So you’ve joined a union

Field notes for white-collar professionals in Australia

2026-04-24 — 2026-04-24

Wherein enterprise bargaining under the Fair Work Act is situated as the central mechanism of Australian unionism, and the practical leverage of the workplace delegate role is set forth.

economics
ethics
incentive mechanisms
institutions
politics
work

tl;dr Australian unions are flawed — affiliated to the political parties, biased toward incumbents, institutionally tired. IMO they are still worth joining, and being a workplace delegate is high-leverage: it opens up points of intervention that are hard to come by otherwise (inside gossip, colleague ties, negotiating room, a widened range of sayable things at work). Simply having a second centre of power in an organisation can be valuable, for the workers for sure, but even for the organisation itself. These notes come from my tour as a delegate with the CSIRO Staff Association at Data61, but draw from the experiences of my colleagues and comrades in the CPSU more broadly.

Figure 1

Context on me: I was a workplace delegate for the CSIRO Staff Association — which is a section of the CPSU, the public-service mega-union — at Data61, CSIRO’s digital research division, during a stretch when Data61 was being systematically under-resourced. My field of view is therefore specific. This is a guide to being a member or delegate in a professional-staff union in Australia, not a guide to unionism in the abstract. People should write guides for meatpacking or wharves, and for other professional unions too. And they have. But this one is mine.

1 The Australian context

Union density in Australia is about 13.1% of employees as of August 2024 — up from 12.5% in 2022, the first recorded increase since 2011. Over the longer term the trend is down: in 1992, density was 40%. That is to say, the movement has been dying for three decades, and recently it, uh, stopped dying. Something interesting is happening at the margin, driven mostly by young workers, but from a low base.

Inside those averages, professionals are overrepresented (around 20% density) and education-and-training workers more so (27%). So if you are a white-collar knowledge worker reading this, you are not colonising a strange industry — there are plenty of us already.

The Australian union system is unusual in a few ways.

Enterprise bargaining is the central mechanism for negotiating conditions. Pay and conditions for most workplaces are set by an enterprise agreement negotiated between the employer and a bargaining representative — usually the union — at the level of the single workplace or company. Since the Fair Work Act 2009, this is how a modern Australian pay rise gets made. Our union’s main job, most of the time, is to negotiate this instrument.

Protected industrial action is only legal during bargaining for a new agreement, and only after a secret ballot via the Fair Work Commission, and only with three working days’ notice (for most workplaces). Striking in sympathy with another workplace is not protected, which is to say, it is effectively illegal. So, don’t be bringing to mind militant historical or european unionism when you think about the modern Australian version.1

Most big unions affiliate to the ALP. “Affiliate” means something specific: the union’s President and Secretary sign a pledge to the Labor platform, pay a per-member affiliation fee, and the union’s members are thereby entitled to 50% of delegate positions at Labor state and national conferences. The ACTU itself does not affiliate — it is a peak body — but most of its largest member unions do. In practice, this means some fraction of our dues is helping to elect a particular political party, and that the union’s political stances will be calibrated against not upsetting Labor. We can think this is fine or not fine. IMO it is a narrow kind of politics — not obviously better or worse than the available alternatives, but I am personally not a fan.

2 The flaws

2.1 Incumbency bias

This one is obvious when you think about it. Unions are disproportionately composed of, and run by, indefinite-tenure staff. The consequences cascade.

In enterprise bargaining, the people who turn up to the member meetings, write submissions, and stress-test the employer’s offer are overwhelmingly indefinite staff. They know how it works because they have been through a round before. The fixed-term, casual, and labour-hire workforce — which in research is often the majority — frequently doesn’t know bargaining is happening, or doesn’t know they can participate, or rationally does not invest in a process whose outputs they may not be around to enjoy. So the agreement gets tuned for the people in the room.

A large fraction of delegate time goes into redundancy negotiation. That is not a criticism of delegates — redundancies are harrowing and someone has to fight them — it is a description of where the effort flows. It also reveals the shape of the service: a union that spends most of its crisis capacity defending tenured jobs is, implicitly, a service for people who have tenured jobs to defend.

For young workers, on a fixed-term contract, or otherwise outside the incumbent class, we feel this as a kind of quiet irrelevance by default. The union is still worth joining. Just go in expecting to push.

2.2 Operational conservatism

Unions per default do not take the operational needs of the employer seriously, or at least make ambit claims that imply they don’t. There is no sense in negotiating a gold-plated employment contract that causes the entire organisation to tank. This is not what they are for, I suppose, but the consequences show up in the agreement. In bargaining I have watched clauses land which would, I could see, make some internal workflow I care about more cumbersome. The classic example is that by making it hard to sack people and hard to hire people on short-term contracts, the union starves the organisation of the ability to experiment with short-term contracts for new skills and capabilities.

