The national ledger for no net loss of wetlands has a debit column and a credit column. In the 1990s the credit column as required in permits exceeded the debit column. Nobody keeps this ledger in one place. The marsh that used to hold your county's spring runoff is in neither column. It is a cul-de-sac now, and somewhere in the same watershed there is a compensating acre that balanced it.
Both of those things are true at once. If you typed no net loss of wetlands, you are probably a planner reading a permit file, a landowner with a wet corner and a question, or a project lead whose site plan just met the Clean Water Act. Here is what the rule means, what it has genuinely done, and the one thing it was never built to do.
what no net loss of wetlands means
No net loss of wetlands is the national goal that wetland losses be balanced by wetland gains, so the country's total wetland base, measured by acreage and function, does not shrink. It came out of the National Wetlands Policy Forum, convened in 1987 at the request of the Environmental Protection Agency and chaired by Governor Thomas Kean of New Jersey. The forum's November 15, 1988 report — convened by the Conservation Foundation at EPA's request — called for "no overall net loss of the nation's remaining wetlands base, as defined by acreage and function." President George H. W. Bush adopted it as national policy in 1989. Every administration since has kept it.
In the regulator's words, the goal runs through a three-step sequence. Avoid the wetland impact if you can. Minimize what you cannot avoid. Compensate for what is left by restoring, establishing, enhancing, or preserving wetland somewhere else. That third step is compensatory mitigation, and it is where credits, banks, and in-lieu fee programs live.
Section 404 of the Clean Water Act is the cousin that carries the sequence: discharging fill into a jurisdictional wetland needs a permit from the U.S. Army Corps of Engineers, issued under guidelines written by EPA, and the permit conditions are where avoid, minimize, compensate gets applied to one site. The bank that sells the compensating credit is a land deal with a register attached, and it already has its own post: a habitat bank is a land deal.
No net loss of wetlands is a national balance between wetland acres lost and wetland acres gained over time. It is not a promise that any particular wetland stays wet.
what the rule actually did
Give it its due, because a lot of people want to skip this part.
In June 1989 the President told Ducks Unlimited that the country was losing wetlands at "nearly half a million acres a year." The most recent U.S. Fish and Wildlife Service status and trends report puts net loss between 2009 and 2019 at about 221,000 acres across all wetland categories — a different method from the 1989 estimate, and not the same as the 670,000 acres of vegetated wetland lost in that window (partly offset by pond gain). USFWS reports both numbers; they do not net cleanly. That is not all the rule's doing. The 1985 farm bill's Swampbuster provision, the Wetlands Reserve Program, state wetland programs, and the plain fact that the cheap acres had already been drained all pulled the same direction. But no net loss turned a set of scattered programs into one target, and the 1990 Corps–EPA mitigation agreement that followed it made avoid, minimize, compensate the explicit order of operations for every Section 404 permit.
The compensate step built something real too. The 2008 Compensatory Mitigation Rule requires a recorded site protection instrument, performance standards, financial assurances, and a monitoring period of at least five years, with most credits gated on performance rather than paperwork. Some mitigation banks and in-lieu fee programs plug ditches, pull tile, remove berms, and reconnect a river to its floodplain under an easement that outlives the sponsor. When that happens, a wet acre exists that would not have otherwise. The people doing that work are not the problem in this story.
where the acre goes missing
The National Research Council studied the program in 2001 and wrote a finding that holds both halves in one line: "The goal of no net loss of wetlands is not being met for wetland functions by the mitigation program, despite progress in the last 20 years." On paper the ledger looked good. From 1993 to 2000, roughly 24,000 acres a year were permitted for fill and roughly 42,000 acres a year were required as compensation, a 1.8-to-1 ratio that reads as net gain — and the committee noted that preserved and enhanced acres were counted as equivalent to restored or created ones. The available records also could not show how much of that required compensation had been built, or was functioning. The 2008 rule tightened much of what the committee flagged. Three gaps remain, and none of them is fraud. They are what a national balance cannot see.
| the acre | what the ledger records | what the floodplain does next spring | who finds out |
|---|---|---|---|
| avoided — the project went around the marsh | Nothing. No permit, no debit, no credit | Fills, slows the water, drains over days. Same as last year | Nobody. That is the point |
| filled, with a credit purchased | A debit and a matching credit; on paper the base is whole or better | The fill sheds water in minutes. The credit acre, if built and mature, holds water somewhere in the service area | Whoever is downhill of the fill, in the years before the credit acre matures |
| wet floodplain, held as itself | Nothing, unless it happens to be the credit acre | Stores the flood, filters the water, feeds the aquifer, in this basin, this year | Everyone downstream, mostly by never noticing |
Type. The 2009 to 2019 numbers net out to 221,000 acres lost, but vegetated wetlands — marsh, shrub swamp, forested swamp — lost about 670,000 acres, and forested wetlands alone lost 426,000, while non-vegetated wetlands such as ponds gained about 488,000. A pond is a wetland on the ledger. It is not a bottomland hardwood swamp, and the Fish and Wildlife Service says plainly that the substitution "alters wetland function," including storm buffering and water quality. The balance is real. The function it balances is not the same function.
