The GDP deflator vs the CPI; deflating aggregates

Detail

1. Nominal GDP, real GDP and the idea of "deflating"

• Nominal GDP: the value of all final goods and services made in the country in a year, counted at current prices (the prices of that same year).
• Real GDP: the same output counted at base-year (constant) prices. The base year is a fixed reference year.
• Why we need both: nominal GDP can rise for two reasons:
• more goods are made (a real rise), or

• the same goods cost more (only a price rise).

• Deflating means removing the price effect from a money total, so only the real change is left.

• MoSPI's bicycle example:
• Last year, 100 bicycles × ₹1,000 = ₹1,00,000.
• This year, the same 100 bicycles × ₹1,100 = ₹1,10,000.

• Nominal GDP looks 10% higher, but real growth is 0%, because no extra bicycle was made [2].

• MoSPI calls the price indices used for this job (CPI, WPI and others) deflators [2].

2. The GDP deflator: definition and formula

• GDP deflator = (Nominal GDP ÷ Real GDP) × 100.
• The quantities are the same on top and bottom. Only the prices differ.

• So the ratio shows only price change since the base year.

• The World Bank defines the GDP implicit deflator as "the ratio of GDP in current local currency to GDP in constant local currency" [3].

• Bread example (Class 12, National Income Accounting):

Year	Quantity	Price	Nominal GDP	Real GDP (at 2000 prices)
2000 (base)	100	₹10	₹1,000	₹1,000
2001	110	₹15	₹1,650	110 × ₹10 = ₹1,100

• Deflator (2001) = 1,650 ÷ 1,100 = 1.50, or 150.
• So bread prices rose 1.5 times (a 50% rise) since 2000.

• Real output rose 10% (100 → 110). Nominal output rose 65%. The deflator explains the gap.

• Useful growth link (an approximation):

• Nominal GDP growth ≈ Real GDP growth + inflation measured by the deflator.
• Example: nominal growth 10%, deflator inflation 4% → real growth ≈ 6%. The exact figure is 1.10 ÷ 1.04 − 1 = 5.77%.

3. Why the deflator is "implicit" and behaves like a Paasche index

• Implicit index: nobody builds a separate basket of goods for the deflator. It comes out of the national accounts by dividing nominal GDP by real GDP.
• It is economy-wide. It covers every good and service made in the country.
• The two index formulas (p = price, q = quantity, 0 = base year, 1 = current year):
• Laspeyres index (uses base-year quantities as weights) = Σp₁q₀ ÷ Σp₀q₀ × 100.

• Paasche index (uses current-year quantities as weights) = Σp₁q₁ ÷ Σp₀q₁ × 100.

• GDP deflator = Σp₁q₁ (nominal GDP) ÷ Σp₀q₁ (real GDP). This is exactly the Paasche form, because its weights are current production.

• The IMF's Quarterly National Accounts Manual (2017) describes national-accounts practice in the same way:
• Laspeyres volume indices measure real growth.

• The matching implicit Paasche price indices measure inflation [5].

• CPI, by contrast, is a Laspeyres-type index. Its basket and weights are fixed in the base year.

4. The GNP deflator

• GNP deflator = (Nominal GNP ÷ Real GNP) × 100.

• GNP (Gross National Product) = GDP + net factor income from abroad.

• NCERT exercise: nominal GNP ₹2,500 crore, real GNP ₹3,000 crore.

• Deflator = 2,500 ÷ 3,000 × 100 = 83.3.

• A value below 100 means prices have fallen since the base year.

• Trap: a deflator below 100 does not mean output fell. It means the price level is lower than in the base year.

5. NCERT's three differences between the GDP deflator and the CPI

Point	GDP deflator	CPI
Coverage	All goods and services made in the country: consumer goods, capital goods and government services	Only the consumer basket that households buy
Imports	Excluded, because imports are not domestic output	Included, because consumers buy imported goods
Weights	Change every year with that year's production (Paasche-like)	Fixed base-year basket (Laspeyres-like)

• Implication 1: crude oil price rise
• India imports most of its crude oil.
• Petrol and diesel are in the consumer basket, so the CPI rises directly.