A related failure: unions rarely favour creative or radical approaches. “We got this last round, let’s get a bit more of it this round” is the default. Arguments for novel structures — more flexible project-based teams, different modes of career progression, profit-sharing arrangements for commercialisable work — tend to fall outside the union’s Overton window. Part of this is presumably institutional memory (those structures have been used as thin ends of anti-worker wedges before). Part of it is organisational inertia.

For the kind of person who likes working on hard interesting problems inside a functional organisation, the union’s default policy stance can feel like being in a 1950s time tunnel, aimed at a very different, less dynamic world of work. Again, this is not what the union is for, but it is a consequence of the institutional setup.

2.3 The professional-organiser problem

Big unions employ professional organisers and officials. These are often good, dedicated people. They are also incumbents in their own sense — they have career paths, personal influence, factional alliances, and a durable stake in the union’s continuing to be run the way it has been run. In the worst cases this produces performative, communications-hoarding, process-heavy unionism that does not scale with the membership’s engagement.

The tell I watch for is the communications architecture. If the union does not host a chat channel where members can talk to each other — if all communication is union-to-member broadcast from the officials, with no lateral member-to-member channel — then that is a union whose professionals are more comfortable talking to members than enabling members to talk to one another. We can draw our own conclusions about why. My guess: an engaged, talkative membership is harder to steer.

I don’t think this is malice. I think it’s an equilibrium that is comfortable for professional staff who are only shallowly engaged with any given workplace’s problems, and that persists because nobody with the power to change it has much incentive to.

2.4 The enervating Fair Work settlement

The Fair Work Act gives Australian workers a floor of rights (minimum wages, leave, unfair dismissal protection, an enterprise-level bargaining mechanism) that is better than nothing. It also gives Australian unions a narrow, procedural, Commission-mediated menu of things they can do. Protected action is protected only within tightly drawn conditions. Pattern bargaining across an industry is largely forbidden. Sympathy strikes are forbidden. The whole apparatus assumes that industrial conflict should be rare, procedural, and bounded.

This produces a curiously enervated unionism. The union rarely needs to mobilise its membership at scale — the Commission is right there, the processes are what they are, and the union’s comparative advantage is procedural expertise rather than mass pressure. Compared to what one reads of unionism in, say, the US or continental Europe, the Australian version can feel strangely administrative. I find this both a relief (nobody is getting arrested) and a limitation (nobody is winning very much, either).

3 Why it’s worth it anyway

Given all the above, why did I sign up as a delegate, and why would I do it again? Because the leverage is huge. Being a delegate got me:

Inside gossip. The delegates’ meetings are where one learns what management is telling the union, which is often a truer signal than what management is telling staff. One sees drafts of redundancy proposals before they become announcements. One hears which divisions are being quietly wound down. Going the other way, the delegate carries back useful information about the workforce that the management chain on its own would not have heard.

The more underrated form of this, IMO, is lateral gossip — what peers know about each other that a hierarchical organisation aggressively filters out. Our formal view of colleagues in other teams is mediated almost entirely by those teams’ managers, who have career incentives to present their units as competent, well-run, and indispensable. The union is one of the few structures in a big organisation where frontline staff from different teams compare notes, and discover: which empire-building manager is winning headcount they don’t have work for, which team is quietly drowning, which senior technical lead is the bottleneck nobody can work around, which department’s nominal stars are coasting, which quiet group does the good work that never shows up in the monthly report. This stuff matters for a lot of mid-career decisions — internal transfers, whether a cross-team dependency is safe to rely on, which up-chain promises are worth trusting — and in an excessively hierarchical organisation almost none of it is available through any other channel. People seriously underestimate how much of their picture of their workplace has been composed for them by interested parties. The union cuts across that.

Strong relationships with colleagues across silos. Research organisations partition into teams that rarely talk to each other, and that compete for resources. The union is one of very few structures where a data scientist talks to a comms officer talks to a facilities manager as peers with a shared interest. This defangs a lot of internal resource competition simply by putting faces on the “other” side of it. The employer benefits from this too, even if they rarely notice: a less siloed workforce wastes less energy on internal turf wars.

Esprit de corps. Not nothing. Knowing that one has a dozen colleagues whose numbers are in one’s phone, and who will pick up, is a psychic resource in a way that is hard to describe until one has had it.

A widened range of sayable things at work. Delegates, by convention and by the adverse-action provisions of the Fair Work Act, have more protection for speaking publicly about workplace issues than ordinary staff do.2 We can say things a rank-and-file employee would be nervous to say, because we are saying them in our delegate capacity. This is not absolute — we can still be sacked for many things — but it widens the range of sayable things considerably. I used this a lot.

