Same Data, Three Directions

The Reserve Bank held. That was the least interesting part of the story.

Within hours, CBA’s economics team was pointing to a slowing economy. ANZ was tipping price falls. Same board decision, same data releases, same short stretch of days: two of the country’s biggest banks reading the road ahead in different directions. Nobody was obviously wrong. Nobody agreed either.

It isn’t just Australia. Capital Economics is now describing the Fed, the ECB, the Bank of Japan and the Bank of England as genuinely “out of sync” in 2026, moving in different directions in response to the same global pressures. Goldman Sachs has the US growing at 2.6% this year. Citigroup has it at 2.1%. Both are serious institutions with serious modelling. Both can’t be right, and there’s no way to know yet which one is.

If you run a business, sit on a board, or are trying to work out whether now is the time to expand, hire, hold or wait, you have just quietly inherited a decision nobody flagged you’d inherited. The people you delegated the “read the economy” job to have stopped agreeing on the answer, and nobody rang to tell you the job is now yours.

The quiet handoff

Most leaders have never consciously decided to delegate economic judgement to the RBA, the big four banks, Treasury, or whichever research house their industry association quotes. It happens by default. Someone senior reads the announcement, forms a view, and that view becomes the planning assumption for the next twelve months. Nobody in the room asks who granted that institution the authority to make the call. It was simply the loudest, most credentialled voice in the room, and credentialled voices are supposed to agree with each other, or at least be close enough that the gap doesn’t matter.

This is what I call a Decision Trust Zone: the invisible boundary around who is actually trusted to make a given call, and on what basis. I wrote about the concept in detail last year, in a piece asking a related question about a different kind of authority: what are decision trust zones, and where do leaders draw the AI line. The AI version of the question gets more airtime right now. But the human version, who do we trust to interpret an uncertain world, and what happens when that trust turns out to be split three ways, is older, and right now it’s the one actually costing people sleep.

Here’s the part that doesn’t get said out loud in most planning meetings: the zone doesn’t dissolve when the institutions disagree. It doesn’t go quiet, or get suspended until consensus returns. It moves.

It drops, without a memo, onto whoever in the room has to act anyway: the CFO who still has to file a budget by the end of the month, the board that still has to approve or defer a capital spend at its next meeting, the household that still has to decide whether to fix a mortgage rate before the next data release. Nobody tells them the zone has landed on their desk. They just find themselves holding a decision that used to belong to someone with a bigger title and a bigger research team.

Why this isn’t new, and why it matters more now

This has happened before, and the last time it mattered enormously, it happened quietly enough that most people didn’t notice until it was too late to matter.

In the years before the 2008 financial crisis, credit rating agencies were assigning top-tier ratings to mortgage-backed securities that a smaller number of analysts inside and outside those same institutions were flagging as dangerously overpriced. Both groups had access to broadly the same underlying data. One group’s institutional incentives and models said the risk was manageable. Another group’s judgement said it wasn’t.

For years, the disagreement stayed inside professional circles, resolved in practice by simply following the rating, because the rating was the institution everyone had delegated the decision to, and questioning it felt like more work than trusting it.

We know how that ended. The point isn’t that today’s central bank divergence is heading for a crisis of that scale. It almost certainly isn’t, and false equivalence between a rate-hold disagreement and a systemic ratings failure would be its own kind of bad judgement. The point is narrower and more useful: disagreement between trusted institutions is not new information about the world. It’s new information about how much weight your own organisation has been putting on someone else’s certainty. The 2008 story isn’t a warning about central banks. It’s a warning about outsourcing the entire judgement, rather than the input, to whichever institution has historically been loudest.

What’s different now is speed. A rating agency’s confidence took years to be tested against reality. A central bank’s confidence, or lack of it, is tested every six weeks against the next data release, in public, with three or four competing institutions publishing their own read within days of each other. The disagreement used to be private and slow. Now it’s public and fast. That should be good news. It’s much harder for a single institutional blind spot to sit unchallenged for years. But it also means the gap between “the experts agree” and “the experts are visibly split” can open in a single news cycle, and most organisations’ planning processes were never built to notice, let alone respond to, a gap that opens that quickly.

What actually moves when the institutions split

A decision made under a false assumption of consensus doesn’t stay contained to the room where it was made. It travels.

A board approves a hiring freeze based on one bank’s slowdown call, while a competitor down the road approves an expansion based on another bank’s growth call, using the same underlying data. Neither board has done anything unreasonable. Both are now running materially different strategies for materially the same set of facts, and neither one will find out which read was closer to right for another two or three quarters, by which point the freeze has already cost the slow-mover its best people, or the expansion has already left the optimist overextended.

A CFO sets next year’s cost base against a forecast that quietly assumes one institution’s number will hold, without ever stating that assumption out loud in the board paper. When the number moves, the budget doesn’t get revisited. It gets defended, because revisiting it means admitting the original number was never as solid as the paper made it sound.

A household, and it’s worth remembering that behind every one of your customers, employees and suppliers there’s a household doing exactly this same calculation, decides whether to fix a mortgage rate based on whichever bank’s commentary they happened to read that week. That’s not naivety. Most people have no framework at all for what to do when the experts they’d normally defer to are visibly split, so they default to whichever voice they encountered first, and treat it as though it were consensus.

