Te Rā Whakamana: Co-production With Homeless People

08/07/2025

I look at a Dutch shelter run by its own residents for nearly 30 years and ask what our delivery reforms would need to give up for that kind of expertise to count.

Thank you for indulging me with a week off; I needed it, as we all do sometimes. This post takes up a challenge most reform papers avoid: what happens when implementation expertise shows up in people the system refuses to recognise? Drawing on a Dutch case in which a homeless shelter has been managed by its residents for nearly 30 years, I argue that legitimacy lives in people, in relationships, and in the messy practice of trust, and not only with ministers, agencies, or professionals. That is where delivery happens, and it is why so many of our reforms fail when they keep trying to retrofit control into systems that work only through shared authority.

In my last post I argued that expertise is not confined to ministers, policy advisers, or central government agencies. I proposed that it is alive across the entire advisory system, among Iwi, Hapū, local government, communities, industry, advocates, lobbyists, academics, and individuals and whānau with lived experience, and I suggested that we treat this as a reality check instead of a dysfunction.

Today I want to push that idea a little further by asking what happens when expertise turns up in people the system is not ready to recognise, and what that means for policy in a country where implementation failure is too often written off as a technical glitch when it is better understood as structural exclusion.

To think this through, I turn to a case that challenges our assumptions about who can hold implementation expertise. Hoppen, Brandsen, and Honingh (2025) document a Dutch homeless shelter that its residents have managed, without fuss, for nearly 30 years.

The model is simple: people experiencing homelessness are put in charge of the shelter, which goes well beyond consulting them. Residents manage rosters, decide who gets a bed, enforce rules, resolve conflict, and even determine who stays and who leaves, and there are no staff on site outside office hours. The governance and daily running of the service are held by people whom the formal system often treats as incapable, and the shelter works.

The case deserves our attention for three reasons.

The first is that it confirms what many frontline workers and navigators already know, that implementation expertise lives in people as well as in systems, even though the system often struggles to see it. Lived, relational, and contextual expertise does not appear in spreadsheets, travel easily up hierarchies, or come with credentials. Hoppen and colleagues show that, with the right structural fit, mutual commitment, and sustained support, deeply marginalised and underserved groups can run complex services themselves, which is a much larger role than participating in them or being subject to them.

The second is that the study helps us challenge a persistent and corrosive trope, the claim that homelessness is a “choice”. I dislike that trope intensely, but it is worth taking seriously for a moment. If homelessness is a choice, the obligation on the system is to offer more and better choices, ones that carry consequences, dignity, and the possibility of ownership, and that is what this empirical work shows. When people are given authority over their own choices and environment, many of them (though by no means all) respond with maturity, care, and commitment, enough to suggest that the problem was the system’s refusal to let go of control rather than any lack of willingness.

The third is that the case undoes the myth that vulnerable people cannot be trusted with the machinery of service delivery. Residents manage one another’s safety, make difficult decisions, and hold themselves to account, imperfectly but with care, and often with more care than formal systems designed around rules and sanctions. As I read the study, the model works because of shared responsibility, relational authority, and collective deliberation rather than because anyone gets it right every time.

There is a quiet challenge here for implementation practice in Aotearoa. For this kind of model to work locally, the people who already hold institutional power must step back. In the Dutch case, social workers facilitated without controlling; they helped hold the culture without overriding it or leading its shaping, and they voiced values while leaving decisions to the group. When things went wrong they stayed and kept showing up, to uphold the mana of the system they had helped build rather than to intervene. I would describe that as institutional humility as much as professional discipline, and as the practical state in action.

That humility is missing from too many of our delivery reforms. We talk about outcomes, co-design, and lived experience, but we rarely create the institutional conditions for those things to matter. We still treat trust as a proxy for control, and we still measure success by how closely the delivery system conforms to Pōneke’s expectations.

What the case shows is that meaningful implementation can rest on a different logic, one that assumes expertise is dispersed rather than centralised, that power should be shared rather than hoarded, and that legitimacy is earned at the edge rather than declared at the top.

In Aotearoa, where authority has been plural from the original design, these lessons are basic rather than optional. Legitimacy here is not bestowed from the Cabinet table. It is negotiated between the Crown and Māori, between the state and its citizens, between agencies and providers, between regulators and regulated parties, and between officials and communities.

For reasons that escape me, we keep trying to retrofit these plural realities into delivery models designed for command and control, and I think that is why those models so often fail.* The difficulty lies less in co-production itself than in systems that are still structured to resist it, and anyone who doubts this need only listen again to the complaints about co-governance.

For those of us working in policy, commissioning, or implementation, the question is therefore no longer whether co-production is possible, since it has been happening for decades. The question is whether we are willing to build systems that let it matter.

*Where those systems do not fail, because outcomes are achieved, trust is built, and people feel served, it is usually because a frontline team has made things work despite the system. Street-level bureaucrats, navigators, and care workers hold the line every day; they bend rigid rules, translate policy into something humane, and carry a relational burden the system will not acknowledge. In my view, it is their discretion, courage, and care that keep the whole arrangement upright.

References

Hoppen, M., Brandsen, T., & Honingh, M. (2025). Conditions for involving vulnerable citizens in co-production: The importance of commitment. Public Management Review, 28(7), 1894–1923. https://doi.org/10.1080/14719037.2025.2523953