Nominal to real data: choose a defensible deflator
Converting nominal observations into real terms requires a deflator matched to the question. A price index is not a generic correction that makes every company series economically comparable. State whose prices, which geography, what period and which base year the transformed figures represent before asking AI to interpret growth.
The BLS CPI questions and answers describes the consumer-price measures and their scope. CPI concerns a defined consumer basket and population; it is not automatically a company’s output-price index or a measure of its exact cost inflation. Select the index for its relevance, not familiarity.
Define the real-value question
Distinguish purchasing-power analysis from a business volume estimate. Deflating company revenue with consumer prices can express revenue in consumer purchasing-power terms under a chosen convention. It does not isolate the quantity of products sold when company prices and product mix move differently from the index.
Record index series identity, coverage, frequency, seasonal-adjustment choice, vintage and reference period. Choose an annual-average index for an annual flow only when that averaging convention fits the task. A year-end index and an annual average answer different timing questions.
Keep currency conversion separate from inflation adjustment. Converting foreign currency amounts to one currency does not remove price-level changes. Conversely, deflating two local-currency series does not automatically put them into the same common-currency purchasing-power unit.
Hypothetical real-growth calculation
Assume fictional nominal revenue rises from 100 to 112. A deliberately selected price index rises from 100 to 108 over comparable periods. Expressing both in first-period prices gives real second-period revenue of 112 × 100 / 108, about 103.70, and real growth of about 3.70%.
Subtracting 8% inflation from 12% nominal growth gives four percentage points, which is an approximation rather than the exact multiplicative result. Write the formula as (1 + nominal growth) / (1 + index growth) − 1 when the rates and periods are compatible.
Suppose a more relevant output-price index rises only 4%. The inferred real change differs. That does not let the researcher choose the more attractive result. It requires explaining which question each deflator answers. All quantities and index movements in this example are hypothetical.
Deflator decision worksheet
| Field | Research decision |
|---|---|
| Nominal measure | Currency, metric and flow or balance date |
| Target interpretation | Consumer purchasing power or another explicit question |
| Index identity | Population, basket, geography and series variant |
| Period alignment | Annual average, endpoint or other justified convention |
| Base period | Unit in which real values are expressed |
| Vintage | Original or revised index history |
The basic level conversion is real value in base-period prices = nominal value × base index / current index. A rebased index with an equivalent relative path should not change computed real growth, but mismatched bases across rows can corrupt a table. Preserve index levels as well as percentages.
Limit the interpretation
Check whether quality adjustments, changing product mix or sector-specific prices weaken the selected index’s fit. State those limitations rather than calling the residual pure volume growth. A financial series can also contain acquisitions or perimeter changes that no deflator removes.
Give AI the deflator choice and its interpretation boundary with the transformed data. Ask it to distinguish observed nominal growth from the calculated real measure. Do not let the model label a CPI-deflated company series real unit sales. The final memo should be explicit about the economic meaning the conversion creates and the questions it leaves unanswered.
FAQ
Is nominal growth minus inflation exact real growth?
Not generally. Use the compatible multiplicative formula and label subtraction as an approximation if used.
Does CPI-deflated revenue measure physical sales volume?
No. It can express a selected purchasing-power convention, but company pricing, mix and reporting scope require separate analysis.