Who wants to found a friendly society?
Let us run ahead of the future and start building the institutions we want to have
2025-10-27 — 2026-03-08
Wherein a friendly society is proposed anew, membership fees being invested in ETFs and such, while AI tools are envisaged to cheapen APRA-style compliance and to bargain for health and unemployment aid.
Here is a design exercise in alternative institutions for citizenry
Do we know what a friendly society is? The term has fallen out of use, but I have an inkling that it might be relevant again.
A friendly society (sometimes called a benefit society, mutual aid society, benevolent society, fraternal organization or “ROSCA”) is a mutual association for the purposes of insurance, pensions, savings or cooperative banking. It is a mutual organization or benefit society composed of a body of people who join together for a common financial or social purpose. Before modern insurance and the welfare state, friendly societies provided financial and social services to individuals, often according to their religious, political, or trade affiliations.
Why might we think it is time for new friendly societies?
- This bit of the Wikipedia page: “Before the development of large-scale government and employer health insurance and other financial services, friendly societies played an important part in many people’s lives.” If we are not confident that state capacity will remain strong in the future, we might want to hedge our bets.
- AI-based tools for collective organization might make them easier to run, and more effective at pooling resources and negotiating with providers of services.
Earlier on this blog, I tried to start a housing cooperative; I’m a co-operative-problem-solving kind of guy. This is way easier than that. I think that co-operative housing is way harder than it sounds, because housing is weird, and housing markets are weird.
We stopped using friendly societies because the state got better at providing services, and we started buying insurance. But now, if we think the state might not provide enough, friendly societies could become useful again.
TBH I don’t think there is a strong argument for a classic-style friendly society, where it just pools income to provide insurance. There are economies of scale to insurance that are hard to beat—e.g. if 10 people in my neighbourhood all pool our income to make our own flood insurance fund this is probably not great, because if we get flooded we will all be flooded at the same time. And the state is, at the moment, pretty good at pooling risk and provided services across the whole population.
Where this might get interesting is a mixture fo mutual aid, and investing in counter-cyclical assets, i.e. things that might still be a store of value at the time when the state cuts bac.
Here’s a model to consider: we start a mutual aid society that takes membership fees and invests them—say, in a diversified portfolio of industry-specific ETFs, or a unit trust. These investments could grow over time, potentially more than government tax income. When financial troubles hit, our society would have assets to negotiate and pay for services like healthcare and unemployment insurance.
This is more risky than classic income pooling, but that is a good design. Essentially, we don’t care if the friendly society does poorly in the good times; where we might get an advantage is one that is mostly likely to be useful to its members in bad times. This thing makes sense as a crisis instrument.
Friendly societies may have seemed expensive to run before because they competed with good state services. But as state services possibly decline and automation lowers running costs, they could become more viable again.
We could make this even better by creating a friendly society start-up kit. By documenting the process—legal steps, tax issues, software tools, meeting tips—we could help others start their own friendly societies. Imagine many of these societies pooling resources together for bigger deals.
What follows is a worked example of one institution we might set up for mutual aid in uncertain time: a friendly society that sets us up for state weakness and geopolitical risk.
I deliberately want to keep each society small—maybe 25 to 50 members, scaling up to a few hundred at most. This dodges coordination problems, keeps trust personal, and means we can iterate fast with AI-agent-assisted workflows rather than grinding through bureaucracy designed for large institutions. Beyond that scale, the answer is replication: publish a how-to guide, let others fork the model and spin up their own groups.
1 Interesting assets
Whatever the retail equivalent is of Catastrophe bonds, for natural catastrophe. Critical minerals, for energy scarcity. You can invest in defence contractors, if you are worried about war, although that is distasteful for various reasons.
2 Interesting services
Could such societies negotiate other deals for their members by using institutional clout?
3 Prior art
3.1 The historical record
Friendly societies have deep roots—burial societies existed in ancient Greece and Rome, medieval guilds extended the idea to illness and other misfortune, and by the 19th century friendly societies were enormous. David Beito’s From Mutual Aid to the Welfare State estimates that roughly a third of adult American men belonged to fraternal lodges by 1910, cutting across race, class, and gender lines. In Australia, the major orders—Manchester Unity Oddfellows, Ancient Order of Foresters, the Druids, the Rechabites—were central to colonial social infrastructure, providing sickness benefits, funeral funds, and medical access decades before the welfare state existed. Australian Unity, which still operates, traces its lineage to these traditions.
The standard narrative is that fraternal societies declined because the welfare state made them unnecessary. Beito’s counterargument is interesting: government programs displaced mutual aid, replacing reciprocity-based solidarity with paternalistic dependency. Whether or not you buy that framing in full, the relevant point for us is that the decline was contingent on the state actually providing. If state provision weakens, the ecological niche reopens. But: we would want to start such projects while we have the prosperity now, not wait until we lose the economic surplus that would have made them possible.
