Why the Good Thing Waits

(Article 14 of The Available Remedies)

You hear it and something in you tightens. Somewhere, someone has shown that a job can be done faster, or cheaper, or with less harm. The proof is in. The trial has been run. And still the country goes on doing it the old way, year after year, as though none of it had happened. Anyone who has watched a better method sit unused knows the feeling.

Walker Briefing is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

When you are standing in that frustration, one explanation is close to hand, and it feels good to reach for. Someone is blocking it. The system has been bought. The good thing threatens somebody’s money, and the money has arranged for the good thing to fail.

Sometimes that is exactly right. More often it is not. And knowing the difference is not a fine point for people with time on their hands. It is the whole practical question. Get the reason wrong and you reach for the wrong fix. The wrong fix does not just fail quietly. It fails in a way everyone can see, and it teaches them that reform is hopeless, when the truth is only that the reform was aimed at the wrong target.

So this chapter does something smaller than accusing, and I think more useful. “The system is slow to take up the better method” is not one thing. It is four things. And the four need four different remedies. Telling them apart is the whole of the work. What follows is not a verdict on who is to blame. It is a way of asking, of any case where a known improvement is stuck at the door, which of the four kinds of waiting this one is. Because the answer is what tells you what would actually open it.

The four kinds of waiting

The first kind is principled caution, or what we might more plainly call careful waiting. The method looks good, but the proof that it holds up over time has not been built yet. There is no long record of how it behaves over decades. There is no settled standard a bank or an insurer can use to price it. The slowness here is not a failure of nerve. It is a fair refusal to move faster than the evidence. The fix is to build that evidence and that assurance, and until it is built, the caution is doing its job.

The second kind is the good that nobody is paid to do. The method works, and everyone who knows the field agrees it works. But the benefit is spread thin across a lot of people and turns up years later, while the cost is concentrated and lands now. No single person or body is paid to carry it, so it gets carried by no one. There is no villain in this one. There is just a set of rewards that quietly pays too little for the very thing it should pay most for. The fix is to line the payment up with the thing you want done, so that someone is actually paid to do what helps everyone a little, later.

The third kind is the inherited category, or in plainer language, the wrong word stuck on long ago. The method is governed by a label that was made for something else, and the label drags in friction that has nothing to do with what the method actually is or does. Nobody sat down and decided to block it. A word attached itself a long time back, for reasons that made sense then and in another setting, and it has gone on doing its work ever since, without anyone stopping to ask whether it still fits. The fix is to look at the label again, which is harder than it sounds, because old labels are exactly the thing large bodies are worst at reopening.

The fourth kind is partial capture: the case where someone really is leaning on the scale, and even here only partly. The incumbents, the ones already in place, have some hand in the delay. But their hand is tangled up with real and legitimate difficulties, and how much weight to give the one against the other is genuinely arguable. This is the kind people reach for first and can prove last, because it is the hardest to show and the easiest to assert.

One warning before the cases. These four are not sealed boxes, and almost no real case sits wholly inside one. Most are mixtures. The question is not which single box a case goes in. It is which of the four is doing the most of the work, because that is the one whose fix has to lead. I will name each case by the kind that leads it, and where a second kind is also at work, I will say so.

Hold those four in mind. Now walk the cases.

Caution that is not blindness

Take modern ways of building: prefabricated, modular, panelised. Made in a factory, in large part, and assembled on site. Australia is a clear laggard here. These methods make up under five per cent of home-building in this country. Sweden prefabricates something like four-fifths of its single-family homes. Japan builds roughly one new dwelling in seven this way, and even there the share has been easing rather than climbing. The Commonwealth Productivity Commission has estimated these methods can cut costs by up to a fifth and build up to half again as fast. On the face of it, this is exactly the story the frustrated reformer wants to tell. A proven method, cheaper and faster, held back.

But look at what is actually holding it back, and the simple capture story does not survive the first contact with the evidence. The barriers are these. Building codes written around the old way of building on site. Construction finance built on staged payments that release money as a house rises on its slab, which does not fit a method that creates most of its value in a factory before it ever reaches the site. And the caution of lenders, insurers and valuers, who have no long track record to price against. A United Kingdom parliamentary inquiry found that new construction systems lacked long-term performance data, and the answer there was not to force the method through. It was to build a scheme that could vouch for it. That is careful waiting. It is the first kind, and it is doing its job.