A direct line to powerful people when it mattered. At bargaining time, at redundancy time, at restructure time, delegates are in the room with the CEO, or the DG, or their direct deputies. One gets to negotiate with the people actually making the decisions. For a mid-career professional this is an unusual amount of face time with senior leadership, and the skills one builds — reading the room, finding the pressure points, knowing when to escalate — are portable.

Friends. I left the delegate role with more friends at CSIRO than I had before I took it. I don’t know a better return on my time.

4 A second centre of power

Most of the above is about what the delegate gets. The less obvious side is what the employer gets from having a functioning union on the premises. IMO a sensible leadership team treats the union as a cheap source of information and a cheap source of accountability that would otherwise be unavailable to it — which is the core of what having a second centre of power inside an organisation buys you.

A back-channel past toxic managers. Every formal complaints mechanism I have encountered routes upward through the chain of command. Which fails at exactly the case that matters most: when the problem is the chain of command. A staff member being ground down by their manager has almost nowhere to go internally — the next rung up is the bully’s mate, or their delegate of authority, or someone institutionally invested in not seeing the bully as a problem. The union is a parallel chain that does not route through the bully. A well-run HR function treats this as a useful signal, because it surfaces problems the formal channel would have swallowed. A badly run one resents it. The ones that resent it tend, in my experience, to be the ones whose formal channel is broken in ways they would rather not admit.

Risk identification. Frontline staff know things that never make it onto a dashboard: safety hazards, project-failure signals, mis-scoped deliverables, customer-abuse patterns, strategies that will not survive contact with reality. A lot of this information is uncomfortable, which is precisely why it gets filtered out as it moves upward through layers of people whose performance reviews depend on appearing to have things in hand. The union is one channel by which those signals can bypass the incentive structures that suppress bad news going up the chain. Same logic as whistleblower protection, operating at lower stakes, more often, on smaller problems — which means the organisation finds out about them while they are still small.

A commitment device. An enterprise agreement, together with an organised workforce that will notice breaches of it, is a way for senior leadership to credibly commit to things. Promises management would like to make but cannot otherwise guarantee — about workload, about career paths, about fair treatment of a category of worker — become more durable when backed by a bargaining process. In principle, the employer gets a more motivated, lower-turnover workforce in exchange for being bound to its own stated intentions.

None of this is free. The union is also, sometimes, the thing that negotiates an inflexible clause into our enterprise agreement that makes something we want to do hard. But on balance — for an organisation that can tolerate being told uncomfortable things by people it cannot easily fire — a functioning union looks closer to an internal audit function than to a hostile external force.

This is not to pretend away structural conflict. Interests do diverge sometimes — about wages, about job security, and above all, who carries the risk when things fail. Conflict, when it arrives, is an important mechanism: the bargaining apparatus is built around exactly those divergences. But most decisions, at least in my workplaces, most of the time, do not need to be zero-sum wars. Some zero-sum fights I was involved in were made to be zero-sum to keep people busy, and those could have been turned into collaborative problem solving. When a bad manager stokes conflict to distract attention from their own failings, this is a power play, not an inevitable state of affairs.

Neither “side” is a monolith. “Workers” and “employers” are each coalitions — junior and senior staff, contractors and indefinite, technical and administrative; shareholders, executives, line managers, finance, operations — with partially overlapping and partially competing interests. Some worker factions are at odds with other worker factions more than with any employer faction. Cross-line alliances — this technical team and the CTO, this cohort of staff and the CFO’s cost-containment agenda — are common and often stable. The two-team framing is convenient shorthand, not an accurate map.

Leaning on the adversarial frame as the default identity of unionism — worker-vs-boss as our core brand — is great for recruiting — people love a fight with the Other. But it has costs: it alienates the moderates on both sides, and leaves win-wins on the table that would otherwise have been obvious. It is important, but it is not the only game.

Most executive teams don’t want union pushback. IMO they should want all the union engagement they can get. A second centre of power inside a dysfunctional organization can be one of the few ways management gets the information and accountability needed to stop making predictable, avoidable mistakes — which is a case of the institutional alignment problem. The union, or at least many unions I have been involved with, can be a flawed instance of that second centre: too procedural, too political, too disengaged, for all the reasons I listed above. But those flaws are not, mostly, irreducible features of unionism. They are habits that unions can fall into, and which we can change by joining the damn union and making it work better.

Members are within their rights to ask that unions be better. Hell, managers are within their rights to want that too. But the union is us. If we want a second centre of power in our organisation that is not captured, not incumbency-biased, and not hoarding communications, the mechanism by which it comes to exist is: we join, and we show up, we pay 0.7% of our salary, we invest our time and sweat equity in making everything else better. The price of admission is unusually low.