3.2 Modern mutual aid revivals
There’s a lot happening here already:
Broodfonds (Bread Funds). Cute modern Dutch variant. Broodfonds started in the Netherlands in 2006 as a mutual sick-leave scheme for freelancers. Groups of 25–50 people each contribute €33–€112/month into individual dedicated accounts. If a member gets sick, the others “gift” them a modest income for up to two years. There are now over 350 groups with 15,000+ members. The gifts structure is a hack to avoid avoids insurance regulation (in the Netherlands, at least), and the small group size keeps trust personal. A UK adaptation, Mutual First Aid, and a UK Bread Funds project are attempting to transplant the model.
COVID-era mutual aid networks. The pandemic saw a massive resurgence of mutual aid groups globally. In Australia, groups like Love Your Neighbour Melbourne (9,000+ members) organised grocery runs and other support. Mutual Aid Australia emerged as an umbrella cooperative. Most of these were informal—no investment function, no durable legal structure—but they demonstrated that the appetite for organised solidarity persists.
ROSCAs and chamas. Rotating savings and credit associations are essentially friendly societies under another name, and they are alive and well globally—Kenyan chamas, Southeast Asian tontines, Latin American tandas. These are mostly informal but some have substantial assets.
The Sustainable Economies Law Center has published a Mutual Aid Toolkit addressing legal structures for mutual aid groups in the US context.
4 AI Angles
The historical friendly societies needed armies of volunteers to do administration—meeting minutes, accounts, claims processing, correspondence, rule-books. Modern financial regulation demands even more: risk management frameworks, actuarial reports, governance documentation, regulatory returns.
The bet here is that AI agents can compress much of this work. Not eliminate it—I imagine we wold still need human judgement for the important decisions—but make it cheap enough that a group of 30–50 people can afford the overhead that used to require institutional scale.
I’ve asked my LLM research assistant for a detailed war-game of what this would look like—itemising APRA compliance costs, estimating what AI can compress, and sketching the shoestring team you’d need. The short version: traditional compliance for a friendly society runs $100k–215k/year, which is death for a small mutual. With aggressive AI automation of documentation, regulatory reporting, and actuarial prep work, looks like we could compress that to $41k–56k/year. Still painful at 50 members, but viable at 100+, especially if the compliance infrastructure is shared across a network of societies.
However, there are some workaround—ultra small boutique funds with only AUD2M, other legal structures that are not APRA-registered, and so on—that could be viable at smaller scale.
That companion post works through the five possible legal structures in Australian law—from full APRA registration down to a Broodfonds-style gift model—and a staged approach that lets you build capability incrementally without taking on institutional-scale costs from day one.
5 The replication kit
I think the most valuable thing we could produce isn’t a single friendly society but a template: legal docs, governance frameworks, software tools, investment strategies, compliance checklists, and lessons learned—all published openly so that anyone can fork the model.
The vision is many small societies (dozens of members each), loosely affiliated, sharing infrastructure and knowledge but independently governed. This keeps each group small enough for trust and agility, while the network as a whole achieves scale for collective bargaining and knowledge-sharing.
This is essentially the societal-scale part of the Broodfonds model—they published their framework and it replicated to 350+ groups. We’d be doing the same thing, adapted for Australian law and with a broader scope than just sick leave.
6 Actually democratizing to all of society
There is a serious gap here, which is that this is all well and good for middle class people who already have some surplus income. This might help keep middle class people above the poverty line, but it won’t help people who are already below it.
How about everyone else? As I learned in that companion post the baseline running costs for these things is high, so you cannot access them if you are already broke. We could massively lower the entry costs with regulatory reform, but I am not holding my breath for that. And if you have less capital to get return from, you get less return.
That said, while recognising this would be an instrument for middle class welfare, I think it is worth considering anyway, because keeping middle class people above the poverty line is also a good idea.
7 What I don’t know yet
- Whether there’s appetite for this among people I actually know (as opposed to people on the internet, where there’s appetite for everything).
- What the right counter-cyclical portfolio actually looks like—catastrophe bonds, critical minerals ETFs, and defence stocks are provocative examples, but designing a real crisis-hedge portfolio needs serious thought.
- What existing Australian co-ops or mutuals are doing that’s close to this vision—and what we can learn from them.
- A bunch of legal and regulatory questions that I’ve tried to work through in that companion post.
If you know about any of this, I’d love to hear from you. And, like, soon. We shoudl probably have started this already, but maybe it’s not too late to start.