When 'no net loss' actually works: a study tries to pin down biodiversity offsetting

On 17 July 2026 a team of researchers published a study that does something the biodiversity offsetting industry has long avoided: it defines, with specifics, when a destroyed habitat can be replaced with something ecologically equivalent, and when it cannot. The paper, summarised by Phys.org the same day, isolates the conditions under which so-called biodiversity net gain projects actually deliver what they promise on paper, and the conditions under which they quietly fail.

For more than a decade, governments and developers have treated biodiversity offsetting as a kind of accounting trick with a green veneer. A road goes through a wetland; a developer pays into a fund; a different wetland, somewhere else, is supposedly restored or created to make up for it. The arithmetic is rarely contested. The biology almost always is. Offset sites are typically smaller, more fragmented, younger, and less ecologically connected than the habitats they are meant to replace, and most monitoring regimes stop measuring long before the ecosystems in question have had a chance to mature.

The new analysis, drawing on a large synthesis of offset case studies, argues that the question is no longer whether offsetting can work in principle. It can. The question is what has to be true on the ground for it to work, and most current programmes fail several of those conditions simultaneously.

What 'no net loss' actually requires

The study's central finding is uncomfortable for ministries that have built planning law around offsetting as a default mitigation tool. Successful long-term biodiversity net gain, the authors write, requires offset sites that match the destroyed habitat in type, condition, and ecological context, not just in nominal area. A restored grassland on a degraded former arable field is not, ecologically, equivalent to species-rich chalk grassland lost to a housing estate, regardless of how many hectares the ledger records.

The paper also pushes back against the standard twenty-five to thirty-year monitoring window that most regulators treat as adequate. Many of the ecological functions that justify offsetting in the first place, soil carbon accumulation, specialist species colonisation, hydrological reconnection, take longer than three decades to fully express. Schemes that close their books at year thirty are, in many cases, declaring victory before the recovery has actually happened.

Connectivity emerges as the decisive variable. Offset sites embedded in larger habitat networks, adjacent to surviving patches of the same ecosystem, with dispersal corridors for the species the offset is supposed to support, perform markedly better than isolated sites of nominally equivalent area. That sounds obvious. In practice, it means the cheapest offset land, usually marginal fields with no ecological neighbours, is also the most likely to fail.

Why governments leaned on offsetting in the first place

The political attraction is easy to state. Developers want planning certainty. Conservation groups want binding commitments that cannot be quietly dropped. Treasuries want a mechanism that does not require additional public spending for every hectare lost. Offsetting, in theory, reconciles all three. A developer who cannot avoid damaging a site pays a calibrated sum; that sum funds habitat elsewhere; the ledger, in principle, balances.

The trouble is that the ledger is almost always drawn up by the party with the most interest in making the numbers come out even. The study does not name any particular jurisdiction, but its findings line up with documented weaknesses in the United Kingdom's Biodiversity Net Gain regime, in the European Union's Habitats Directive compensation practice, and in Australian state-level offset schemes that have been audited and found wanting in successive reviews.

There is also a structural problem the paper only partly addresses. Offsetting assumes that ecosystems are interchangeable at the margin, that one hectare of functioning wetland is roughly equal to another. Ecologists have argued for years that this is not how nature works: habitats are place-specific, species are site-faithful, and the loss of a particular patch can break a population dynamic that no amount of distant restoration will repair. The new synthesis gives that argument empirical weight.

The counter-argument the industry will make

Offsetting's defenders, and there are serious ones inside conservation NGOs as well as in the development sector, will point out that the alternative to offsetting is often no mitigation at all. If a regulator tells a developer that there is no acceptable substitute for a damaged site, the developer may simply walk away, or push for a political override. The result is unmitigated loss rather than imperfect replacement.

That defence holds where the alternative is genuinely binary. It holds less well where offsetting has become a planning convenience, a way for authorities to approve destruction they would otherwise have to refuse. The study's findings suggest that the sharpest improvements come not from abandoning offsetting but from tightening its preconditions: matching habitat type, requiring connectivity, extending monitoring to ecologically meaningful timescales, and publishing the data so that failures cannot be quietly buried.

What changes if the findings are taken seriously

If governments accept the study's framing, three shifts become likely. Planning regimes will need to score potential offset sites on ecological context, not on cost per hectare. Monitoring obligations will extend well past the standard quarter-century horizon. And the offset market, currently structured around cheap, distant, marginal land, will need to reprice toward more ecologically defensible locations, with the predictable consequence that offset costs for developers will rise.

That last effect is the politically awkward one. Higher offset costs feed directly into housing and infrastructure budgets, and ministries under fiscal pressure will resist them. The study does not resolve that tension, and it does not pretend to. What it does is make the trade-off legible. Governments that want offsetting to function as advertised will have to pay for it as advertised. Governments that are not willing to pay will have to admit, in public, that the green arithmetic on their planning ledgers does not actually balance.

What remains uncertain

The synthesis leans heavily on case studies from temperate, well-instrumented ecosystems, particularly in Western Europe and parts of North America. Whether the conditions it identifies translate to tropical, arid, or marine offsets is less clear; the underlying biology is similar, but the institutional capacity to monitor and enforce is not. The authors flag this as a limitation. Practitioners in lower-capacity settings should treat the headline thresholds as provisional rather than as portable standards.

There is also the question of whether political systems built around offsetting as a planning convenience will accept the discipline the study implies. Past experience, from forestry carbon credits to wetland mitigation banking in the United States, suggests that rigorous standards tend to erode under industry pressure once offsetting becomes a routine revenue stream. The paper cannot prevent that. It can, however, give regulators a defensible empirical baseline the next time a developer offers to make the books balance for less than the ecology actually costs.

Desk note: Monexus treated this as a policy story, not a pure ecology piece. The wire lede focused on the conservation biology. The story worth reading is the one about which planning ministries are willing to enforce the conditions the study identifies, and which will quietly let the cheaper, weaker version keep operating under the same label.