The United States is approaching one of the most important monetary-policy turning points since the inflation shock began. When the Federal Reserve meets on 17–18 September, financial markets overwhelmingly expect it to cut interest rates for the first time since 2020, bringing an end to the aggressive tightening cycle that took the federal-funds target from effectively zero in early 2022 to 5.25–5.50%. The reason is not that inflation has disappeared. US consumer prices were still 2.5% higher than a year earlier in August, above the Fed’s 2% objective. What has changed is the balance of risks. Inflation has fallen dramatically from its 9.1% peak in June 2022, while unemployment has risen from extraordinarily low post-pandemic levels and hiring has cooled. The central bank is therefore confronting the problem that makes monetary policy much harder than simply “raise rates when inflation is high”: how do you know when enough economic pain has already been created to return inflation towards target without continuing until the cure produces a recession?
The inflation decline is important because it challenges two of the strongest narratives that dominated economic debate in 2022. One predicted that inflation would remain entrenched unless the United States experienced a severe recession and much higher unemployment. The other argued that inflation was almost entirely temporary and would disappear once pandemic supply disruptions normalised, making aggressive monetary tightening unnecessary. The outcome so far sits somewhere between them. Supply chains recovered, energy-price shocks moderated and pandemic distortions faded, helping inflation decline substantially. But the Federal Reserve also tightened financial conditions dramatically, weakening housing demand, credit growth and other interest-sensitive expenditure. Disinflation therefore appears to have come from both supply normalisation and demand restraint rather than validating either extreme explanation completely.
This matters because the economic cost has so far been much smaller than many conventional models feared. US real GDP continued expanding, unemployment remains low by historical standards and the economy has avoided the recession repeatedly predicted during the tightening cycle. That is what economists mean by a soft landing: inflation falls towards target without the economy experiencing a major contraction. The concept sounds simple but is extremely difficult operationally because monetary policy works through long and uncertain lags. When the Fed raises rates today, the full effect on mortgages, corporate investment, hiring and consumption can take months or years to appear. Policymakers therefore need to make decisions partly according to economic conditions that do not exist yet.
This is essentially control theory under delayed feedback. Imagine steering a ship where the wheel responds several minutes after being turned. If the captain waits until the ship visibly reaches the desired direction before straightening the wheel, the vessel overshoots. Central banks face the same problem. Interest rates may already be slowing the economy considerably before unemployment data reveal it. Waiting until inflation actually reaches exactly 2% before cutting rates could therefore mean maintaining restrictive policy for too long and producing unnecessary unemployment afterwards.
The Fed’s current problem is consequently different from the one it faced in 2022. Then, the dominant risk was under-tightening: allowing inflation expectations to become entrenched and forcing even more painful adjustment later. Now the risk is increasingly over-tightening. Interest rates remain at their highest level in more than two decades even though inflation is far below its peak. The same nominal rate therefore becomes more restrictive as inflation declines because the approximate real interest rate, nominal interest minus expected inflation, rises. A 5.5% policy rate when inflation is 9% is very different economically from 5.5% when inflation is around 2.5%.
This is why monetary policy can tighten automatically even while the central bank does nothing. Falling inflation raises real borrowing costs unless nominal rates fall alongside it. Maintaining the existing rate therefore is itself an active policy choice rather than neutrality. The Fed now needs to determine how quickly to reduce that restraint without reigniting demand sufficiently to stop disinflation.
The labour market provides the strongest argument for beginning. US unemployment has risen to around 4.2%, compared with 3.4% at its post-pandemic low, while job creation has slowed and vacancies have declined substantially from their 2022 peak. Yet unemployment remains historically low and layoffs have not exploded. This suggests much of the labour-market adjustment has occurred through fewer vacancies rather than mass job destruction. That is close to the ideal outcome policymakers hoped for: companies stop competing so aggressively for scarce workers, wage pressure moderates and inflation falls without millions of people becoming unemployed.
This matters for the debate surrounding the Phillips curve, the relationship between labour-market tightness and inflation. A simplistic interpretation suggests reducing inflation requires raising unemployment substantially. The post-pandemic period demonstrated that the number of job vacancies relative to unemployed workers had become exceptionally high, creating the possibility that labour demand could fall considerably before firms needed to begin large-scale layoffs. The economy could therefore move down the vacancy curve before moving dramatically up the unemployment curve.
