The industry — when "solving" addiction is a salary, the incentives don't align with solving it

The thesis in one paragraph

The interventions that work — opioid agonist therapy in primary care, structural environmental changes, decriminalisation paired with funded treatment, GLP-1 prescribing at scale, contingency management for stimulants — are under-deployed by an order of magnitude. The interventions that do not work — abstinence-only residential rehab, classroom drug-education, faith-based recovery as the only mode of treatment, unaccountable housing-first programmes — are well-funded, institutionally entrenched, and politically defended by constituencies whose income, identity, or ideological commitments depend on them continuing. This is the addiction industrial complex. It is not a conspiracy. It is an equilibrium. The equilibrium is the reason the lever stays untaken.

Shellenberger — the steelman

Michael Shellenberger published San Fransicko: Why Progressives Ruin Cities in October 2021. The book is the most-cited recent diagnostic of US West Coast harm-reduction failure modes. The position is anti-progressive and politically aligned with the US centre-right, but the empirical observations on which the diagnostic rests are mostly correct and worth engaging with at full strength before they are answered.

The Shellenberger thesis, voiced at full strength:

San Francisco, Los Angeles, Portland, Seattle and a handful of other West Coast cities have run the largest natural experiment in harm-reduction-without-accountability that any developed country has attempted. The results are visible on the street, and they are catastrophic. Open drug scenes in the Tenderloin and Skid Row. Tens of thousands of people living and dying in unmanaged encampments. Overdose deaths concentrated in the populations the policy was supposed to help. Theft, public defecation, untreated psychosis, and a steady out-migration of the working population that pays the taxes that fund the system. None of this is denied by anyone with eyes.

The progressive ideology that produced this is not stupid or malicious. It is well-intentioned and theoretically coherent. It holds that addiction is a disease, that disease is not a moral failure, that compulsion violates the rights of the addicted, that housing must come first and without sobriety conditions, and that the punitive frame is itself the cause of much of the harm. Each of these premises is defensible in isolation. The combination, applied without correction at the city scale, produces the visible street outcomes I have just described.

The institutional structure that maintains the ideology is what I call the addiction industrial complex. It has several parts. The treatment-industry side: residential programmes funded by Medicaid and county contracts that produce poor outcomes but persist because the funding model rewards bed-nights, not recoveries. The advocacy side: NGOs whose grant pipeline depends on the problem continuing and on the rejection of accountability frames. The political class: city councils and prosecutors whose careers depend on signalling progressive virtue and who cannot publicly endorse compelled treatment without losing their primaries. The patient side: a population whose addiction is being maintained, not treated, by the absence of structured intervention.

The countries that have solved or substantially reduced street addiction — the Netherlands, Switzerland, Portugal in its funded decade — combined harm reduction with accountability. Open drug scenes were tolerated only inside supervised consumption rooms. Housing came with treatment expectations. Drug courts combined sanction and treatment. Refusing treatment had consequences. The US progressive model removed the accountability half and kept the harm-reduction half, and the outcomes show what happens when you do that.

The honest progressive answer is to restore the accountability layer. Mandatory drug-court routing for street users. Compelled residential treatment for the floridly psychotic and the seriously addicted who cannot consent functionally. Housing-first with sobriety-progress conditions for continued residency. A return to the European model that the progressive movement claims to admire but operationally opposes. Until that happens, the West Coast cities will continue to be a humanitarian disaster operated under a humanitarian slogan, and the rest of the country will continue to draw the lesson that harm reduction itself is the problem.

Where Shellenberger is right

Four things, load-bearing.

One: the West Coast street outcomes are genuinely terrible. San Francisco overdose deaths rose roughly 4× between 2018 and 2023. The Tenderloin and South of Market open drug scenes are not the harm-reduction-success picture the policy promised. Los Angeles County's homeless population grew through every cycle of policy intervention. Public-data measures — overdose, encampment counts, untreated mental health prevalence, retail theft — have all moved in the wrong direction across the West Coast progressive cities during the period the progressive harm-reduction model has been operationally dominant. The denial of this by some advocacy voices is itself a sign that the framework is fragile.