And notice, before we move on, that this first case is already a mixture. Those building codes, written for one way of building and now taxing another, are a small piece of the third kind, the wrong word stuck on long ago, sitting inside a case whose main driver is the caution of the first. I file housing under caution, because that is what leads, and that is what any fix has to answer. But I flag the second thread, because this is how the tool is meant to be used. Not to force each case into one box, but to name the kind that leads and to notice when more than one is present.

More than that, the reform is already moving, and it belongs here at full weight. New South Wales introduced legislation in May 2026 meant to make it the first Australian jurisdiction to recognise modern methods of construction in law. The Australian Building Codes Board is developing a national voluntary certification scheme. The Commonwealth has announced about fifty-four million dollars for prefabricated and modular housing. A major bank has brought in a construction-loan pathway built for prefabrication. An argument that set out to prove the system cannot see this solution has to reckon, here, with a system that is beginning to see it clearly, and to move.

What this case settles clears away an error before it can spread through the rest. Adoption here is tracking the benefit that is concentrated, immediate and easy to see. Once finance can price the method and the law recognises it, the money follows, because now the reward is visible to someone. The slowness that is left is not suppression. It is the ordinary and defensible lag of assurance catching up with something new. This case comes first for a reason. The discipline that refuses to see a plot where there is none also refuses to dress ordinary prudence up as persecution, and refuses to reward those who would profit when honest caution about an under-tested product gets recast as the system conspiring against them.

The method with no one to pay for it

Preventive health is the cleanest case of the second kind, and I come to it as a general practitioner, which shapes how I read it but does not stand in for the evidence.

There is a figure that gets quoted a great deal: that it takes, on average, about seventeen years for new evidence to reach the consulting room. I will not lean on that number as though it were solid, because it is not. The original work drew it from a small set of specialties, with different people meaning different things by “reaching practice,” and later commentary has rightly picked holes in it. What the same body of work does show, and shows more usefully, is that the lag is not fixed in stone. Where someone actually resources the job of putting the evidence to work, and does it together rather than alone, uptake of around four-fifths in about three years has been reached. I offer that as an illustration and not as a law, from the same literature and held to the same caution. It shows the mechanism at work.

And the mechanism is this. Prevention works, but its benefit is spread out and delayed. It is shared across a whole population and shows up years or decades from now, while treatment pays a doctor today for the patient sitting in front of them. Nobody blocks prevention. No powerful player has to capture a regulator to keep it down. The system simply pays too little for it, because the reward for prevention is arranged, by the very way care is funded, to arrive too late and too spread out for anyone to gather up. This is a real failure, and it is not capture. What it shows is that the speed of adoption tracks where the benefit falls, not how good the method is. The fix is not to expose a villain, because there is none. It is to line the payment up with the work, so that someone is at last paid to do the thing that helps everyone a little, later.

A label built for the room

The third kind is the inherited category, or in plainer language the wrong word stuck on long ago, and before I come to the case I know best, here is one I can hand you without defending any interest of my own. For most of Medicare’s life, the thing the schedule was built around was the visit: a doctor and a patient in the same room. Telehealth, care given over a phone or a video link, was not so much banned as barely allowed for, let through only by narrow gates of distance and circumstance that had little to do with whether care at a distance was any good. A consultation, the schedule quietly assumed, was a thing that happened in a room. Before 2020 it ran at well under one per cent of services. The friction that assumption produced had nothing to do with the actual worth of speaking to a patient over a video link, and everything to do with a label that had never been asked to imagine the room might be optional.

Then it was asked. When the pandemic made the room dangerous, the Commonwealth created two hundred and eighty-one temporary telehealth items on the Medicare schedule between March and May of 2020. The quickest of them were built by copying the wording of the existing face-to-face items, because the consultation was the same consultation. The department itself later called the change roughly ten years of reform in ten days. In January 2022, two hundred and eleven of those items were kept for good. A method that had been a sliver of a per cent of general practice became, inside a couple of years, a standing part of how the country sees its doctors.

Two things need saying here, and I put them in rather than leave them out. First, I am a general practitioner and I bill under this schedule, so I note the interest, though it is a broad one shared with every doctor in the country and nothing peculiar to me. Second, the old caution was not empty. There are real questions of quality and of oversight in care given at a distance, and the government was right that telehealth does not suit every problem. That is the first kind of waiting doing a little work inside the third, exactly the sort of mixture the tool is built to name. But the main fact is plain. A label built around being in the room governed, for years, a method whose worth had nothing to do with being in the room, until an emergency forced the label open and showed how quickly it could have opened all along. There was no villain to expose. The fix was to look at the label again, and under pressure that is just what the system did.