If that interpretation survives, it would be one of the most important lessons from this inflation episode. The cost of disinflation depends on where labour-market imbalance begins. An economy with ten vacancies for every available worker can reduce demand substantially before unemployment rises as much as an economy beginning closer to equilibrium. That means historical relationships between inflation and unemployment cannot simply be applied mechanically without considering the structure of the shock.
The original inflation itself was similarly unusual. COVID simultaneously disrupted supply and transformed demand. Households shifted spending from services towards goods, governments provided enormous fiscal support, supply chains struggled to respond and labour markets experienced unusual exits and reallocations. Russia’s invasion of Ukraine then intensified energy and food pressures. Inflation therefore emerged from several shocks interacting, not one conventional demand boom.
That is why the argument over whether Biden’s American Rescue Plan caused inflation has always been too binary. The $1.9 trillion package almost certainly increased aggregate demand and therefore contributed to price pressure, particularly because it arrived when the economy was recovering rapidly and households had already accumulated substantial fiscal support. But American fiscal policy cannot explain why inflation accelerated simultaneously across many economies that did not implement identical stimulus. Supply constraints, energy markets and global reopening clearly mattered too.
Good macroeconomic analysis therefore asks for marginal contribution, not ideological exclusivity. Fiscal stimulus contributed. Supply disruption contributed. Monetary accommodation contributed. Energy shocks contributed. Labour-market restructuring contributed. The proportions changed over time.
The same principle should govern credit for disinflation. Biden cannot plausibly claim that administration policy alone brought inflation down. Powell cannot plausibly claim the Fed did it alone. Supply chains normalised independently. Energy prices changed. Firms adjusted production. Workers returned to employment. Monetary policy reduced demand. Fiscal deficits evolved. Macroeconomic outcomes are jointly produced.
Politics dislikes this because elections reward simple attribution.
The November presidential election makes the problem especially visible. Democrats can point towards strong growth, low unemployment and falling inflation as evidence that the economy has performed remarkably well relative to predictions. Republicans can point towards the cumulative increase in the price level since 2021 and argue that households remain materially worse off in the prices they encounter every day.
Both claims can be true.
This is the distinction between inflation and the price level. If inflation falls from 9% to 2.5%, prices are still rising; they are simply rising more slowly. The price increases accumulated during 2021–23 do not reverse unless the economy experiences deflation. A grocery item that rose from $4 to $5.50 does not return to $4 because inflation normalises.
This explains why public economic sentiment can remain weak even after economists celebrate disinflation. Households remember the previous nominal price as a reference point. Wages may eventually catch up in real terms, but the psychological experience remains: everything became expensive.
Central banks do not normally try to reverse that price-level increase because sustained deflation can be economically damaging. The objective is stabilising the rate of future increase, not restoring 2019 prices. That creates a permanent political problem after inflation shocks: successful monetary policy stops additional damage without undoing the visible evidence of previous damage.
Housing intensifies the frustration. Higher interest rates reduce inflation partly by making borrowing more expensive, but this means households trying to buy homes face mortgage rates dramatically above pandemic-era levels. A policy designed to stabilise prices can therefore make one of the largest household purchases less affordable in the short term.
This is a distributional effect of monetary policy. Savers can benefit from higher yields. Existing homeowners with fixed low-rate mortgages are partially insulated. First-time buyers suffer from high financing costs. Highly indebted firms suffer more than cash-rich firms. Central banks use one national interest rate, but households experience it very differently.
Rate cuts will begin reversing some of those effects, but they should not be confused with a return to the near-zero-rate world that followed the global financial crisis. Markets expect a gradual easing cycle, not necessarily an immediate restoration of extraordinarily cheap money. The neutral interest rate itself may be higher than before COVID because fiscal deficits, investment demand, demographics and global capital patterns have changed.