• The deflator moves only through domestic output, such as refinery products made in India.

• Implication 2: machinery price rise

• Machinery is a capital good that India makes. So the deflator rises.

• Households do not buy machines, so the CPI hardly moves.

• Implication 3: fixed weights vs changing weights

• Fixed weights can make the CPI overstate living costs, because people switch away from goods that became costly.

• The deflator's changing weights take this switching into account.

• Frequency: the CPI comes out every month. The GDP deflator comes only with the quarterly or annual GDP estimates. So the RBI's inflation target uses the CPI, not the deflator.

6. The current CPI base (update to NCERT)

• MoSPI released the CPI with base 2024 = 100 on 12 February 2026. Its basket and weights come from the Household Consumption Expenditure Survey (HCES) 2023-24 [4][6]. (NCERT: base 2012.)
• The number of weighted items at all-India level rose from 299 to 358 [4]:
• goods: 259 → 308;

• services: 40 → 50.

• The 358 items are grouped under the COICOP 2018 classification into 12 Divisions, 43 Groups, 92 Classes and 162 Sub-classes [4].

• COICOP (Classification of Individual Consumption According to Purpose) is the UN's system for sorting household spending.

• The new IIP series (base 2022-23) was scheduled for May 2026 [6].

7. Deflating aggregates

• Deflating aggregates means using price indices to remove price change from money totals, such as national income and capital formation (money spent on new machines, buildings and other capital assets).
• Class 11 (Index Numbers) says the WPI is used for this. In practice, CPI and WPI components each deflate the sectors they match.
• Formula: Real value = Nominal value ÷ (Price index ÷ 100).

• Example: nominal output of a sector ₹550 crore, price index 110 → real output = 550 ÷ 1.10 = ₹500 crore at base-year prices.

• Single deflation: deflate a sector's output with one price index, and treat value added as moving the same way.

• Double deflation: deflate output and inputs separately, each with its own price index. Then:

• Real value added = real output − real inputs.

• Single extrapolation: move the base-year real value forward using a volume indicator, for example the quantity produced, instead of deflating money values. MoSPI uses it for sectors where double deflation is not applied [2].

8. Why single deflation can overstate growth: a worked example

• Base year: output ₹100, inputs ₹60 → value added (GVA) ₹40.
• Current year: quantities are the same. Output prices +5%, input prices −10%.

• Nominal output = ₹105. Nominal inputs = ₹54. Nominal GVA = ₹51.

• Single deflation (deflate GVA by the output price index 1.05):

• Real GVA = 51 ÷ 1.05 = ₹48.6. That shows 21.5% "growth", but nothing extra was made.

• Double deflation:

• Real output = 105 ÷ 1.05 = 100.
• Real inputs = 54 ÷ 0.90 = 60.

• Real GVA = 100 − 60 = ₹40, so 0% growth, which is correct.

• Lesson: when input prices fall faster than output prices (for example, when commodity prices crash), single deflation inflates real value added.

9. The measurement debate on the 2011-12 series

• No services in the WPI: the WPI covers only goods. Services are now more than half of GDP. So a goods index was deflating a large services economy.
• Low or negative WPI:
• In 2015–16, WPI inflation was very low or negative.
• Dividing nominal output by a very low deflator gives a high real number.

• So critics argued that real GDP growth was overstated.

• Single deflation: critics said the 2011-12 series deflated output with one index. In manufacturing, the same WPI was used for both output and inputs. MoSPI now accepts this was a weakness of the old series [2].

10. The fix: the 2022-23 GDP series (update to NCERT)

• MoSPI released the new GDP series with base year 2022-23 on 27 February 2026. It covers annual and quarterly estimates for 2022-23 to 2025-26 [2][6]. (NCERT/scaffold: 2011-12 base; revision "to be verified".)
• Why 2022-23:
• The Advisory Committee on National Accounts Statistics (ACNAS), set up in 2024, judged it a "normal" year with the survey data needed.