A crop governed by the wrong word

The case I know best in this third kind is low-THC industrial hemp, and I have put it after telehealth on purpose, because it is the one I have to handle most carefully. The reason is awkward, and I will not dress it up. I lead a political party whose purpose touches this crop, and I would gain from a conclusion that the rules around it are out of proportion. That is exactly why I hold this case to the same weight as the others and no more, why I will not let it carry the argument, and why every fact that has to bear weight in it comes from the government’s own mouth rather than from mine. A declared interest is still an interest. The guard against it is not confession. It is discipline.

Here is what the government’s own record shows. Under the Single Convention on Narcotic Drugs of 1961 and the Narcotic Drugs Act of 1967, cannabis is classed as a narcotic by its category. The Commonwealth’s own Office of Drug Control says plainly that the law is silent on the THC level of cannabis extracts, and that all extracts are treated as drugs whatever their actual cannabinoid content. The rule, in other words, turns on what the plant is called, not on what a particular crop does. And the regulators themselves describe this particular crop as non-intoxicating. A South Australian primary industries agency states that industrial hemp’s low THC level means it has none of the mind-altering effects of the illicit varieties, and the Office of Drug Control’s own security guidance reasons that low-THC material is not attractive for recreational use. A crop the regulators call non-intoxicating is governed by a category that is silent on the very thing that makes it non-intoxicating.

Two corrections belong here, both of which cut against a looser version of this case, and both of which I carry at full weight, because leaving them out would be the easy and dishonest path. First, the word narcotic is a legal term, not a medical one. The category lists cocaine, a stimulant, alongside the depressants, which tells you the label carries a policy decision about control, not a claim about how a substance acts. That is true, but it is colour, not proof. It shows only what no informed person argues about. The real mismatch is narrower and stronger. The word narcotic, in its ordinary and its medical sense, suggests an opioid-like power to kill, and it has attached to a crop that does not cause the deadly slowing of the breath that defines an opioid overdose. That is a claim about what the substance does and does not do, and it holds. Put as a claim about the fine detail of brain receptors it would not hold, and I do not make it there.

Second, and this genuinely surprised me when I checked it, Australia’s own limit is generous, not tight. Our industrial-hemp threshold sits at or below one per cent THC, which is at the higher, more permissive end by world standards, while major markets including the United States, Canada, China and the European Union commonly use a 0.3 per cent line. A higher threshold means more of the plant counts as hemp, not less. The friction that threshold causes is not at the growing but at the border, because a crop lawfully grown here at up to one per cent goes over the 0.3 per cent ceiling of the big buyers overseas. And even that friction has a recorded history of being noticed and removed. In 2020, Parliament passed legislation whose own explanatory notes called the rules blocking exports of these narcotic-classed goods unnecessary and unintended, and set out to stop treating them differently from other farm goods that carry similar risks. The word in the record is barriers, the scope is exports, and it was the government that named the disproportion.

So here is what this case shows. A crop the regulators themselves call non-intoxicating, governed under a category in which its cannabinoid level is treated as beside the point, with a home threshold that is generous and one specific export barrier that was recognised and lifted. And here is what it knocks down. Three tempting overstatements: that the one per cent limit holds back the growing of it, that the safety case rests on receptors being missing from the base of the brain, and that the cocaine-in-the-category point proves the whole rule is mad. The fix is not to accuse anyone. Nobody lied. A category was inherited from a treaty that used the word loosely, and it has gone on doing work that no one has stopped to look at again. The fix is to look at the label again. That is a harder and a more honest thing to ask for than a villain.

The one case where the word half-fits

Renewable energy is where leaning on the scale actually happens, and it is the case that tests whether this whole exercise is disciplined or just noise, because it is the case where I could reach for the strong story and where I am going to refuse to.

Both explanations are genuinely backed by the record, and that is the whole point. The real difficulties are real and well documented. The engineering of feeding a wind-and-sun supply that comes and goes into a grid built for the steady output of coal and gas. The transmission lines that have to be built, and the queues of projects that bank up behind them waiting to connect. The rise in cost that the market operator has recorded in real terms. The fact that the wind drops and the sun sets. The long approvals. And the resistance of communities to big infrastructure built near them. None of that is anyone leaning on the scale. It is the difficulty of a large change. And alongside it, the incentive factors are documented too. Fossil-fuel subsidy arrangements estimated in the billions a year. A policy framing that peer-reviewed analysis describes as favouring the fossil-fuel generators already in place. And the influence of the incumbent industry and of concentrated media ownership. That is where the word capture earns a hearing.

