Reading Economic Data: Levels, Growth and Base Effects

The level and the change tell different stories

An economic release often reports both a level and a rate of change. A price index summarises movements in a specified basket of prices relative to a reference period. A growth rate compares two observations of that index. Confusing these quantities can turn “inflation has fallen” into the very different claim that “prices have fallen”.

Inflation is a rise in the general price level measured by the chosen index. Disinflation means that the rate of inflation slows. Deflation means that the measured price level falls. These definitions concern a price index: slowing growth in an employment or production index should not automatically be called disinflation. The object being measured matters as much as the arithmetic.

A year-on-year rate compares an observation with the corresponding period a year earlier. A monthly rate compares it with the preceding month. Annualising a monthly rate asks what repeating that pace would imply, rather than describing what actually happened over a year. Always identify the period and any seasonal adjustment before interpreting a headline percentage.

The comparison base influences the rate

The example changes two things: the absolute increase falls from ten points to five, and the starting level rises from 100 to 110. It therefore does not isolate a pure base effect. To see denominator dependence separately, a five-point increase from 100 is 5%, whereas a five-point increase from 110 is about 4.55%. The same absolute movement can produce different percentage changes.

In actual year-on-year data, an unusually large move in the comparison period can affect the current annual rate as that earlier observation enters or leaves the calculation. The ECB discussion below illustrates this mechanism with historical energy prices. Those historical contributions are not forecasts for current inflation, and a base-effect explanation does not establish that present price pressures are absent.

Comparability comes before interpretation

A headline index summarises a basket; it need not match any individual household’s spending. Components can move in opposite directions, and expenditure weights affect the aggregate. A change in the overall rate alone cannot identify which prices moved or why. Supply conditions, demand, taxes and measurement choices require additional evidence.

Economic series may be revised as information improves. Seasonal adjustment can also change earlier observations. Comparing two reports without recording their release dates risks treating a revision as a new economic event. The example assumes fixed annual observations with no revisions. Real analysis benefits from noting both the period described and when the number became available.