according to the resource base of the parties and who benefits. Diseases are assigned to categories that determine the split. For foot-and-mouth disease, government pays 80% of response costs and the affected industries 20%.
Two design features stand out. First, the industry share is underwritten: the Commonwealth pays it at the time and industry repays it through producer levies over up to ten years. That removes the problem of raising money in the middle of a crisis. Second, compensation to owners is explicitly designed to remove disincentives to report. Because industries are parties to the deed, they also take part in response decisions and commit to biosecurity planning — a seat at the table in exchange for a share of the bill.
New Zealand: sharing readiness as well as response
New Zealand’s Government Industry Agreement for Biosecurity Readiness and Response (GIA), grounded in the Biosecurity Act 1993, extends the principle to prevention. Industry signatories and the Ministry for Primary Industries make joint decisions and share costs under operational agreements that cover both readiness — planning, surveillance, preparation — and response. That matters for the prevention paradox, because it is readiness, not response, that suffers most from the misaligned ledger. A response is funded because it has to be. Readiness is funded only if someone chooses to.
England: considered and deferred
England came close to a comparable arrangement. In January 2010 the government published a draft Animal Health Bill that proposed an independent body for animal health and a framework for sharing the costs of disease control with the livestock industry. The proposals were contentious with the industry, and after the 2010 election the new government changed course. In April 2011 it established the Animal Health and Welfare Board for England, bringing farmers, vets and officials together to advise ministers, and stated that cost-sharing would be taken forward in a staged approach once responsibility-sharing arrangements were in place, with the Board to advise on funding. A standing arrangement comparable with EADRA or GIA has not followed.
Much has changed since. The Public Accounts Committee reported in November 2025 that Defra and APHA have been in “semi-permanent response mode” for years, that surveillance has been reduced as a result, and that there is no overarching long-term strategy for animal-disease resilience. Those are symptoms the flagship traced to the misaligned ledger. They strengthen the case for revisiting cost-sharing — not primarily to raise money, but to give the parties who benefit from prevention a stake in funding and governing it.
What cost-sharing buys, and what it risks
Cost-sharing arrangements are not simply a way of moving money from one pocket to another. They change who is in the room. Industries that pay a share of response costs have a legitimate claim to a voice in how responses are run, and a direct interest in preventing the outbreaks they would otherwise help pay for. That is the main prize.
The risks are real too. Industries with a voice may resist measures that are in the public interest but costly to them. Levies can fall heavily on small producers. Sectors outside the agreement — smallholders, hobby keepers, wildlife — may be neglected. And a formal split can itself become a subject of dispute during an emergency, precisely when disputes are most costly. A well-designed arrangement anticipates these problems. None avoids them entirely.
Five design principles
Drawing these threads together, five principles would help close the gap between who pays for prevention and who benefits.
Make the distribution visible before arguing about it. Most discussions of cost-sharing start with a percentage. They should start with a map: who bears the costs of prevention, who receives the benefits, and who bears the residual risk. Step 6 of the OHS Prevention Ledger is designed to produce that map, including the payer’s own ratio alongside the ratio for society.
Share readiness, not only response. Response is always funded eventually. The prevention paradox bites hardest on readiness — surveillance, laboratory capacity, contingency planning — which is where New Zealand’s model is strongest and where England’s current arrangements are weakest.
Treat compensation as an incentive, and design it as one. The level, conditions and timing of compensation should be set with explicit reference to reporting and biosecurity, and evaluated against them. Full market value may be fair; the evidence suggests it may not be efficient. Linking part of compensation to demonstrated biosecurity standards is one option worth testing.
Pay for the weakest link. Where the value of surveillance depends on its least capable part — a region, a sector, a neighbouring country — the strongest parties have a direct interest in funding it. That applies within the UK, between the devolved administrations, as well as internationally.
Agree it before the emergency. The single most important feature of EADRA and GIA is not the split. It is that the split is agreed in advance. Cost-sharing negotiated during an outbreak is slow, adversarial and damaging to the trust that reporting depends on.
What would prove this wrong?
If standing cost-sharing arrangements like EADRA and GIA produced no measurable improvement in readiness spending, reporting speed or response outcomes compared with taxpayer-funded systems, the claim that aligning payers and beneficiaries changes behaviour would be weakened.
If empirical studies of farmer behaviour found that compensation levels had little effect on either reporting or biosecurity investment, the case for designing compensation as an incentive would be weakened, and fairness would be the only consideration.
And if a proper distributional map of prevention’s costs and benefits showed that the primary payer in fact captured most of the benefit, the payer’s paradox would not be the explanation for underinvestment, and the flagship’s other mechanisms would have to carry more weight.