The question that would settle the strong story is which of these weighs more, and here I have to keep two things apart that a looser argument would smear together. That the incentive factors exist is established. The subsidy arrangements, the policy framing that favours the incumbents, the concentrated ownership, they are all on the documented record. That they exist is not in doubt, and it is enough on its own to show that partial capture is real here, a mechanism genuinely present in the system and not just alleged. What is not established is their weight against the real difficulties, which of the two is the bigger cause of the delay. The record I have does not let me settle that, so I will not pretend it does. To manufacture a verdict, to declare capture the main cause because it would make the more gripping chapter, would be to do the very thing this whole discipline exists to stop. So I claim what I can defend and no more. Partial capture is present and real, its weight is contested, and the contest is open. Stating the existence as a fact and the weight as unsettled is not a dodge. In this case, it is the argument.

The ledger, set out plainly

Gather the four cases together and a kind of stocktake appears. The stocktake is the finding.

The record establishes several things. The speed of adoption tracks where the benefit falls, not how good the method is: benefit that is concentrated and lands now gets taken up quickly, as in housing finance once the method became legible to it, while benefit that is spread out and delayed waits, as in preventive health. The system can see and can move: housing shows it now, and so does the in-the-room consultation category, reopened in days once it was finally asked, and largely kept afterwards. A crop the regulators call non-intoxicating is nonetheless governed under a narcotic category that treats its cannabinoid content as beside the point. And in renewables, the incentive factors are documented as present, so partial capture there is real as a mechanism, whatever its weight against real difficulty turns out to be.

The record knocks several things down too, and the corrections put right live misreadings rather than scoring points. The idea of regulator capture by incumbents everywhere does not explain these cases. It fits none of the four cleanly, and only renewables even partly. The generous home threshold on hemp does not hold back the growing of it. The safety of low-THC hemp does not rest on any claim about missing receptors. And the striking point that a stimulant sits in the narcotic category proves the label is legal rather than medical, which is true but modest, and not the proof of a mad rule it is sometimes asked to be.

And the record leaves things open, which is where the discipline shows. The weight of partial capture against real difficulty in renewables is unsettled, and I have left it there. So too, at the edge of the hemp case, is the exact cause of one long delay, which I come to now.

What you can now do

The tool is simple to state and demanding to use. Faced with any claim that the system will not take up the better thing, you can now ask which of the four kinds of waiting this is. Caution that is doing its job. A good that no one is paid to carry. A label inherited from another purpose. Or the partial capture where the weight is genuinely arguable. The fix is different in every case. Build the assurance, or line up the payment, or reopen the label, or contest the weight. Apply the fix for one kind of waiting to another kind, and the effort is not just wasted. It hands the doubters their proof that reform does not work.

Two facts sit at the close of the hemp case, and they carry their own labels rather than borrow a certainty they have not earned. Hemp as a food was banned in this country until late 2017. The food regulator’s own science had assessed low-THC hemp foods as safe and recommended they be approved by 2012, and the effort to get there stretched back to an application lodged in 1998. So the delay, close to two decades from that first application and a further five after the science was settled, is not explained by any unresolved question of safety. That much is established, and it is established by elimination, because the safety question had an answer and the answer was yes. What the delay tracked instead, the crop’s link with the drug strain and the whole apparatus of drug testing built around it, is a strong inference and not a fact of the same standing. The regulator did require that approved hemp foods keep clear of any link with the drug strain, down to the pictures on the packet, which fits that reading. But an ordinary account of bureaucratic caution and of getting the states and the Commonwealth to move together could play a part too, and I cannot close the gap between the elimination, which I can prove, and the attribution, which I can only strongly suspect.

The system, then, is not blind. Blindness was the wrong first reach, the single explanation the frustrated reformer grabs for, and the cases have earned the correction. The system sees what it is built and paid to see, and does not see what nothing rewards it for seeing, or what an old label has told it not to look at. The work is not to accuse it of blindness. The work is to tell the four kinds of not-seeing apart, because only then does the question of what to do about any one of them have an answer. There is a harder question underneath all of this, which is whether, beneath the noise of our disagreements, there is more settled agreement about what should be done than the system lets us believe. That is the ground the next chapter walks.

Walker Briefing is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.