Inflation, Honestly

What this article does

Inflation is the most politically loaded number in macroeconomics. Politicians blame each other for it. Pundits weaponise it. Most people experience it but couldn't define how it's measured. And the gap between what "inflation" means in headlines vs how the Fed actually measures it creates years of misunderstanding.

This article fixes that. By the end:

• You'll understand what CPI is and how it's actually measured
• You'll know the difference between CPI and PCE — and why the Fed targets PCE not CPI
• You'll understand "core" inflation and why excluding food and energy actually makes it MORE useful, not less
• You'll know what shelter costs do to inflation persistence (a lot)
• You'll understand why the post-COVID inflation spike happened and why it's been so hard to bring fully under control
• You'll be able to read inflation reports without falling for either side's political spin

This is article 5 of 9 in the Lighthouse foundation series. The Fed targets inflation; understanding what they're targeting makes everything else click.

What inflation actually is

Inflation is the rate at which the general price level rises over time. If a basket of goods cost £100 last year and costs £103 this year, you have 3% inflation.

Sounds simple. Two devils hide in the details:

Devil 1: Which basket? Different households buy different things. Petrol is 5% of the average household budget but 15% for someone who drives 30,000 miles a year. Healthcare is 8% of average spending but 25% for retirees. Whose basket counts?

Devil 2: How do you handle change? A basic mobile phone in 2010 cost £100. A basic mobile phone in 2025 costs £100. But the 2025 phone has features the 2010 one couldn't dream of. Did the price stay the same? Or did it effectively fall (more value for the same money)?

Statisticians have to make calls on these. The two main US inflation measures handle them differently.

CPI vs PCE — why we have two

The two main US inflation measures are CPI (Consumer Price Index, published by the Bureau of Labor Statistics) and PCE (Personal Consumption Expenditures, published by the Bureau of Economic Analysis). The Fed officially targets 2% Core PCE.^[1] The two measures are:

CPI (Consumer Price Index) — published by the Bureau of Labor Statistics monthly. Surveys ~80,000 prices across 23,000 retail outlets in 75 urban areas. Uses a fixed basket weighted to what urban consumers spend on. Updated every 2 years.

PCE (Personal Consumption Expenditures Price Index) — published by the Bureau of Economic Analysis monthly. Tracks prices weighted by what consumers ACTUALLY spend on. Updates weights more dynamically. Captures rural and business spending CPI misses.

CPI is the more famous measure. PCE is the Fed's preferred measure.

Why does this matter? Because they don't agree. PCE typically runs 0.3-0.5 percentage points lower than CPI on average due to differences in weighting methodology, scope of goods, and how substitution behaviour is captured.^[2]

A typical year might see:

• CPI year-over-year: 3.2%
• PCE year-over-year: 2.8%

Same economy, different methodology, different number. When you see "inflation hit 4%" or "inflation cooled to 2.5%" in a headline, the journalist might be using either measure. Always check which.

The biggest reason for the gap: PCE captures substitution behaviour (when beef gets expensive, consumers buy chicken; the basket weight shifts) while CPI assumes the basket stays the same for 2 years.

The Fed targets 2% Core PCE specifically. The Fed adopted a formal 2% inflation target in January 2012, defining 'stable prices' operationally. Before that, the Fed had no explicit numerical target, though 2% had been informally understood as the goal since the Greenspan era.^[3] Not 2% CPI. So when the Fed says "inflation is moving toward target" they're talking about a number that's typically 0.3-0.5pp lower than the CPI you've been seeing in headlines.

What "core" actually means

Most inflation discussions mention core inflation — and a lot of people misunderstand why it exists.

Core inflation excludes food and energy prices. People often hear this and think: "But food and energy ARE inflation! Excluding them is rigging the numbers."

That's not what core does. Core inflation excludes food and energy prices because they are extremely volatile. Federal Reserve research has shown core inflation is a better predictor of where headline inflation will be in 6-12 months than headline inflation itself.^[4] Here's the genuine reason:

Food and energy prices are extraordinarily volatile. Oil can swing 20% in a month on a Middle East crisis. Wheat can swing 30% on a drought in Ukraine. Gas prices spike on hurricanes. These swings are usually:

1. Driven by external shocks (geopolitics, weather, supply disruptions) 2. Largely unaffected by Fed policy (raising rates won't end a war) 3. Often reversed (oil spikes today, settles back in 6 months)

Including these in the headline number means the headline can spike 2 percentage points on a single oil shock and then drop 2 percentage points when the shock reverses. That's noise, not signal.

