I have carried this idea for a long time, but it stopped being abstract in the spring of 2020. During France’s first lockdown, within a single week, police interventions for domestic violence rose by roughly a third: 32 percent in gendarmerie zones, 36 percent in the Paris prefecture, according to the interior minister at the time. UN Women named the global surge the “shadow pandemic.” I remember sitting in my own confinement doing the arithmetic that the numbers implied. The same walls that were protecting the rest of us from a virus were locking thousands of people in with the person who hurts them.
Let me state the premise of this article without softening it, because everything else follows from it. Raising a hand against a partner or a child is an act of absolute malice and absolute weakness, and the people who survive it have the right to a new life, a real one, without the total fear of reprisal. That is not what they get today. Today, the state acts on the aggressor only when the violence is known, only when he can be found, and only when the courts can process him, three filters that each leak. And the survivor, meanwhile, is accompanied partially, unevenly, rarely for the long term. My conviction is simple. A far greater share of the effort and the money should be pointed at the person who was harmed.
A word on pronouns, because they will be read. I write he for the aggressor and she for the survivor because that is what the numbers overwhelmingly describe, not because the program has a gender. Victims come in every gender, and so do aggressors. The door described below is open to anyone who has to flee.
Every human being brings something to the world. No one should be held hostage by fear, suffering in silence, convinced they are the problem.
So here is the idea, stated plainly. We know how to protect people whose lives are in danger. We have done it for decades when the endangered person was a witness in a criminal trial. The Haven Program applies the structure of witness protection to survivors of domestic violence: risk assessment, extraction, confidential housing, a dedicated human protector, and months of wraparound support, until the danger is legally contained or the new life stands on its own.
The stakes, measured
This section exists because the danger this program addresses is one of the best-documented facts in violence research, not a feeling. Peer-reviewed work (Campbell et al., American Journal of Public Health, 2003) found that leaving a cohabiting abusive partner roughly tripled the risk of femicide, and that the combination of a highly controlling abuser and a separation multiplied it more than fivefold. The moment a victim finally does what everyone tells her to do, leave, is precisely the moment her risk of being killed peaks. In France, the interior ministry’s annual study counts around a hundred women killed by a partner or ex-partner every year: 118 in 2022, 96 in 2023, 107 in 2024. Roughly one every three days, year after year, many of them in that exact window this program is designed to cover.
What already exists, in fragments
Nothing in this idea starts from zero, and honesty requires the map. France’s Téléphone Grave Danger gives high-risk victims a phone with a direct 24/7 line to responders. More than 6,000 were deployed in 2024, triggering some 3,000 police interventions that year. The anti-rapprochement bracelet, operational since late 2020, GPS-tracks the aggressor and alerts police when he approaches. Over 2,500 had been ordered by end-2023. Judges can evict the violent partner from the family home and reserve it for the victim. Spain has gone furthest. The VioGén system risk-assesses every reported case and assigns graduated police follow-up, with roughly 104,000 cases under active monitoring at the end of 2025. In the United States, about 45 states run Address Confidentiality Programs that give survivors a substitute legal address. And everywhere, shelter networks operate confidential addresses, chronically saturated, and built for emergency, not for the long arc.
Look at that list and you see the shape of what is missing. Every fragment is real and every fragment is partial: a phone but not a protector, a bracelet but only after a court order, a shelter but only for the acute weeks, police follow-up but not a rebuilt life. No one has assembled the fragments into a program, one door that a person in danger walks through, behind which risk assessment, safe housing, human protection, and reconstruction happen as a single continuous service.
The program, concretely
The Haven Program is that assembly. A survivor whose danger is assessed as serious enters the program and receives five things. First, a structured risk assessment, in the spirit of VioGén, repeated over time rather than done once. Second, if the risk justifies it, relocation to secure housing that only she knows, deliberately dispersed, ordinary apartments in ordinary buildings, never clustered into an identifiable target. Third, a named protection officer, one trained human being responsible for her case, who checks in, coordinates police and justice contacts, and is reachable at 3am. Fourth, the reconstruction layer: legal support through the proceedings, psychological care, help with employment, schooling for children, the administrative surgery of a life moved. Fifth, a graduated exit, on her timeline, when the aggressor is contained by justice or the new life is established, with the door remaining open.
The duration is measured in months, sometimes longer. The unit of success is not a closed file but a person who no longer scans the street before leaving home.
The tension this idea must face
There is a moral inversion at the heart of any protection-by-hiding model, and this article will not walk past it. It is the victim who disappears while the aggressor keeps his name, his home, and his freedom. That is upside down, and the modern direction of policy, the bracelet on his ankle, his eviction from the home, is the correct one, because it constrains the person who created the danger. I hold both of these ideas at once, because the data forces me to. Constraining the aggressor is justice, and it is slow, requiring him to be known, found, and processed. Hiding the victim is unjust, and it is fast, and in the window where the femicide risk multiplies, fast is what keeps people alive. The Haven Program is therefore a bridge, never a destination. It protects during exactly the interval when justice has not yet done its work, and its own success metric, the one that should be printed in its annual report, is how short it can make the average stay. A society where this program is large and permanent has failed twice.
Why this is not a startup, and what it is instead
I will not insult this subject with a freemium model. The cost per protected person, housing, staff, security, months of support, is high, and the person being protected can never be the customer. This is a public mission. Realistically, it would be a delegated public service or a state-funded nonprofit, financed the way shelters and witness protection already are, by the state, supplemented by foundations. Call it an enterprise only in one sense. It would employ thousands (social workers, trained protection officers, jurists, psychologists, housing managers), and it would need the operational discipline of a serious company, because its failures would be measured in lives.
Two hard boundaries. The program works with police and justice, never instead of them. Protection officers are not vigilantes, not armed escorts, and the legal limits of private protection must be respected to the letter, or the program becomes a militia with a mission statement. And it must be built with the organizations who have carried this fight for decades, in governance and not as decoration. Years ago, my first sketch of this idea was an app. The more I learned, the clearer it became that an app on a phone the aggressor may inspect is not safety, and the people who taught the world that lesson are the specialized associations. They hold the expertise this program would run on.
Closing
The Haven Program is witness protection, extended to the people who need it most and get it least. It is built as a bridge across the deadliest months of a survivor’s life, pointed at a simple outcome, a new life, without fear, for people whose only mistake was trusting someone who was weak enough to use violence on those he claimed to love.
This is not a weekend build, and for once I will not end with a playful invitation. If you hold power over budgets or policy, in a ministry, a foundation, an NGO, this idea is yours, free, like everything on this site, and I would gladly help you think it through. And if you are living what this article describes: in France, call 3919; in immediate danger anywhere in Europe, 112. You are not the problem. You never were.