This creates another important question: what is the normal interest rate in the post-pandemic economy? Nobody observes the neutral rate directly. Economists infer it from behaviour. If the Fed cuts too far below neutral, demand can accelerate and inflation return. If it remains too far above neutral, unemployment rises unnecessarily.
Central banking therefore is not simply data-driven in the sense politicians sometimes imply. Data do not announce the correct interest rate. Policymakers interpret data through models containing uncertainty. Two competent economists can observe the same unemployment and inflation numbers and recommend different policies because they hold different estimates of neutral rates, lags and inflation persistence.
This is why central-bank independence matters. Monetary decisions create immediate political winners and losers. A government approaching an election has obvious incentives to prefer lower rates because cheaper credit can support asset prices, investment and household sentiment. If politicians controlled interest rates directly, markets would reasonably suspect policy was being manipulated around electoral calendars.
The Fed’s independence therefore functions as a commitment device. Politicians surrender some short-term control to an institution with a narrower mandate so households and investors believe inflation policy will not change according to electoral convenience.
Independence does not mean the Fed is apolitical in consequences. Interest rates affect employment, wages, mortgages, government debt service and asset prices. The institution makes choices with enormous distributional effects. Independence means those choices are supposed to be made according to legislated macroeconomic objectives rather than the electoral interests of whichever party controls the White House.
That distinction will become particularly important if Donald Trump returns to office. Trump has repeatedly criticised Powell and previously argued that presidents should have greater influence over monetary policy. A serious attempt to subordinate the Fed politically could increase inflation expectations because markets would begin pricing the possibility that rates remain too low when tightening becomes politically inconvenient.
Institutional credibility therefore has direct economic value. If households believe the Fed will ultimately return inflation to 2%, wage negotiations and price-setting behaviour remain anchored. Firms have less reason to raise prices pre-emptively. Workers have less reason to demand wages assuming permanent high inflation. Expectations make the central bank’s job easier when credibility is high.
This may be one reason the inflation shock did not become a repeat of the 1970s. Long-term inflation expectations remained comparatively contained despite the enormous short-term price surge. The Fed therefore did not need to recreate the extreme unemployment associated with Paul Volcker’s early-1980s tightening to convince Americans that inflation would eventually fall.
Credibility accumulated over decades became a form of policy capital.
Like all capital, it can be spent.
The Fed initially described inflation as “transitory” for too long, damaging some credibility when inflation persisted. Its subsequent aggressive tightening demonstrated willingness to accept economic risk to restore price stability. The soft landing, if it survives, would rebuild much of that reputation.
The phrase “if it survives” matters. Declaring victory before the lagged effects of tightening are fully visible would repeat the forecasting confidence that produced errors earlier in the cycle. Unemployment could rise faster over the coming months. Inflation could become sticky above target. Energy shocks could return. Financial vulnerabilities created by high rates could surface.
Macroeconomic success should therefore be assessed probabilistically rather than ceremonially.
The Fed appears much closer to a soft landing than many economists expected two years ago.
It has not completed one until the economy actually stabilises around low inflation and sustainable employment.
The next decision therefore is not whether inflation has been defeated.
It is whether monetary policy should begin moving from restriction towards normalisation before restriction itself becomes the next shock.
That is the difficult part of central banking.
Raising rates when inflation is 9% eventually becomes politically obvious.
Cutting them while inflation remains above target and unemployment remains relatively low requires much more judgement.
Move too early and demand may accelerate again.
Move too late and workers lose jobs unnecessarily.
There is no indicator that tells policymakers the exact day the balance changes.
That is why the September rate cut, if delivered next week as expected, will matter beyond the number of basis points involved.
It will represent the Federal Reserve saying that the dominant macroeconomic risk has changed.
For two years, the central question was how much demand needed to be destroyed to restore price stability.
The question now is how quickly that restraint can be removed without destroying the stability it created.
If the Fed manages that transition while unemployment remains near historical lows, the soft landing will deserve to be considered a genuine policy achievement.
But the most important lesson will not be that central bankers discovered a perfect formula.
It will be almost the opposite: economic shocks are heterogeneous, policy works with uncertain lags, and good institutions need enough credibility to change direction when the balance of risks changes rather than remaining trapped by the strategy that was correct six months earlier.
