In my previous post on monetary policy transmission asymmetries, I leaned too heavily on the mortgage repayment channel – the "cash-flow" story – when explaining why rate hikes transmit fast. Peter Tulip rightly called this out: the consumption research suggests the cash-flow channel is less important in aggregate than the public discussion implies.
He's right. And it's worth a more detailed treatment, because understanding which channels actually matter – and how much – is essential background for evaluating the alternative instruments people periodically propose for managing inflation.
The Main Channels
The RBA identifies four main transmission channels in its April 2025 Bulletin article on monetary policy transmission through the lens of its MARTIN and DINGO models (Mulqueeney, Ballantyne and Hambur 2025). Christopher Kent, as Assistant Governor (Financial Markets), added a fifth – the credit channel – in his October 2023 address to Bloomberg. These aren't exhaustive, but they're ones the RBA can quantify in its models. I'll come back to some others later.
1. The exchange rate channel. A rate rise increases the return on Australian-dollar assets, attracting foreign capital and pushing up the exchange rate. A stronger dollar makes imports cheaper (directly lowering inflation) and makes Australian exports less competitive (reducing net exports and demand). This channel is fast and powerful. In the RBA's MARTIN model, the exchange rate channel accounts for between one-quarter and two-thirds of the peak GDP response to a rate change, and an even larger share of the inflation response. This is the single most important channel for inflation in the RBA's modelling – a finding that often surprises people who associate monetary policy primarily with mortgages.
2. Asset prices and wealth. Higher rates weigh on asset prices – housing, equities, bonds. Falling asset prices reduce household wealth, which reduces consumption via the wealth effect. They also reduce the equity available for borrowing, tightening access to credit. Research by May, Nodari and Rees (2020) and Windsor, Jääskelä and Finlay (2015) confirms that Australian households do adjust consumption in response to changes in housing wealth. The housing sector in particular is a sensitive part of the transmission mechanism – dwelling investment responds strongly to rate changes, and the RBA's DINGO model shows housing notably increases the duration of the GDP response to monetary policy (Gibbs, Hambur and Nodari 2018). Put simply: higher rates depress housing's contribution to GDP for longer.
3. The savings and investment (intertemporal substitution) channel. Higher rates increase the return on saving and the cost of borrowing. This shifts the calculus between spending now and spending later – what economists call intertemporal substitution. With higher rates, households save more and consume less; businesses face a higher hurdle for investment. In structural New Keynesian models, intertemporal substitution via the real interest rate is the core transmission mechanism, with asset price and exchange rate movements operating partly through that same real-rate channel. In the RBA's empirical models, the asset prices channel and the savings/investment channel are hard to separate because they overlap significantly. Together they account for about one-quarter of the peak GDP response in MARTIN, with most of the effect flowing through dwelling investment and, over time, consumption.
4. The cash-flow channel. This is the one everyone talks about. Higher rates mean borrowers pay more on their debt and savers earn more on deposits. Because Australian households are net borrowers (interest-bearing liabilities exceed interest-earning assets), higher rates reduce aggregate household cash flow. And because borrowers tend to have higher marginal propensities to consume than savers, the net effect on spending is negative.
But here's the key finding: this channel is smaller than the public discussion implies. The RBA's MARTIN model suggests the cash-flow and related channels produce a consumption decline of only about 0.15 per cent in response to a 100 basis point rate increase. La Cava, Hughson and Kaplan (2016) found a similarly modest effect: 0.1–0.2 per cent of aggregate household expenditure. Updated RBA estimates (Jennison and Miller 2025) suggest the cash-flow effect hasn't changed substantially. As the April 2025 Bulletin puts it: "although the cash flow channel gains a lot of public attention and can have a large effect on individual households, it has a smaller role in aggregate transmission."
The reason is partial offset: higher rates hurt borrowers but help savers. The aggregate flows can be large for each group, but they partially cancel. The net economic effect is modest.
5. The credit channel. Changes in rates affect not just the demand for credit but its supply. When rates rise, banks tighten lending standards – borrowers become less creditworthy both directly (higher repayments) and indirectly (lower asset values reduce collateral). This amplifies the other channels. Brassil, Major and Rickards (2022) developed a banking-augmented version of MARTIN showing that the overall size of monetary policy effects can vary depending on the state of the economy – the credit channel introduces state-dependence into transmission.