Two: the US progressive harm-reduction model is not the European model. Shellenberger is right to draw the distinction. The Swiss HAT programme runs in supervised consumption rooms, with case management, with re-engagement pressure on patients who disengage. The Dutch coffee-shop model is licensed, registered, and conditional. Portugal at its peak had Commissions for the Dissuasion of Drug Addiction that combined a fine, a warning, or a treatment referral — that is, a structured accountability layer alongside the decriminalisation. The US progressive model removed the structured accountability and kept the substance tolerance. That is a real ideological choice, with real consequences, and Shellenberger is not wrong to name it.

Three: the institutional incentives in US addiction services are misaligned. County-contracted residential treatment funded by bed-night billing produces an institutional incentive to maintain beds, not to produce outcomes. Medicaid reimbursement for the 28-day model has kept the model alive long past its evidence base. NGO grant cycles depend on the visible problem continuing. Shellenberger names this honestly. The structure is real even where his political conclusions from it are contestable.

Four: the accountability frame is empirically defensible. Contingency management — see the Treatments page — produces measurable behaviour change through structured incentives, which is a softer version of the accountability frame Shellenberger calls for. Drug courts in multiple US evaluations produce better outcomes than treatment-only or sanction-only arms. The "no consequences" version of harm reduction is not what the European countries he cites are actually doing. The accountability layer is a real thing and the evidence base is real.

Where Shellenberger is wrong

Four things, equally load-bearing.

One: he generalises from the West Coast US to the entire harm-reduction project. The book's title generalises further than the evidence supports. Portugal in its funded decade was a harm-reduction success on every measurable axis. France's primary-care buprenorphine rollout (see Countries § France) was a harm-reduction success and it had no compulsion layer. The Estonian and Ukrainian OST programmes are harm-reduction successes operating under conditions far worse than San Francisco's. The honest claim is narrower: US-style progressive harm-reduction-without-accountability has failed in specific US cities. Generalising from that to "progressives ruin cities" or "harm reduction does not work" is a political move dressed as an empirical one.

Two: he under-weights housing supply as a confounder. San Francisco's homelessness crisis is partially a function of the most restrictive housing-supply regime in any major US city — single-family zoning over most of the residential map, decades of NIMBY litigation against multifamily construction, the median single-family home price above $1.5 million. Tokyo, Vienna, Helsinki, and Singapore all run substantially more permissive harm-reduction regimes on substances and produce nothing like SF's street outcomes, because the cost of housing in those cities makes "couch-surf with a friend" a viable option that SF eliminated. The structural housing failure is a larger contributor to the visible street picture than the addiction policy is. Shellenberger's framework does not engage with this seriously.

Three: his preferred compelled-treatment alternative has its own mixed evidence base. Coercive residential treatment in the US has been evaluated and the results are not what the rhetoric suggests. Drug courts work for the population that is mandated into treatment that they would not have chosen — but they also raise serious due-process concerns and underperform voluntary OST on mortality endpoints for the comparable populations. The Cochrane review of compulsory drug treatment finds no consistent evidence of effectiveness ([Werb et al., International Journal of Drug Policy, 2016](https://pubmed.ncbi.nlm.nih.gov/26790691/)). The case for compulsion is not as evidentially clean as the case for voluntary access to OST in primary care. Shellenberger treats the two as comparable; the evidence base does not.

Four: he conflates the rehab industry with the harm-reduction NGOs. The 28-day Minnesota Model residential industry — which is the most evidence-light, most lucrative, and most institutionally entrenched part of US addiction services — is not a progressive harm-reduction industry. It is a commercial abstinence-only industry that pre-dates the progressive ideology by decades and that the progressive harm-reduction movement has, on balance, criticised. Shellenberger's "addiction industrial complex" framing rolls these together; in the specific budget and political-coalition data, they are separate constituencies with different interests and different failure modes. The conflation makes the rhetoric stronger and the analysis weaker.

Narconon — the cleanest cautionary case

Where Shellenberger's industrial-complex thesis is most defensible, with the cleanest evidence base, is in cases of actual institutional capture by parties whose stated purpose ("anti-drug education") is a cover for a different actual purpose. The Scientology-affiliated Narconon and Foundation for a Drug-Free World programmes are the textbook case.

The structure: Foundation for a Drug-Free World (FDFW) and Narconon are operated by the Church of Scientology. Both present themselves as secular anti-drug-education and rehabilitation programmes. The Scientology affiliation is systematically concealed from the schools, parents, and students they engage with ([Wikipedia FDFW](https://en.wikipedia.org/wiki/Foundation_for_a_Drug-Free_World); [Wikipedia Narconon](https://en.wikipedia.org/wiki/Narconon)).