None of this is anyone behaving badly. It’s what happens by default when a Decision Trust Zone quietly changes shape and nobody in the organisation was ever told to watch for the change. This is the pattern I described in a piece last month about AI verification overhead, what I called the Checking Tax, except the checking tax there was about people re-verifying an AI system’s confident output. Here it’s the same tax, levied by three confident institutions instead of one confident system. The bill still lands on whoever has to act, not on whoever produced the number.

The economy is also a mood

None of this is purely a modelling problem, either.

An economy doesn’t just get measured. It gets believed into being, at least partly, by the collective mood of everyone inside it. If enough people decide to spend cautiously because they’ve absorbed a story about hard times ahead, hard times tend to follow, not because the fundamentals demanded it, but because enough people acted as though they did. The economist Robert Shiller built an entire body of work, narrative economics, around exactly this: the stories a population tells itself about money move the numbers, not only the other way round.

That’s part of what’s sitting underneath this particular split. Three institutions aren’t only reading spreadsheets differently. They’re also making different judgement calls about how much weight to give the public mood itself, about whether the caution people are feeling is a leading indicator or a lagging one, about whether what they’re seeing is data or narrative dressed up as data. I wrote at length recently about the low-grade global mood almost everyone has been sitting in without quite having a name for it, the funk, and why it isn’t permanent or random, and the same mechanism is doing quiet work here. When the thing you’re forecasting is partly made of belief, disagreeing about the forecast is also, whether anyone says so or not, disagreeing about how much to trust the belief.

That doesn’t make the numbers meaningless. It makes them one input into a system that’s also, always, partly psychological, and it’s a reasonable part of why three well-resourced institutions can look at the same release and walk away thinking different things without anyone being incompetent.

What Decision Trust Zones actually gives you

The framework isn’t a call to distrust institutions, and it isn’t a licence to ignore expert forecasting altogether and run on gut feel. Both of those are just different ways of avoiding the actual work.

What it gives you is a simple, uncomfortable question to ask before the next planning cycle locks in: which decisions in this organisation currently rest on the assumption that a specific external institution’s judgement is reliable, and does anyone actually own the call if that institution turns out to be wrong, or if two institutions disagree?

Most organisations, when they actually run this exercise, find the honest answer is nobody. The economic forecast sits in a slide in the strategy deck, cited without a source, treated as background fact rather than as one institution’s judgement among several plausible ones. Nobody owns it because nobody ever formally accepted ownership of it. It arrived by default, the way most Decision Trust Zones do, and it will keep sitting there uninterrogated until the gap between forecasts becomes too large to politely ignore.

This is the same underlying gap I keep finding in client work, across sectors that have almost nothing else in common. A corporate organisation weighing whether to trust an internal AI tool’s output over a senior analyst’s contrary read. A tertiary education provider deciding whether to follow a national enrolment projection or its own more cautious regional data. A local government weighing a state-level growth forecast against what its own planning team is actually seeing on the ground.

Three different sectors, three different kinds of “expert,” and the same missing piece every time: nobody had pre-agreed, before the disagreement arrived, whose number would win. I explored the AI-adoption version of this same governance gap in a recent piece on who is actually deciding when AI is in the room. The mechanism is identical whether the second voice in the room is an algorithm or a rival bank’s chief economist.

The test, before the next number lands

Not a forecast. Not a prediction about which bank is right this time. I have no special insight into that, and anyone claiming otherwise is selling something. What’s useful here isn’t a better forecast. It’s a better process for the moment the forecasts split.

Before your organisation’s next planning cycle, budget round, or board paper, it’s worth running one short exercise. Take the two or three numbers your current plan quietly assumes, the rate outlook, the growth outlook, whatever external judgement is sitting uncredited in your model, and ask, in the room, out loud: if this number turns out to be wrong within the next two quarters, who notices first, who decides what changes, and how long does that decision take once someone does notice? If the honest answer is “nobody’s really watching for that” or “we’d probably just keep going,” you’ve found your Decision Trust Zone. It’s been sitting there the whole time. It just took three disagreeing banks to make it visible.

That’s not a criticism of the RBA, CBA or ANZ, any more than it would be a criticism of the Fed or the ECB. Institutions built to interpret genuine uncertainty are supposed to sometimes land in different places. That’s not a failure of the system. It’s the system doing exactly what a system built for judgement under incomplete information is meant to do. The failure, when there is one, sits closer to home: in planning processes that were built on an assumption of institutional consensus that was never actually guaranteed, and that quietly stop working the moment consensus breaks.

Choose Forward

The organisations that come through this year’s forecasting split in reasonable shape won’t be the ones that guessed correctly which bank was right. They’ll be the ones who noticed, early, that the zone had moved, and who had already worked out, before the gap opened, whose job it was to decide what happens next.

The Reserve Bank held. Somewhere in your organisation, a decision that used to belong to someone else just quietly became yours. It’s worth finding out before it finds you.

Choose Forward.