The government is trying to define a startup, and finding out how hard that is.

After the 2026-27 Budget moved to replace the 50% CGT discount with cost-base indexation and a minimum tax, the backlash from founders and venture investors was loud enough that the government started looking for a way to carve startups out of the change. As of June 2026 it’s consulting on exactly that, aimed at low-capital, high-growth businesses.

But watch what happens the moment you try to write the definition. A tech startup, fine. A med-tech or biotech, fine. But what about a fashion designer launching a label, a scalable brand in the making, or a sole trader sewing to order? A craft brewery? A boutique consultancy with three founders and a clever method? A mining explorer with no revenue and a lot of upside? A trendy new coffee shop with a twist, where the founder swears it’s a franchise concept in waiting? Or the property developer at the start of a project? Every business starts up. “Startup” isn’t a category the law can pin down cleanly, it’s a judgement, and the moment a rule rests on judgement you’ve created a thousand edge cases and an argument at every one of them.

I raise it because there’s a corner of the law that confronted this exact problem decades ago and chose the opposite path. It refused to be clever. And the result is one of the more elegant pieces of regulatory design you’ll come across, even though almost everyone who meets it gets it backwards.

The thing everyone gets backwards

I hear a version of this at events fairly often. Someone has decided to start angel investing, they’ve been to their accountant, and they tell me (with the small pride of a box ticked) that they’ve got their sophisticated investor certificate now, so they’re allowed to invest.

It’s a reasonable thing to believe. It’s also the wrong way round.

The certificate doesn’t give you a right to invest. You always had that. What it does is let the company take your money without breaking the law. The permission flows to the issuer, not to you. And once you see it from the company’s side of the table, the whole apparatus – the blunt wealth test, the grey area in the middle, the caution everyone shows around it – stops looking like friction and starts making sense.

Where the disclosure went

The relevant words in Australia sit in section 708 of the Corporations Act, and they’re worth reading literally. The provision begins: an offer of a body’s securities does not need disclosure to investors under this Part if the investor clears one of the tests. That’s the whole frame. It isn’t a section about who is clever enough to invest. It’s a section about when a company is excused from producing a disclosure document.

I spent a long time in finance before I started doing this (buy side as a fund manager, sell side in equity capital markets) and on listed deals, disclosure is the pillar everything else rests on. A prospectus is a regulated document. Every statement in it has to be sourced. Legal and accounting teams work through comprehensive checklists so they can sign off on their obligations, because their names are on it and their professional indemnity sits behind it. It’s slow, it’s adversarial in the right way, and it’s expensive: tens of thousands at the low end, more often hundreds of thousands once the lawyers and accountants are done. For an IPO that’s the cost of doing business. For a startup raising a few hundred thousand to a couple of million, it’s uneconomic; the disclosure can eat a meaningful slice of the raise itself.

So the law offers a carve-out. If the people you’re offering to clear a threshold, you don’t have to produce the document. That’s why so many investors arriving from public markets are startled to find themselves wiring money on the strength of a pitch deck. They keep waiting for the prospectus. There isn’t one, and there was never going to be one, because the whole point of the exemption is that the company doesn’t have to write it.

So what replaces the prospectus? Nothing, legally

Here’s the part that creates confusion. The law takes the disclosure document away, but it puts nothing in its place. It imposes no positive duty on you to do diligence instead. It’s simply silent on the question. Your obligation, such as it is, is to sign an acknowledgment that you were given no disclosure document. That’s it.

Which is why the popular narrative: that sophisticated investors are qualified to do their own due diligence. It gets things slightly out of shape, however well it’s meant. “Doing your own due diligence” sounds like a positive qualification, a thing you’re equipped for. It’s really the upbeat way of stating a negative fact: nobody prepared a disclosure document, and you’ve acknowledged as much. The phrase puts a confident face on an absence. It describes a protection you’ve gone without, not a capability you bring.

And because the law is silent, the diligence done runs the full range, from prospectus-grade at one end (which, as we’ve established, can’t happen for commercial reasons) to none at all at the other. Nobody is checking which. That silence is exactly the space I was writing into in Due Diligence: Enough, But Not Too Much: the law won’t tell you how much diligence to do, so the right-sizing judgement is yours. Good angel DD is what you choose to put in the gap the regime deliberately left open, right-sized to the cheque, hunting down what you can in the time you have. What it can never be is the prospectus process, with its checklists and its hunt for the unknown knowns. That gap is real, and it matters later in this story.

The beauty of a test you can’t argue with

Here’s what makes the wealth test elegant: it’s deterministic. You either pass or you fail.