The Relative Importance Hierarchy
Putting this together from the RBA's own modelling:
Exchange rate: dominant for inflation, very important for GDP. One-quarter to two-thirds of the peak GDP effect; an even larger share of the inflation effect. This is the workhorse channel, and it operates fast.
However, there is an important practical caveat. The exchange rate responds to relative interest rates – Australia's rates versus the rest of the world. When global tightening cycles are synchronised, as they were in 2022–2023, everyone hikes together and exchange rate movements are muted. The channel that MARTIN says is dominant can be largely neutralised in practice. Conversely, when Australia is out of step with global cycles – as it is now, being one of a small cohort of economies in a tightening phase (alongside Japan and Brazil) while most major central banks are holding or still easing – the exchange rate channel operates at full power. The Australian dollar appreciates (or holds up) relative to peers, compressing import prices and tightening trade competitiveness. So the model hierarchy is most informative precisely when Australia's cycle diverges from the global cycle, and least informative when cycles are synchronised. The current episode is one where the exchange rate channel should be doing heavy lifting – and that means the RBA may be getting more inflation-reducing work per basis point than it would during a synchronised tightening.
Asset prices + savings/investment: important, especially over longer horizons. About one-quarter of the peak GDP effect, with housing particularly sensitive. These channels are slower but persistent.
Cash flow: visible but small in aggregate. Around 0.1–0.2 per cent of household expenditure per 100bp of policy movement. Gets the most media attention relative to its actual macroeconomic importance. The RBA itself says it plays a "smaller role in aggregate transmission."
A note on intertemporal substitution and the gap between theory and evidence. In New Keynesian models, intertemporal substitution is the core mechanism – the real interest rate is the price of consumption today versus tomorrow, and households optimise across time accordingly. That said, the textbook picture of households smoothly shifting consumption across time in response to rate changes sits uneasily with the empirical evidence.
Estimates of the elasticity of intertemporal substitution (EIS) are all over the place, with many clustered near zero. Hall's (1988) foundational work found little evidence households meaningfully shift consumption in response to rate changes, and a meta-study by Havranek (2015) based on 2,735 estimates from 169 papers reported a distribution ranging from −5 to +5. More recent survey-based estimates (Crump et al., NY Fed; Christelis et al., ECB) are more plausible but still leave wide confidence bands. When researchers relax the textbook preference assumptions, the real-rate elasticity of output falls into the 0.05–0.20 range – far below what log-utility NK models implicitly assume.
This matters because it suggests the theoretical backbone of monetary transmission may be doing less work in practice than models imply. The intertemporal substitution channel probably matters more for investment – where hurdle rates and cost-of-capital calculations are explicit – than for household consumption, where habits, liquidity constraints, and inattention dominate. The channels that do appear to do heavy empirical lifting are the exchange rate, asset prices, and credit conditions. This is another reason why the channel hierarchy matters: if you're designing alternative instruments, you need to know which channels actually move the needle in the data, not just which ones are elegant in theory.
Credit: amplifying and state-dependent. Harder to quantify but acts as a multiplier on the other channels, especially in stressed conditions.
Who are the "constrained households"? The cash-flow channel story implicitly assumes mortgage holders, but it's worth asking who is actually cash-flow-constrained in Australia. The most constrained households are more likely to be lower-income renters – often with consumer debt – than mortgage holders, who are at least asset-rich on paper.
But renters experience monetary policy very differently. Their rent is set by housing supply and demand – vacancy rates, population growth, the construction pipeline – not by the cash rate. In a tight rental market, rents can be rising while rates are rising, squeezing renters from both directions without any offsetting wealth effect. And rate hikes may perversely push rents higher by slowing construction (dwelling investment is among the most rate-sensitive components of GDP in both MARTIN and DINGO), constraining future supply and keeping vacancy rates low. As for renters carrying credit card debt: credit card rates are already north of 20%, so a 25 basis point cash rate move is noise against that base. The risk premium swamps the policy signal.