Core inflation strips out the noise. What you're left with is the persistent, broad-based inflation that:

1. Is driven by domestic demand and labour markets 2. CAN be affected by Fed policy 3. Tends to be stickier and harder to reverse

Federal Reserve research has consistently shown core inflation is a better predictor of where headline inflation is going than headline inflation itself. Core is not a politically convenient number — it's a more useful one for forecasting.

The shelter problem

There's one massive component of inflation that deserves its own treatment: shelter costs.

Shelter costs (rent and owners' equivalent rent) make up approximately 33% of the CPI basket and 16% of the PCE basket. Because shelter prices change slowly, this category is a major driver of why inflation is sticky and slow to fall.^[5] Shelter (rent + owners' equivalent rent) makes up roughly 33% of the CPI basket and 16% of the PCE basket. That's huge — bigger than any other category.

And here's the issue: shelter is a sticky prices category. Rents are typically annual contracts. Mortgages are fixed for years. Property values change slowly. When inflation spikes for other reasons, rents take 12-18 months to catch up. When inflation cools elsewhere, rents take 12-18 months to follow.

This means shelter inflation often LAGS the broader cycle by over a year. When you see headlines like "core inflation refuses to come down," it's often because shelter inflation is still working its way through, even after every other category has normalised.

In 2022-2023, when overall inflation was spiking, shelter inflation lagged. Then in 2024, even as overall inflation had cooled, shelter remained sticky high — keeping core PCE above the 2% target longer than expected.

This is the genuine reason the Fed has been so cautious about declaring victory on inflation. The shelter category alone can keep core PCE elevated for a year after the underlying inflation problem is solved.

The 2022 spike — what actually caused it

The June 2022 CPI reading of 9.1% was the highest US inflation rate since November 1981. The post-COVID inflation spike was driven by supply chain disruptions, fiscal stimulus, energy shocks from the Russia-Ukraine war, and pent-up demand from reopening.^[6] The June 2022 CPI reading of 9.1% was the highest US inflation rate since November 1981. That's 4 decades.

Why? Multiple causes converged:

1. Massive fiscal stimulus. Three rounds of stimulus checks (March 2020, December 2020, March 2021) plus expanded unemployment benefits put trillions of dollars into household balance sheets while supply was constrained. Pent-up demand met limited supply.

2. Supply chain breakdown. COVID disrupted manufacturing and shipping for years. Container shipping costs spiked 800%. Semiconductor shortages cascaded through industries. Just-in-time inventory systems failed under stress.

3. Energy shock. Russia invaded Ukraine in February 2022. Oil prices spiked above $120. Natural gas prices in Europe rose 10x. Energy is a major input to almost everything else.

4. Labour market shifts. The "great resignation" tightened labour markets. Wage growth picked up. Service-sector inflation (which is largely wages) accelerated.

5. Fed slow to respond. The Fed maintained near-zero rates through 2021, calling inflation "transitory." When they finally hiked in March 2022, they were already months behind.

These factors compounded. Headline CPI peaked at 9.1% in June 2022. Core CPI peaked slightly later at 6.6% in September 2022. The gap is the food and energy story — headline ran higher because of the energy shock.

The journey back to 2%

US Core PCE inflation peaked at 5.6% in February 2022 and gradually declined to approximately 2.8% by late 2024 — still above the Fed's 2% target. The 'last mile' of bringing inflation from 3% to 2% has historically been the hardest.^[7] Core PCE peaked at 5.6% in February 2022 and gradually declined to about 2.8% by late 2024. Still above the Fed's 2% target.

The "last mile" — bringing inflation from 3% to 2% — has historically been the hardest part of disinflation. Several reasons:

Shelter persistence. Shelter inflation continues to lag. Even after market rents stabilise, the official measure takes 12+ months to catch up.

Service inflation. Services (restaurants, healthcare, education) tend to have more wage embedded, and wages don't fall easily.

Wage-price spiral risk. Workers, having seen inflation, demand higher wages. Higher wages get passed through to higher prices. The cycle continues. The Fed has to break this dynamic without breaking the labour market.