Time. Fill takes a season. The NRC found that restored and created wetlands often do not reach functional equivalence with reference sites within five years and can take up to twenty for some attributes. A credit can release on construction milestones while the swamp it replaces is already a slab, so the gap is measured in years, and the water does not wait for it. The 2008 rule's five-year monitoring floor narrows that gap. It does not close it.
Place. A credit is valid anywhere in the bank's service area, which can be an entire watershed. Flood storage is not portable. A restored floodplain forty miles upstream is a good thing, and it does not hold water for the subdivision built where the marsh was. Since Sackett v. EPA in 2023, federal jurisdiction also reaches only wetlands with a continuous surface connection to other covered waters, so many prairie potholes, playas, and other geographically isolated wetlands sit outside Clean Water Act jurisdiction. Isolated waters largely left the federal ledger with SWANCC in 2001; Sackett narrowed it further. Swampbuster still reaches many farmed potholes, and many states regulate what the federal ledger drops. Not counted federally is not the same as not lost — and it is not the same as unregulated.
A pond dug to compensate for a filled forested swamp balances the acreage ledger. It does not replace the swamp's flood storage, and it does not put that storage back where the flood is.
Fens and bogs are the extreme case. The NRC recommended straight avoidance for them because no ledger can replace peat on a human timeline. That is a different wet stock, with its own series.
nobody in this chain is the villain
Walk the file. The farmer who tiled the field in 1974 was doing what the county extension office recommended and what the Swamp Land Acts had encouraged since 1849. The developer who bought credits followed the rule exactly, and paid for it. The Corps district applied the sequence, found the impact unavoidable, and required compensation at ratio. The bank sponsor restored real acres under a real easement. The county approved a plat that matched the flood map.
Every decision in that chain was defensible, and the marsh that held the water for that neighborhood is gone. That is not a scandal. It is what happens when a national accounting goal is asked to do a local hydrological job. No net loss was built to hold a number steady across a country. It was not built to hold one floodplain wet, and it is not fair to the rule, or to the people who run it, to pretend otherwise.
keep the rule. hold the acre
Here is the honest position. Keep no net loss. It slowed the bleeding, it forced avoidance, and it created a funding channel for restoration that did not exist before. Then stop expecting it to keep the floodplain in your basin wet, because that is a different object and it needs a different hold.
The people paying for the missing acre already know this, even if they have not named it. Flood insurance claims, storm sewer upgrades, treatment plant capital, drainage district assessments, the buyout program that arrives ten years late: those are receipts for wetland function that left the basin. You don't have a flood problem. You have a missing wetland walks the claims data.
What holds a specific acre wet is a position in that acre's present condition, funded now rather than after the permit, priced on what the floodplain is doing this year, and routed to whoever keeps it doing it. That is what ensurance is for. A certificate is issued against one named wetland agent; buying it routes proceeds to that agent's account. Coins fund the wet condition more broadly. Neither is a credit, neither settles a permit, and neither pretends to be the wetland. The marsh exists whether or not anyone buys anything. What changes is who is paying to keep it a marsh.
Our own stage here is small and real. inland-wetlands.ensurance is the stock agent. marble-wetlands-preserve.avlt is a named 54-acre beaver wetland on the Crystal River in Colorado. A handful of place coins carry specific wetlands. Volumes are modest. The mechanism is live.
| no net loss | a wet floodplain, held | |
|---|---|---|
| what it is | A national goal and a permit sequence | A living acre, in this basin, this year |
| unit | Acres, netted across the country over time | The condition of one named place |
| when it acts | At the permit, after an impact is proposed | Now, before anyone applies to fill anything |
| what it commits to | The national base does not shrink on paper | This floodplain is funded now — not guaranteed wet |
| what it cannot do | Keep a specific marsh wet | Balance the national ledger |
Same reader, two jobs, two instruments. The rule is the floor. The acre is the point.
read next
- a habitat bank is a land deal — what the compensating credit is when you look at it as real property
- a wetland credit is not a wetland — credit versus stock, for the people buying and selling them
- what wetlands actually are — the pillar: the floodplain that still works
- explore the live agents and coins — see which wetland places are funded today