5 Advice I wish I had on day one

A few things I had to work out the hard way.

Turn up. A startling fraction of the influence in the delegate body accrues simply to the people who turn up to every meeting. This is unglamorous and correct. Showing up is 80% of the job. If there are literally too many meetings to attend, ramp it back. If the union insists on stupid held-over structures (e.g. place-based organising for an online organisation with distributed teams, looking at you CSIRO Staff Association) we could make it a project to try to modernise.

Ask for the bargaining log of claims early. Every round of enterprise bargaining starts with a list of things the union is asking for. If you are not on the bargaining committee you may not see it until it is near-final. Ask for it. If there are claims we care about, we have to fight for them early in the process, not at the end.

Cultivate union communications the union staff cannot control. Ideally you want to trust your organiser (for the record, I have always trusted my organiser’s intent and competence). But baking too much control into the union staff’s roles sets up bad incentives for them to hoard power just like the bosses. If the union doesn’t run a members’ chat, start a group chat among the delegates you trust. If the union’s email list is one-way, the ground truth of what members want must reach decision-makers through some other channel.

Learn the instrument. So easy these days that there is no excuse. Load the current enterprise agreement into an LLM. The last one too. Also, Part 2-4 of the Fair Work Act. Ask questions. Read specific passages. Learn the stupid terms of art. We become much more useful to our colleagues the moment we can answer basic questions about the instrument instead of referring them upward.

Practise communication skills. Delegate work is the best-calibrated training environment I have found for the workplace-communication skills that white-collar work otherwise hides from us until we are mid-career and already expected to have them: writing a clear update for members, chairing a meeting where people disagree and we need a decision at the end of it, negotiating with a senior executive who has lawyers (we don’t), delivering hard news to a colleague, pushing back on a bad idea without burning the relationship. The stakes are high enough that sloppiness has consequences, and low enough that we get to try again next meeting. Take the opportunity.

Do not let this become your whole life, or even your whole job. Union work is high-leverage precisely because the baseline engagement of everybody else is so low. Which means there is infinite demand for your time. Decide in advance what fraction of your working week you will spend on it, and stick to it. Extension projects are good, but don’t pick too many.

Recruit more delegates. The more delegates we have, the more influence we have, and the less work each delegate has to do. Don’t solve everything yourself; solve one thing and use that as a recruiting pitch for the next delegate. Showing people they can make life better is hugely encouraging.

You get so much love for doing this. People, mostly, value and appreciate the work of delegates. I got given so much respect from my colleagues when I’d shown up for a tough time, but they would make sure the line manager was not around. Then the line manager would also thank me, when their direct reports were not around. Double thanks!

6 If we were starting from scratch

A sceptical reader will have noticed that almost everything in this post is a second-best argument. The union is a patch on a firm shape that has employees and shareholders on opposing sides of an ownership boundary, with managers in between. A lot of what the union does — surfacing risks, piercing silos, keeping managers accountable, sometimes even committing the enterprise to its stated intentions — is only necessary because the firm is designed the way it is.

Other shapes are possible. Let me get galaxy-brained for a second: Worker-owned firms, consumer cooperatives, and various flavours of co-determination redistribute the ownership boundary and fold the “second centre of power” into the first. I have written separately about how one might divide equity in a firm being built along those lines. More broadly, this sits inside the frame of institutional alignment: any institution we build is, in effect, a trained system whose incentives and reporting structures have to be tuned for the work we want out of it, and most organisations have simply never been tuned for the work of collectively caring for the people inside them.

A ground-up redesign is not on the menu for most employees of most organisations. But the frame is useful for explaining why even a flawed union is high-leverage: unions are one of the viable ways to retrofit a power distribution into an organisation that was not designed with one. If we could start over, we might build the distribution in at the foundation. We mostly can’t, so we patch.

7 Resources

Background and statistics

The legal instruments

Peak bodies and some relevant unions

8 Things I am less sure about

I don’t know how much of the communications-hoarding phenomenon is specific to CPSU-style professional unions versus a general feature of mid-sized Australian unions. Anecdotes from other delegates suggest the latter; I haven’t confirmed.

Footnotes

  1. The history is pretty cool though, esp. Paddy Troy↩︎

  2. Strictly, the Fair Work Act’s general protections (Part 3-1) are what actually do the legal work; the constitutional “implied freedom of political communication” is a much narrower doctrine that rarely maps cleanly onto workplace speech. Conflating the two is sloppy but a useful shorthand for the lived experience of being a delegate.↩︎