Core Inflation

Core inflation is not a separate survey; it is the same consumption basket with two component groups removed and the remaining weights rescaled so they again sum to one. If food and energy together account for roughly a fifth of a consumer price basket, the residual components are each divided by about 0.8, which mechanically raises the influence of shelter, services and core goods on the reported figure. The inflation rate itself is then computed the usual way: the core index level in the current period divided by its level twelve months earlier, minus one, expressed as a percentage.

Because the reweighting amplifies whatever remains, core is not a neutral filter. A basket dominated by rents and medical services will produce a core reading that moves slowly and with long lags, while a basket heavy in tradable core goods will respond faster to exchange rates and global supply conditions.

Food and energy prices are set largely in global commodity markets and respond to weather, harvests, OPEC decisions, pipeline outages and geopolitical events. These shocks are frequently mean-reverting: a spike raises the year-over-year rate for roughly four quarters, then drops out of the base and mechanically pulls the rate down again. Filtering them removes noise that a policy rate, which acts with a lag of several quarters, cannot influence anyway.

The filter breaks down when energy shocks propagate. Fuel is an input to freight, airfares, plastics and food processing, so a sustained oil move eventually shows up inside core. If workers bargain for compensation after a grocery-price surge, or if firms reset prices on the expectation that inflation will stay high, the shock becomes embedded in wages and services. At that point core rises after headline rather than instead of it, and can remain elevated once headline has already fallen back.

The Federal Reserve's formal objective is defined on headline personal consumption expenditures inflation, yet core PCE is the series most closely watched in communications and projections. Core PCE and core CPI differ in scope and weighting, notably in how shelter and medical costs enter, so the two can diverge by several tenths of a percentage point for extended periods. The European Central Bank publishes HICP excluding energy and food alongside its headline target measure, and most inflation-targeting central banks in emerging markets do the same.

Statistical agencies also produce alternatives to simple exclusion, such as trimmed-mean indices that discard the largest price increases and decreases in each period, and weighted median inflation. These are designed to strip volatility wherever it appears rather than assuming it always resides in the same two categories. Markets track all of them: breakeven inflation rates and swap curves are quoted on headline indices, while the persistence signal that drives rate expectations comes largely from core and trimmed measures.

Expect to be handed a table of headline and core rates over several periods and asked what it implies about the nature and durability of the inflation episode. The reasoning to demonstrate is directional: headline above core with core stable suggests a commodity or supply shock that will likely unwind; core rising toward or past headline suggests broadening pressure and a stronger case for tighter policy; both falling with core lagging is consistent with disinflation still working through services.

You should also be able to link core inflation to the demand-pull versus cost-push distinction, to the output gap and short-run aggregate supply framework, and to expectations formation. Related mechanics worth being fluent in include base effects, the difference between a falling inflation rate and outright deflation, and how index construction choices, such as fixed-weight Laspeyres formulas and substitution bias, affect any measured rate, core included.