The channel that actually reaches these households is the labour market: tighter policy cools demand, eventually softens hiring, and their hours or job security deteriorate. But that's the slowest channel of all – see the Okun persistence story from the previous post. So the people the public most associates with monetary policy pain are actually the people least directly affected by the rate instrument, and reached last, at the longest lag. This also reinforces the case against proposed alternatives: an SGC adjustment or tax lever would hit those same constrained workers directly through their payslip, concentrating the pain on the people least able to bear it while weakening the channels that do the actual macroeconomic work.
The popular narrative – "the RBA raises rates, mortgage holders suffer, spending falls" – captures one channel. It's not wrong, but it is perhaps the least important channel for the aggregate economy. The exchange rate, asset prices, and (for investment) intertemporal substitution do more of the heavy lifting.
The Expectations Channel
There is a channel that cuts across all the others and arguably matters more than any of them individually: expectations. Monetary policy works partly through signalling. A rate move – or even the credible expectation of one – changes expectations about future rates, future inflation, and the central bank's resolve. This is what makes the yield curve work: long-term borrowing rates embed expected future cash rates, and those expectations shift the moment the RBA signals a change in direction. It's also the mechanism behind foreshadowing – the reason the February 2026 hike tightened financial conditions before it was even delivered, as I argued in the previous post.
The expectations channel is what distinguishes a credible central bank from a mere interest rate setter. If firms and households believe the RBA will do whatever it takes to return inflation to the 2–3% band, they anchor their wage bargaining and pricing decisions to the target rather than to realised inflation. That anchoring does much of the disinflationary work without requiring the full demand destruction that would be necessary if expectations were unmoored. Lose credibility, and you lose this channel – which is why purposefulness matters so much, and why erratic policy is so costly. It's not just about looking competent; it's about preserving the single most powerful transmission mechanism the central bank has.
Other Channels Worth Noting
The channels above are the ones the RBA's models can quantify. But there are others that matter and deserve at least a mention.
The fiscal channel. Higher rates increase government debt servicing costs. Post-COVID, with a large stock of government debt, this is not trivial. In effect, higher rates put pressure on governments to contain the growth of other spending or accept larger deficits – a kind of indirect fiscal restraint that the RBA doesn't control but sometimes benefits from.
The risk-taking channel. Low rates push investors into riskier assets in search of yield; high rates pull them back toward safer ones. This affects financial conditions broadly – credit spreads, venture capital availability, asset allocation. It also links to productivity growth: easy monetary policy can sustain low-productivity firms that would otherwise exit ("zombification"), while tighter policy forces capital toward more productive uses. The relationship between monetary policy and long-run productivity is contested, but the channel exists.
The borrower-saver redistribution channel. Rate changes redistribute between borrowers and savers. Auclert (2017) argues that the aggregate demand effect depends on the difference in marginal propensities to consume across these groups. This is related to the cash-flow channel but conceptually distinct – it's about who gains and loses, not just the net aggregate effect. It also feeds a political economy dynamic that shapes public attitudes to monetary policy: many Australians perceive the RBA as primarily supporting asset-rich baby boomer retirees (who benefit from higher deposit rates and accumulated housing wealth) at the expense of younger borrowers. Whether or not that perception is fair, it affects the political sustainability of the institution – and it's a distributional consequence that the proposed alternatives (SGC, tax brackets) would not resolve, since they create their own distributional problems.
How the Overnight Rate Becomes the Rate That Matters
All of this presupposes something important: that the RBA's overnight cash rate actually transmits to the borrowing and lending rates the economy faces. This is not automatic. It requires a functioning yield curve anchored by confidence in the central bank's inflation target.
The mechanics are straightforward. The cash rate is the rate banks pay to borrow overnight from each other. Short-term money market rates track it closely – as Kent (2023) explained, if a bank faced a much higher one-month rate, it would simply borrow overnight and roll daily. So the cash rate propagates along the short end of the curve almost immediately.