Inflation expectations. Inflation expectations are tracked through both market-based measures (TIPS breakeven rates, 5y5y forwards) and survey-based measures (University of Michigan consumer survey, NY Fed consumer expectations survey). The Fed monitors both because anchored expectations are critical to inflation control.^[8] Once people EXPECT inflation, they negotiate accordingly. Long-term inflation expectations becoming "unanchored" is the Fed's nightmare scenario. Anchored expectations (people still believing 2% is the long-term norm) is critical to disinflation.

The Fed's challenge in 2024-2025 has been finishing the job without overshooting. Cut too soon → inflation re-accelerates. Hold too long → unnecessary recession. This is genuinely hard, and people who are confident the Fed is "obviously" doing it wrong usually don't appreciate the trade-offs.

How to read inflation reports

Practical guidance for reading CPI/PCE releases:

1. Note month-over-month vs year-over-year. Year-over-year is the headline number. Month-over-month is what the Fed actually focuses on for forward direction.

2. Look at core vs headline. If headline is much higher than core, an energy/food shock is responsible. Probably temporary. If core is much higher than headline, the inflation is broad-based. Harder to fix.

3. Pay attention to shelter specifically. "Inflation excluding shelter" sometimes called "supercore" — strips out the sticky shelter component. If supercore is at 2% but headline is at 3%, you know exactly where the persistence is.

4. Watch services vs goods. Goods inflation (electronics, cars, furniture) often deflates during cycle slowdowns. Services inflation (restaurants, healthcare, recreation) tends to be sticky upward. If services inflation is high, that's a Fed concern.

5. Annualised month-over-month is useful. Multiplying a month-over-month rate by 12 gives you the implied annual rate at current pace. If month-over-month was 0.2%, annualised is 2.4%. Better forward indicator than year-over-year (which lags).

6. Don't react to one print. Inflation data is noisy month-to-month. Trends matter more than individual readings.

Inflation expectations — the secret weapon

Inflation expectations are tracked through both market-based measures (TIPS breakeven rates, 5y5y forwards) and survey-based measures (University of Michigan consumer survey, NY Fed consumer expectations survey). The Fed monitors both because anchored expectations are critical to inflation control.^[8] The Fed pays enormous attention to inflation expectations because they can become self-fulfilling.

Two main types tracked:

Market-based: TIPS breakeven rates (yield difference between regular Treasuries and inflation-protected ones). 5y5y forwards (the market's expected average inflation rate from 5 years out for 5 years). These reflect what bond traders are betting on.

Survey-based: University of Michigan consumer survey, NY Fed Survey of Consumer Expectations, Survey of Professional Forecasters. These reflect what people and economists expect.

When expectations are "anchored" — meaning people believe long-term inflation will stay near 2% — the Fed has more room to manoeuvre. Wage demands are moderate, business pricing decisions are restrained, expectations don't drift higher.

When expectations become "unanchored" — meaning people expect inflation to stay high — disinflation becomes much harder. Wage demands spiral. Businesses raise prices preemptively. Bond yields rise on inflation premium. The Fed has to hike harder and longer to break the cycle.

The 1970s inflation experience traumatized central bankers because expectations became unanchored. Volcker had to crush the economy with 20% interest rates to restore credibility. Modern Fed policy is largely about preventing that situation from ever recurring.

The Lighthouse takeaway

If you remember nothing else from this article, remember:

CPI is the most-cited inflation measure but the Fed targets Core PCE specifically. PCE typically runs 0.3-0.5 percentage points lower than CPI. Core inflation excludes volatile food and energy prices because they obscure the persistent inflation trend the Fed can actually influence. Shelter costs (33% of CPI) lag the cycle by 12-18 months, which is the main reason inflation refuses to fall quickly. The 2022 spike was caused by simultaneous fiscal stimulus, supply chain disruption, energy shock, labour market tightness, and slow Fed response. The "last mile" from 3% to 2% is the hardest part of disinflation. Anchored inflation expectations are what gives the Fed room to manoeuvre.

The Lighthouse dashboard tracks core inflation alongside the federal funds rate so you can see the relationship the Fed cares about. The next time you see "inflation hit X%" in headlines, you'll know which measure they mean, what core would say, and where shelter is in the picture.

The next article in the foundation series is Liquidity Plumbing — because to understand why some periods of high rates cause crises and others don't, you need to understand how money actually flows through the financial system.

Test your understanding