The income test asks whether you earned at least $250,000 in each of the last two financial years. Your accountant prepared those returns; they can certify it in an afternoon. The net assets test (at least $2.5 million) is a little more work, but the same kind of work: countable, checkable, done. The professional investor limb (an AFSL holder, or someone controlling $10 million in assets) is deterministic in the same way. There’s a number, and you’re above it or below it.

The thresholds have a telling side effect, though. They were set in 2001 and have never been indexed. So as wages and, above all, house prices have climbed, with the family home counting toward the asset test, the net has widened dramatically. When the test came in it captured around 1.9% of adults; today it’s about 16%. That’s over three million Australians, and on current projections it passes 40% by 2041. Nobody decided to let more people in. The number just sat there while the world inflated around it. That’s the deal you make with a deterministic test: it’s cheap to apply and impossible to argue with, but it measures wealth, not competence, and over time it measures even that more loosely.

Now put yourself on the company’s side. You’ve raised your round and you need to be satisfied you did it lawfully. What a certificate gives you is a clean binary: this person is wholesale, here’s the accountant’s signature, the file is complete. The thing the company is buying is certainty.

And the law wasn’t written with angels in mind. Angel cheques are usually small. The abuse the disclosure regime exists to prevent tends to live in things like mezzanine property deals, where the headline rate is high, the risk is buried in the fine print, and people write big cheques believing they’ve bought something close to fixed income. The law doesn’t sort by type of investment, though. A wholesale investor is a wholesale investor whether they’re backing a seed-stage software company or a second-mortgage property syndicate. Angels simply live under a rule drawn for a different room.

The grey area in the middle

Then there’s the limb that causes all the trouble: the experienced investor.

This is for the person who plainly understands what they’re doing but can’t clear the deterministic tests. They’re asset-rich in ways that don’t count, or they may simply not be wealthy yet, though genuinely capable. The law makes room for them in section 708(10), and the wording repays close reading. The offer must be made through a financial services licensee. The licensee must be satisfied on reasonable grounds that the person to whom the offer is made has previous experience in investing in securities that allows them to assess the merits, value and risks of the offer. The licensee then has to give the investor a written statement of its reasons. And the investor signs an acknowledgment that they were given no disclosure document.

Read that again and notice what’s changed. “Satisfied on reasonable grounds”. That’s not a number but a judgement of the one competence call the law otherwise refuses to make anywhere. And judgement is where the exposure lives.

Why the cautious people are right to be cautious

Picture the deal that goes wrong. The investor loses their money (which, in this asset class, is the base case for any individual deal) and decides they never really understood what they were buying. They argue they were retail all along, that the sign-off was generous, and that nobody explained the risk. Now the worse version, and it needn’t be a startup at all: a company raises under the exemption and turns out to be a fraud. This kind of bad deal might have been caught by a full prospectus process, and the diligence done here never would have. (I told one version of that story in The Patent That Killed an IPO.)

There is always a lawyer willing to test the argument. And here’s the part that’s easy to miss: if they win, the money has to come from somewhere. The investor isn’t suing the broke founder. They’re suing the financial services provider and the AFSL who signed off, and, behind them, their insurers. Those are the only parties left with a balance sheet. Worse, the legal costs land on them whether they win or lose; defending the case is expensive even when the sign-off was perfectly reasonable. The experienced-investor limb is exactly the soft target a plaintiff’s lawyer looks for, because it turns on a judgement that can be picked apart with hindsight. The deterministic test, by contrast, is a fortress: there’s a certificate, there’s a number, the investor was above it, end of argument.

So an AFSL asked to sign experienced-investor statements is right to be cautious. It’s being asked to put its judgement where a number would otherwise sit, and to carry the consequences when that judgement is second-guessed in a courtroom.

But couldn’t we just test for it?

The obvious fix comes up constantly. If the problem is competence, test for competence. This debate runs hottest in the US, where the accredited investor rule is the direct equivalent of our sophisticated investor test and has been largely unchanged since 1982: $200,000 of income or $1 million of net worth, numbers never indexed, which is why the US net has drifted exactly as ours has. Jason Calacanis makes the case regularly that the rule should work less like a wealth bar and more like a driver’s test: something you qualify for by passing, that everyone sits, rich included, paired with a cap on how much of your income you can put at risk. The logic is hard to argue with: a person can lose the lot at a blackjack table but can’t put $500 into a startup unless they’re already wealthy.