Longer-term rates – the ones that matter for fixed-rate mortgages, business loans, and bond yields – embed expectations about the future path of the cash rate plus a term premium. When the central bank is credible, the term premium stays low because markets believe rates will return to a level consistent with the inflation target over time. The yield curve becomes a smooth bridge between today's policy rate and tomorrow's borrowing costs.
When credibility breaks, this bridge collapses. Long rates start reflecting the real cost of money plus an inflation risk premium the central bank doesn't control. You can cut the cash rate and long rates don't follow – or they rise. The overnight rate disconnects from the rates the economy actually faces. The UK experienced a version of this during the Truss mini-budget episode in September 2022: gilt yields spiked not because of BoE rate decisions but because markets lost confidence in the fiscal framework, which undermined confidence in the inflation path.
The Trump administration’s public pressure on the US Federal Reserve to lower rates regardless of the inflation backdrop is exactly the sort of behaviour the credibility literature warns about. If markets come to believe the central bank will subordinate its inflation mandate to political preferences, the expectations channel breaks down: long rates start pricing in inflation risk the central bank has chosen not to fight. Powell has so far held that line but his term expires in May 2026, and what follows will be a live test of the institution’s independence. This isn’t a partisan point; it’s a structural one. Any central bank that lets the electoral cycle dictate its reaction function will eventually lose the expectations anchor that makes the instrument work.
This has a direct bearing on the purposefulness argument from the previous post. The reason the RBA can move the cash rate by 25 basis points and have it ripple through to mortgage rates, business lending rates, bond yields, and the exchange rate is because the institution is credible. The transmission mechanism presupposes the credibility. Erratic policy doesn't just look bad – it literally degrades the instrument's effectiveness by widening the gap between the overnight rate and the rates that matter.
Purposefulness isn't institutional vanity. It's a structural precondition for the instrument to work.
Why the Proposed Alternatives Are Narrower, Not Broader
Understanding the full channel map makes it possible to evaluate the alternative instruments people periodically propose. The common thread: they are typically narrower than interest rates, not broader.
The superannuation guarantee charge (SGC) as a macro lever. Raise the SGC to cool demand, lower it to stimulate. This operates through exactly one channel: household disposable income. It's a forced saving rate adjustment. To put some scale on it: the SGC is currently 12% of ordinary time earnings. A 1 percentage point change shifts roughly 0.5% of GDP between household consumption and compulsory saving – a meaningful amount, but delivered through a single, narrow channel. There's no exchange rate effect, no asset price channel, no cost-of-capital signal to firms, no intertemporal substitution. The distributional incidence is also narrow – it only hits employed workers, and disproportionately lower-income ones already consuming most of their income. Retirees, the self-employed, and asset-rich households are untouched.
Tax adjustments tied to macro conditions. The idea has several versions, ranging from adjusting only the top marginal rate to moving all brackets simultaneously. Take the strongest version: an independent body – perhaps the RBA itself – with legislated authority to add or subtract basis points across all personal income tax brackets in response to macroeconomic conditions. The institutional design is solvable. You could model it on the existing central banking framework: regular meetings, published minutes, a clear mandate.
The economic case is more problematic. Even across all brackets, you're operating through a single channel – household disposable income. There's no exchange rate effect, no asset price channel, no cost-of-capital signal to firms, no intertemporal substitution beyond the direct income effect. You'd get a demand impact, certainly – especially from lower-bracket adjustments where marginal propensities to consume are highest – but it would be one-dimensional.
More importantly, you can't substitute this for interest rates. You'd still need rates for the exchange rate, asset prices, credit, and cost-of-capital channels – the channels the RBA's own research says do the heavy lifting. So the tax lever runs alongside rates, not instead of them. And then borrowers cop it twice: the rate hike and the tax adjustment. The people most exposed to mortgage costs – who motivated the proposal in the first place – end up bearing a heavier load than under rates alone.
The calibration problem is also difficult. If you do less on rates because you're also tightening via tax, you weaken the exchange rate, asset price, and cost-of-capital channels. If you keep rates the same and add tax on top, you're overtightening on household income relative to everything else. There's no obvious split that improves on rates alone.