The history. A 1998 Boston Herald series exposed Narconon and the World Literacy Crusade using anti-drug and learn-to-read programmes to gain public-school access without disclosing their origins. In 2005, the California Department of Education commissioned a review by fourteen independent substance-use and health-education experts; the review found the programme inaccurately stated that "drugs are stored in fat and later released", a pharmacological claim with no scientific basis but considerable load-bearing weight in Scientology's "purification" theology. Santa Monica High School cancelled the programme after parents discovered the affiliation. The Hollywood Reporter's investigation documented the same pattern at scale: undisclosed religious recruitment routed through ostensibly secular school programmes. Narconon's former president publicly confirmed the recruitment angle: getting children "anti-medicine, anti-doctors, anti-psychiatrists" was part of the curriculum design.

The McGill Office for Science and Society describes the operation in clean terms: a religious organisation with a doctrinal commitment against psychiatry and pharmaceutical medicine uses anti-drug education as a Trojan horse to reach children whose families and schools would not consent to the underlying programme if its origin were disclosed. The "education" content is fabricated pharmacology. The downstream funnel is into Narconon's residential programmes, which are themselves a Scientology recruitment vector and have produced multiple deaths in care over the years (litigation and regulatory action documented at the Wikipedia case index).

The case matters for this page because it is the cleanest possible illustration of the broader pattern. The institutional surface of "anti-drug education" is being used as a delivery mechanism for an interest entirely disconnected from reducing addiction. Narconon is the extreme case; the milder versions — rehab chains that bill for unproven services, advocacy NGOs whose funding depends on the perpetuation of the constituency they advocate for, public-health bureaucracies whose authority is preserved by maintaining the diagnostic-perimeter expansion — all run the same logic at lower amplitude.

The 28-day rehab industry — insurance reimbursement as the binding constraint

The dominant US treatment-industry model is the 28-day residential abstinence-only programme, often built on the "Minnesota Model" first developed at Hazelden in the 1950s. The model is so dominant in US discussion of addiction that for many Americans "treatment" and "a 28-day inpatient programme" are synonymous. They should not be.

The evidence base for the 28-day model is weak. The Recovery Research Institute review is candid about the gap between market dominance and outcome evidence. Outcomes are mixed-to-poor relative to the costs ($30,000–$80,000 for a 28-day stay in the US market). Relapse rates after discharge are high. Post-discharge mortality during the tolerance-drop window is a documented phenomenon — patients leave the programme drug-free, attempt to use at pre-treatment doses they no longer tolerate, and overdose. The model excluded opioid agonist therapy for decades on the doctrinal grounds that "replacing one drug with another isn't recovery", a position with no evidentiary support and a substantial body count.

The reason the model persists is structural: US insurance preferentially reimburses it. Medicaid and many private insurers will pay for residential treatment but not for outpatient OST at the same intensity. The reimbursement architecture was set in the 1980s when the Minnesota Model was the dominant clinical paradigm; it has not been re-set since. The industry that bills against the reimbursement codes has organised politically to defend those codes. The codes therefore determine the treatment people receive, not the evidence base.

This is the Shellenberger "addiction industrial complex" critique in its strongest form. The 28-day rehab industry is not a progressive harm-reduction industry. It is a commercial abstinence-only industry that pre-dates the progressive movement, is institutionally entrenched, has a worse evidence base than the OST alternative, and continues to dominate the US treatment landscape because the billing architecture rewards it. The political coalition that defends it includes commercial operators, the recovery-industry advocacy structures, families of patients (whose lived experience of the programmes is often positive even where the outcome data is poor), and the insurance bureaucracies that have built their utilization-management around the model.

The fix is the same fix the Lever page calls for: shift reimbursement to OST in primary care, recognise contingency management as a reimbursable behavioural intervention, and de-fund the residential model where its outcome data does not justify the bed-night. The political coalition to do this does not exist in the US. It does exist in France, the Netherlands, and most of Northern Europe — which is why the European OST outcomes are an order of magnitude better than the US's.

The progressive failure mode, named

Returning to the steelman with the teardown applied. The narrower, defensible Shellenberger claim is this: some specific US progressive cities have, since around 2015, run a version of harm reduction that omitted the accountability layer that European harm-reduction systems include, and the outcomes have been bad enough to deserve naming. The cause is not "progressives ruin cities" — it is a particular ideological move within US progressive drug-policy circles to treat any structured pressure on the addicted patient as a violation of rights, which omits exactly the structural lever the European systems use.