And it turns out a version already exists. In 2020 the SEC added a knowledge-based path: pass certain securities exams and you qualify regardless of net worth: competence rather than wealth. It’s narrow and bureaucratic, but it’s real. The competence test isn’t hypothetical.

Run it back through the company’s lens, though. The company’s nightmare isn’t an under-qualified investor, but a regulator pinging it for an illegal raise. Does a certificate from an angel investing course give the company the comfort a deterministic test or a recognised licence does? Probably not. A course certificate is still a judgement call dressed up as a credential. Judgement, as we’ve seen, is the thing that gets tested when the money’s gone.

Reviewed, and left alone

Almost every developed market runs a version of this carve-out: the US accredited investor, the UK and Singapore tests, the same architecture under different numbers. And in most of them the thresholds have been formally reviewed and then, pointedly, left where they are.

The Americans were the most candid about why. When the SEC last overhauled its definition it looked squarely at the inflation drift and chose not to lift the numbers. Part of the reasoning was that information is far more available than in 1982, so the average participant is more capable than the threshold assumes. But the blunter part: there had been no wave of fraud at the existing thresholds, so a large-scale change couldn’t be justified. The harm the rule exists to prevent simply wasn’t showing up.

Australia hasn’t said it so plainly. Our review recommended raising the bar; the government, facing a startup sector that depends on this capital, has so far declined to move it. But the same logic is at work underneath. Ask who the disclosure regime exists to protect, and the answer is the average worker and below: the people for whom a bad private placement isn’t a setback but a catastrophe. Those people were never clearing a $2.5 million asset test. The drift has swept in the comfortable middle, who can absorb a loss and who, tellingly, don’t generate the lawsuits. The genuine abuse still clusters where it always did: the high-yield, fixed-income-looking products sold to people chasing safety, not the angel rounds. So the cost-benefit of tightening looks poor. You’d choke off legitimate startup capital to protect a group that mostly isn’t being harmed. Which is, I suspect, exactly why nobody has.

Which loops back to where we started. If the government does manage to write a clean definition of “startup” for its CGT carve-out (and that’s a real if) you could imagine the same definition doing other work. A startup-specific experienced-investor carve-out, say: let capable people without the wealth back qualifying startups, even if they can’t clear the asset test. It’s an appealing thought, and it would answer the access problem the drift has created. But notice it doesn’t escape anything. A carve-out scoped to “startups” is still a category built on judgement, sitting on top of another category built on judgement. You’ve now got two definitions a lawyer can pick at instead of one. And before any of it reaches an investor, it still must pass the one test nobody legislates for.

The boss at the back of the room

Which brings me to the thing that actually governs all of this, and it isn’t the law. It’s insurance.

Every service provider in this chain needs cover, and insurance is bought by answering a questionnaire. The questionnaire asks about every activity that carries risk, and if you answer it inaccurately your cover can be voided when you need it most. Now think about how you answer a question about experienced investors. It’s a category built on judgement, so it can’t be answered cleanly, which means the insurer prices in the uncertainty, attaches conditions, or wants to see your process before writing the policy.

And remember the lawsuit. It only takes one case – one aggrieved investor, one sympathetic judge, one settlement – for “do you deal with experienced investors under 708(10)?” to become a line on every questionnaire in the country. That’s how these things propagate. The deterministic test answers that line in one word. The judgement-based one starts a conversation, and conversations with insurers cost money.

So here’s where it lands. An angel investing course with a certificate might survive an ambulance-chasing lawyer. You could argue the investor was properly educated. What’s much harder is surviving the insurer, and the insurer holds the real lever. If cover for servicing experienced investors gets too expensive, or the questions too hard to answer honestly, or the policy is simply withdrawn, the provider stops offering it. Not because the law forbids it, as the law makes room for it, but because nobody will insure it at a price that works. Retail premiums run at multiples of wholesale for the same reason.

The deterministic test, awkward as it feels when you’re the capable investor stuck on the wrong side of it, is the thing that survives all three pressures at once: the regulator, the lawyer, and the insurer. That’s why versions of it have lasted in so many countries for so long. It isn’t clever. It’s the opposite of clever. And that’s exactly why it works.

So the next time someone tells you they’ve got their certificate and they’re allowed to invest now, you’ll know what to gently correct. The certificate was never about you. It’s the company’s permission slip, blunt on purpose.

Richard Moore — MooCoo Ventures

Richard Moore is co-founder of MooCoo Ventures, an angel syndicate that co-invests alongside Brisbane Angels, one of Australia’s most active angel groups. He has made over eighty personal angel investments since 2013.

moocoo.vc