There are also practical complications. Tax adjustments hit employed workers but miss retirees, the self-employed, those on government transfers, and capital income earners – significant portions of aggregate demand that rates reach but tax brackets don't. And the political economy is harder than it looks: even with institutional independence, an independent body that visibly raises taxes on ordinary workers' payslips concentrates the political cost in a way that rates – which arrive diffusely through mortgage statements, asset prices, and exchange rate movements – do not. That indirection isn't a bug in monetary policy; it's what makes institutional independence politically sustainable over decades.
Business tax surcharges. A temporary business tax increase intended to cool demand. But this is self-defeating: firms treat it as a cost increase and pass it through to prices. Your anti-inflationary instrument becomes inflationary. You're fighting demand-pull inflation by creating supply-push inflation – feeding both sources at once. This circles back to the "two sources of inflation, one instrument" problem from the previous post: the interest rate works on demand without creating supply-side inflation. A business tax lever does the opposite.
The Additionality Problem
The common thread across all these proposals is the additionality problem. None of them can replace interest rates – they can only supplement them. The exchange rate channel, the asset price channel, the cost-of-capital signal, and the credit channel are all unique to monetary policy. No fiscal lever touches them. (Macroprudential tools – APRA's loan-to-value caps, serviceability buffers, and the like – are sometimes raised as an alternative, but they target housing credit and asset prices to protect financial stability, not consumer prices; since house prices aren't in the CPI, these tools complement rates by managing systemic risk rather than substituting for them as inflation instruments.)
So any new instrument runs alongside rates, not instead of them. The combined effect of two demand-tightening instruments hitting simultaneously is harder to calibrate, harder to communicate, and distributes the burden less evenly than rates alone. The people the proposals are typically designed to help – mortgage holders bearing the visible cost of rate moves – end up bearing that cost plus the fiscal tightening.
Why Interest Rates Survived
There's a reason monetary policy ended up as the primary demand management tool, and it's not because nobody thought of alternatives. It's because the alternatives either operate through fewer channels, hit the wrong end of the marginal propensity to consume distribution, create coordination problems with the instrument you can't give up, concentrate the political cost in ways that make institutional independence unsustainable, or – in the case of business taxes – are actively self-defeating.
Interest rates survive as the primary instrument because of a rare alignment:
Economic breadth: they touch the exchange rate, asset prices, saving/investment decisions, cash flows, and credit supply simultaneously. No other single instrument has this reach.
Political viability: the transmission is diffuse. The pain arrives on mortgage statements, in gradually falling house prices, in a slightly stronger dollar – spread across the economy rather than concentrated on a payslip. This indirection is politically load-bearing. It allows an independent institution to make unpopular decisions without immediate electoral consequences.
No supply-side inflation: unlike business tax instruments, rate changes do not mechanically raise firms' marginal production costs in the short run. They cool demand without creating cost-push inflation. The instrument fights one source of inflation without feeding the other. (Over longer horizons, sustained high rates can reduce supply via lower investment – but that's a second-order effect, not the direct cost pass-through that makes business tax surcharges self-defeating.)
Yield curve architecture: a credible central bank's overnight rate propagates through the entire term structure of interest rates, reaching every corner of the financial system. This amplification mechanism is unique to monetary policy and depends on the credibility that purposeful policy-making sustains.
The popular critique of interest rates focuses on the one channel – mortgage pain – that the research says matters least in aggregate. And the proposed alternatives typically only operate through that same narrow channel. That's the irony worth understanding: the thing people dislike about rates is their breadth, but that breadth is exactly what makes them work.
A defender of fiscal alternatives might argue that tax changes have higher fiscal multipliers than rate changes. Even if that's true for some instruments in some conditions, it doesn't resolve the fundamental problem: fiscal multipliers operate through narrower channels and lack the exchange-rate, asset-price, and cost-of-capital amplification that monetary policy uniquely provides. A high multiplier on a single channel is still a single channel.
The reason interest rates dominate macroeconomic stabilisation isn't ideological inertia. It's because they are the only instrument that simultaneously moves relative prices across borders (the exchange rate), across time (intertemporal substitution), and across asset classes (wealth and portfolio rebalancing). No proposed alternative touches all three.
No comments:
Post a Comment