The defensible response is not Shellenberger's full-spectrum compulsion programme. It is the European synthesis: harm reduction plus structured re-engagement, supervised consumption inside designated facilities not on the open street, housing-first with case-management expectations, drug-court routing as an option not a default, OST scaled in primary care, decriminalisation paired with the funded treatment side. This is what worked in Portugal, France, Switzerland, the Netherlands. The US progressive movement has imported half of it. The other half is what this page is naming.

The political problem is that naming this puts the speaker into the Shellenberger camp from the left's perspective and into the harm-reduction camp from the right's perspective. The fair-investigation register of this microsite is the wrong tribal alignment for either side. Which is presumably why the synthesis is operationally rare even where the evidence base for it is the strongest in the field.

When solving addiction is a salary

The unifying pattern across Narconon, the 28-day rehab industry, the West Coast progressive failure mode, and the broader political economy of addiction policy is the same. Whenever an institution's revenue, identity, or political coalition depends on the addiction problem continuing in roughly its current shape, that institution will produce interventions that maintain the problem rather than solve it. The institution will not see itself this way. The individual practitioners will be sincere, competent, often well-intentioned. The systemic outcome will be the equilibrium.

Five examples of the pattern at different amplitudes:

1. The Narconon / FDFW case. Institutional revenue (Scientology recruitment) requires the addiction "education" mechanism to continue. Outcomes for the addicted are secondary at best. Most extreme case on the page.
2. The 28-day Minnesota Model rehab industry. Commercial revenue requires bed-night billing under the existing reimbursement codes. Outcomes are mixed but the institutional pressure is to maintain the model, not to migrate to the evidence-based alternative.
3. The US harm-reduction NGOs operating without accountability layers. Grant funding and political coalition depend on positioning that rejects compulsion. The European synthesis is operationally rare even where it is empirically superior.
4. The tobacco-tax-dependent fiscal architecture. The New Zealand smokefree-generation repeal (see Countries § NZ) is the clean case. State revenue depends on tobacco duty; structural elimination of the consumer base threatens the revenue; the political coalition aligns to preserve the revenue and reverses the structural intervention.
5. The professional addiction-treatment guild structure. Specialist gatekeeping of pharmacotherapy access — addictologists' protected scope of practice for buprenorphine, methadone clinics' federal regulation, the long resistance to GP-level prescribing in the US — is defended by the professional bodies whose members benefit from the gatekeeping. France 1995 broke this; most countries have not.

The point is not that any of these are evil. The point is that the equilibrium is structurally robust against evidence, because the evidence-based alternatives would put the existing institutional structures out of business or out of their current scope. The willpower frame on the Steelman page persists at the public-discourse level for related reasons: it provides a politically convenient diagnosis that justifies the current institutional architecture and absolves it of the failure to produce outcomes.

The honest position

What survives from the Shellenberger steelman after the teardown is sharper and more defensible than the original.

1. The West Coast US progressive harm-reduction model is failing on outcomes — overdose deaths, encampments, untreated psychosis, the cleanly measurable street picture. The European harm-reduction model is not. The difference is the accountability layer.
2. The accountability layer is recoverable without abandoning the harm-reduction frame — supervised consumption inside designated facilities, drug-court routing as an option, housing-first with structured re-engagement, OST scaled in primary care.
3. Multiple US addiction-services constituencies have revenue or political dependencies on the current architecture continuing, and the equilibrium is robust against evidence-based reform precisely because of that dependency structure. Naming the dependency structure is the prerequisite for changing it.
4. The clean cautionary cases — Narconon at the religious extreme, the 28-day Minnesota Model at the commercial extreme — are the most visible illustrations of the broader pattern. The pattern operates at lower amplitude elsewhere and is more politically uncomfortable to name because the relevant constituencies are mainstream rather than fringe.
5. The fair-investigation framing of this microsite does not align with Shellenberger's political camp and does not align with the US progressive harm-reduction movement either. The synthesis position — European harm reduction with restored accountability, OST in primary care, structural environmental intervention, evidence-based reimbursement reform — is empirically defensible and politically homeless.

The Policy page picks up from here with the broader pattern of how policy designed on the willpower frame produces forty years of failed interventions. The Industry page argues that the failure is not only ideological — it is also institutional, and the institutional layer is the harder thing to change.