The cheapest way to save a predator is a better fence

There is a version of conservation fundraising in which the threat is always a criminal with a rifle. It exists because it is easy to narrate and easy to feel strongly about. In many inhabited landscapes it is also a minority of the problem.

The larger and quieter pattern is this: a household loses goats overnight, the loss represents a meaningful share of their assets, and the household responds in the way that reliably prevents a repeat. Sometimes that is a spear, sometimes a snare, and increasingly often it is poison — which kills indiscriminately and takes vultures, hyenas and scavenging birds along with the intended animal.

The household economics that drive the outcome

To understand why prevention works better than enforcement here, it helps to look at the decision from inside the household rather than from the perspective of a conservation budget.

• Livestock is often the household’s savings account, insurance policy and liquid capital simultaneously. Losing several animals is not an inconvenience; it can be the difference between paying school fees and not.
• The loss is concentrated and immediate; the benefit of wildlife presence is diffuse and usually captured by others — tour operators, distant governments, international donors.
• Enforcement raises the cost of killing but does nothing about the loss, which means the pressure remains and the response becomes more covert.
• Poison is cheap, silent, requires no confrontation and is therefore the rational choice once enforcement makes visible methods risky. Enforcement without prevention can shift the method towards the one with the worst collateral effects.
• Prevention changes the underlying arithmetic: if livestock does not die, there is no loss to retaliate against.

The interventions, and what determines whether they work

The pattern across that table is that the cheap physical interventions have maintenance failure modes and the expensive administrative interventions have institutional failure modes. Maintenance is a solvable problem with a small recurring budget. Institutional failure usually is not.

Why compensation schemes disappoint so reliably

Paying for verified livestock losses is intuitively fair and is the mechanism most often proposed by people encountering the problem for the first time. It has a consistent record of underdelivering, for reasons that are structural rather than local.

1. Verification requires someone to visit, identify the predator from the carcass, and do so before scavengers remove the evidence. In practice this means a subset of genuine losses is never verified.
2. Payment delay destroys the mechanism. A household compensated four months later has already made its decision about the predator.
3. Partial valuation feels like insult. Paying market value for an animal that was also breeding stock and social status does not restore the household position.
4. Funds run out. Compensation is an open-ended liability funded by a finite grant, so schemes tend to pay well in year one and poorly in year three, which is when trust breaks.
5. Incentives can invert. Where verification is weak, compensation can reward carelessness with herding; where it is strong, the administrative cost approaches the payout.

None of this makes compensation useless — as a supplement to prevention, for losses that occur despite good practice, it can maintain tolerance. As the primary mechanism it substitutes a permanent administrative obligation for a one-off physical fix.

Crop raiding is a different problem

Predation and crop raiding are routinely bundled together as human–wildlife conflict, and the interventions do not transfer. Losing a season’s maize to elephants is an income event rather than an asset event, it happens on a predictable seasonal schedule, and the animal involved is large enough that deterrence carries real physical risk to the person attempting it.

The toolkit is correspondingly different: barrier crops the animals dislike, beehive fences that exploit a genuine aversion, timed harvesting, communal watch systems, and land-use planning that avoids placing farmland across a known movement route. The last of these is the most effective and the least fundable, because it is a planning decision rather than a purchasable object.

Assessing a coexistence programme

1. Ask what the baseline loss rate wasWithout a before figure, a reported reduction is unverifiable. Good programmes record livestock losses per household before building anything.
2. Ask who maintains the structuresThe most common quiet failure is an enclosure built by a project and never repaired after it left. Maintenance responsibility and funding should be explicit.
3. Ask how households were selectedPrevention concentrated on the highest-loss households produces most of the benefit. Even distribution is politically easier and less effective.
4. Look for retaliatory-killing data, not just predator countsThe outcome the programme claims to change is retaliation. If it is not measured, the programme is reporting activity.
5. Ask what the community gets that is not conditional on wildlifeProgrammes whose only benefit stream depends on tourism collapse when tourism does, and the collapse arrives at the worst moment.

WildCare Trust funds endangered wildlife rescue and animal welfare work rather than a landscape-scale institution. We describe what a donation buys as a defined input — materials for enclosure reinforcement, veterinary supplies, feed for a stated period — and publish a wallet per campaign with raised and spent totals, because an input we can document is worth more than an outcome we would have to assert.