Conclusion
The prevention paradox is not a failure to understand that prevention is valuable. It is what happens when value is real but scattered, and cost is concentrated. A veterinary service asked to fund surveillance whose benefits flow mostly to others is behaving rationally when it underfunds it; so is a farmer who under-invests in biosecurity that mostly protects the neighbours.
The answer is not to exhort either of them to behave less rationally. It is to change the ledger they are working from: to make the distribution of costs and benefits visible, to ask those who benefit to share the cost of readiness as well as response, and to design compensation as the incentive it already is. Those are institutional choices. They were available to England in 2010. They are still available now.
Questions & Answers
What is the payer’s paradox?
An intervention can be clearly worthwhile for society and clearly unattractive to the body that has to fund it. The benefits of biological prevention are spread across government, farmers, supply chains, health systems, rural economies, insurers and households, while the cost usually falls on one budget.
Why does the gap persist rather than correcting itself?
Three features of biological risk keep it open: prevention is partly a public good, much of its value is an externality that accrues to people who did not pay, and surveillance produces information whose benefit is diffuse. None of those is fixed by better intentions.
Has anyone actually solved this?
Australia and New Zealand have built standing cost-sharing arrangements between government and industry, which is the closest thing to a worked answer. England considered a comparable scheme and deferred it, which is why the question is live here rather than settled.
References
- Animal Health Australia (n.d.) Emergency Animal Disease Response Agreement (EADRA) and A quick guide to compensation and valuation in an EAD response. animalhealthaustralia.com.au/eadra/
- Defra (2010) Draft Animal Health Bill. Cm 7784. London: The Stationery Office. www.gov.uk/government/publications/draft-animal-health-bill
- Defra (2011) ‘Sharing responsibility for animal health and welfare policy’, press release, 26 April 2011. www.gov.uk/government/news/sharing-responsibility-for-animal-health-and-welfare-policy
- Defra and APHA (n.d.) Compensation for Animals Culled to Control Animal Diseases. www.gov.uk/guidance/compensation-for-animals-culled-to-control-animal-diseases
- Fraser, R. (2018) ‘Compensation payments and animal disease: incentivising farmers both to undertake costly on-farm biosecurity and to comply with disease reporting requirements’, Environmental and Resource Economics, 70(3), pp. 617–629. doi.org/10.1007/s10640-016-0102-7
- Gramig, B.M., Horan, R.D. and Wolf, C.A. (2009) ‘Livestock disease indemnity design when moral hazard is followed by adverse selection’, American Journal of Agricultural Economics, 91(3), pp. 627–641. doi.org/10.1111/j.1467-8276.2009.01256.x
- Hirshleifer, J. (1983) ‘From weakest-link to best-shot: the voluntary provision of public goods’, Public Choice, 41(3), pp. 371–386.
- House of Commons Committee of Public Accounts (2025) Resilience to Threats from Animal Disease. HC 885. publications.parliament.uk/pa/cm5901/cmselect/cmpubacc/885/report.html
- House of Lords Library (2022) Bird Flu 2022: Dealing with the UK’s Largest Ever Outbreak. lordslibrary.parliament.uk/bird-flu-2022-dealing-with-the-uks-largest-ever-outbreak/
- Ministry for Primary Industries (n.d.) Government Industry Agreement: Biosecurity. www.mpi.govt.nz/biosecurity/about-biosecurity-in-new-zealand/government-industry-agreement-biosecurity
- Olson, M. (1965) The Logic of Collective Action: Public Goods and the Theory of Groups. Cambridge, MA: Harvard University Press.
- Rose, G. (1981) ‘Strategy of prevention: lessons from cardiovascular disease’, British Medical Journal, 282, pp. 1847–1851.
- Thompson, D., Muriel, P., Russell, D., et al. (2002) ‘Economic costs of the foot and mouth disease outbreak in the United Kingdom in 2001’, Revue Scientifique et Technique (OIE), 21(3), pp. 675–687.
Key Takeaways
- The benefits of biological prevention are spread across government, farmers, supply chains, health systems, rural economies, insurers, households and other countries. The costs usually fall on one budget.
- This produces a payer's paradox: an intervention can be clearly worthwhile for society and clearly unattractive to the body that must fund it. Much underinvestment lives in that gap.
- Three features of biological risk make the gap persistent: prevention is partly a public good, much of its value is an externality, and surveillance is a weakest-link good that is only as strong as its least capable participant.
- Compensation is the hinge on which the private incentives turn. Set too low, it discourages reporting; set at full value, it can discourage biosecurity. The economics literature suggests there is a range that does both jobs, and that full market value may sit outside it.
- Australia and New Zealand have built standing cost-sharing arrangements between government and industry. England considered one, and deferred it in 2011. The case for revisiting it is stronger now than it was then.