• The years 2017-18 to 2021-22 were rejected because of the GST rollout and COVID-19 [2].

• Deflation changes, recommended by an ACNAS sub-committee on deflators [2]:

• Single deflation has been completely eliminated.
• Double deflation is now used in manufacturing and agriculture.
• Single extrapolation is used in the other sectors.
• Deflators are used at a granular (item) level. Over 300 item-level indices are used.
• Manufacturing uses separate item-level WPIs for output and for intermediate consumption (inputs).

• Using item-level WPIs also avoids distortions from changing weights when items are added up.

• Quarterly GDP: CPI, WPI and the Unit Value Index (a price measure for exports and imports) have moved from aggregate level to item-group level [2].

• Still pending (as of 2026):
• The WPI base revision is still in progress, so the old 2011-12 WPI continues as a deflator for now [2].
• MoSPI plans to add the Producer Price Index (PPI) once DPIIT officially releases it [2].

• A PPI measures the prices producers receive, and it can cover services too. This is the long-term fix for the "no services in WPI" problem.

• Back series (old years recalculated with the new methods, back to 2011-12 and then spliced to 1950-51) is expected by December 2026 [2].

• Standards: India follows the SNA 2008 (System of National Accounts). It plans to move to SNA 2025 at the next base revision. India also subscribes to the IMF's Special Data Dissemination Standard (SDDS) [2].

Prelims Hooks

• GDP deflator = Nominal GDP ÷ Real GDP × 100. It is an implicit index with no fixed basket, and its weights make it Paasche-type (current-year quantities) [3][5].
• CPI is Laspeyres-type (fixed base-year basket). The GDP deflator is Paasche-type. This pairing is a common trap.
• Imports: in the CPI, yes; in the GDP deflator, no. A crude oil price shock shows up more strongly in the CPI.
• Capital goods (machinery): in the GDP deflator, yes; in the CPI, no.
• A deflator below 100 (e.g. 2,500/3,000 × 100 = 83.3) means prices fell since the base year. It says nothing about output falling.
• Double deflation deflates output and inputs separately. In the 2022-23 series it applies to manufacturing and agriculture, and single deflation has been completely eliminated [2].
• New GDP base 2022-23 was released on 27 Feb 2026. New CPI base 2024 was released on 12 Feb 2026, with 358 items (goods 308, services 50) [2][4][6].
• The WPI still has base 2011-12 and covers no services. DPIIT is preparing the PPI, which MoSPI will use once it is released [2].
• ACNAS, constituted in 2024, guided the GDP base revision [2].

Mains Points

• Which index to use for which job.
• The CPI measures cost of living and is the RBI's inflation target, because it is monthly and tracks household prices, including imported goods.
• The GDP deflator measures the price level of all domestic output, so it is the right tool for turning nominal growth into real growth.

• When the two diverge (for example, after oil shocks, when the CPI rises more than the deflator), policy readings differ. An answer should name the gap and its cause.

• Credibility of GDP data.

• The 2011-12 series drew criticism for single deflation and for using a goods-only WPI to deflate a services-heavy economy.
• The 2015–16 low WPI was said to have overstated real growth.
• The 2022-23 series answers this by removing single deflation, bringing in double deflation, using item-level deflators and using Supply and Use Tables to reduce the statistical discrepancy [2].

• This links to data transparency and India's IMF SDDS commitments.

• Unfinished reform.

• The WPI is still on base 2011-12, and the PPI is still awaited [2].
• Until then, services and non-manufacturing sectors depend on single extrapolation and older indices.

• Useful arguments: a services PPI, regular five-yearly base revisions [2], and a move to SNA 2025.

• Policy linkage.

• Real GDP drives the debt-to-GDP and deficit-to-GDP ratios, Finance Commission devolution, and RBI growth-inflation forecasts.
• So deflator errors spread into fiscal and monetary decisions. Getting the deflator right is a governance issue, not only a statistical one.