Two inflations, and why they disagree

A packet of arhar dal costs about ₹180 a kilo in your neighbourhood kirana. The wholesale price that the dal mill charged the wholesaler may have been ₹140. That gap captures the difference between the Consumer Price Index and the Wholesale Price Index.

The CPI measures the retail prices households actually pay. It includes goods and also services: the rent for your flat, the school fees you pay, a doctor’s consultation. The WPI captures prices of goods at the wholesale level, before they reach you. No services, no rent, no school fees, no doctor visits. That makes the two indices fundamentally different animals.

CPI is built by MoSPI from a survey of what households spend, so its weights reflect actual consumption. WPI, compiled by the Office of the Economic Adviser in the Ministry of Commerce and Industry, weights goods by their importance in wholesale trade. Because WPI is goods-only and heavy on fuels and commodities, it swings with global prices. CPI, anchored by food and services, moves to a different beat.

So India doesn’t have one inflation number. It has at least two. And as we will see, they often shout in opposite directions.

In February 2026, India launched a new CPI series with base year 2024, built on the Household Consumption Expenditure Survey of 2023-24. The basket now tracks 358 items, up from 299 in the old 2012-base series. For the first time, it includes e-commerce prices, telecom data plans, and OTT streaming subscriptions, a recognition of where Indians actually spend their money today.

The weights reveal what modern Indian households prioritise. Food and beverages, at 36.75%, remains the single largest chunk but is nowhere near the 45.86% it commanded in the 2012 basket. This drop is a textbook case of Engel’s law: as people get richer, they spend a smaller share of their income on food and more on other things. Housing, water, electricity, and fuels now weigh 17.66%, transport and communication 12.41%, health 6.1%, and clothing and footwear 6.38%. Restaurants and accommodation get 3.35%, education 3.33%, and recreation 1.52%.

Here is how the CPI basket looks in the new 2024 series:

The basket is no longer frozen in 2012 spending patterns. It reflects what Indians buy now, not what they bought a decade ago. The new series recorded headline inflation of 2.75% in January 2026.

WPI measures price changes at the wholesale level: the mandi, the factory gate, the fuel terminal. It is split into three broad groups: primary articles (cereals, pulses, vegetables, minerals), fuel and power (coal, electricity, petrol, diesel), and manufactured products (textiles, chemicals, metals).

In April 2026, WPI inflation stood at 8.3%, driven largely by fuel and power, which was up 24.7% year-on-year. Primary articles rose 9.2%, and manufactured products a milder 4.6%. Notice what is absent: rent, healthcare, education, and personal services. A hike in school fees or a rise in doctor’s charges is invisible in WPI. So is the cost of an OTT subscription or a mobile data recharge.

WPI is thus a producer’s inflation, not a household’s. It sends early signals about pipeline costs: when global crude prices spike, WPI fuel and power leaps, often well before that pain shows up at the petrol pump or in higher transport fares embedded in CPI. But because WPI is so sensitive to global commodity swings, it can also give false alarms about domestic overheating. A surge in crude oil because of geopolitical tensions will light up WPI like a Diwali rocket, even if Indian demand is tepid.

The two indices diverge because their baskets and weights are worlds apart. CPI is dominated by food and services; WPI by globally traded goods like fuels and metals. Consider the 2012-base CPI (the most recent detailed series until the new 2024 base ramps up). In December 2025, overall CPI inflation was just 1.3%, dragged down by food and beverages. Here is how key food items performed:

Meanwhile, core inflation (excluding food and fuel) was running at 4.6%, and services like health (3.4%) and education (3.4%) ticked along.

Now look at WPI: all-commodities inflation was 8.3% in April 2026, with fuel and power soaring 24.7%. That spike had little to do with Indian vegetable prices. So while your grocery bill might have felt lighter, factories were paying much more for energy inputs. The two indices are measuring different economies: the CPI captures the consumption of households; the WPI captures the cost of production.

The divergence matters because if a policymaker glances only at WPI, they might conclude inflation is running hot and tighten policy, even as households experience relief. Conversely, a benign CPI could mask building cost pressures upstream that will eventually feed into retail prices.

In April 2026, WPI all-commodities was at 8.3%, while the most recent detailed CPI reading (December 2025, under the 2012 basket) came in at 1.3%. Even the newly launched 2024-base CPI for January 2026 showed only 2.75%. That is a gap of nearly 6 percentage points, one index suggesting serious inflation, the other near-zero.

The main driver is fuel. WPI fuel and power inflation hit 24.7% in April 2026, reflecting a surge in global energy costs. CPI, by design, gives fuel and light only a 6.8% weight (in the 2012 basket), and that weight is for direct household expenditure on LPG cylinders and electricity, not the indirect energy costs embedded in everything else. Your LPG cylinder may have become 5.5% more expensive, but the potato in your jhola is 35% cheaper. Meanwhile, food price deflation is pulling the CPI headline deep into calm territory. In the 2012 basket, food and beverages commanded 45.9%, so a sharp fall there dwarfs everything else.

What makes this twist particularly vivid is that businesses are squeezed between the two. A truck operator pays more for diesel (captured in WPI fuel) but the vegetables he transports are selling for less (captured in CPI food). The same economy, two completely different stories. And which story you hear depends entirely on which index you are looking at.

Until 2014, the Reserve Bank of India anchored its monetary policy to WPI. That meant when the central bank set interest rates, it looked primarily at wholesale price movements. But the Urjit Patel Committee, in its landmark report, argued this was fundamentally wrong. Monetary policy should target the inflation that households actually experience, not a producer price index that ignores services, which by then accounted for more than half of India’s GDP.

WPI excludes rent, school fees, healthcare, transport fares, and a host of personal services. A rise in the cost of living that comes from higher service prices would be invisible in WPI. WPI is also highly sensitive to volatile global commodity prices that may have little to do with domestic demand conditions.

The committee recommended switching the policy anchor to CPI (Combined), and the RBI adopted that in 2014. Now the RBI has a flexible inflation targeting framework: keep CPI inflation within a target band, with the repo rate as the main instrument. The repo rate stood at 5.25% in May 2026, with core CPI inflation at 4.6% in December 2025, suggesting the central bank was still watchful. Meanwhile, a bank fixed deposit offering 6.6% was delivering a real return of 5.35 percentage points after subtracting the headline CPI, a marked improvement over the negative real rates seen in the early 2010s.

Switching to CPI brought monetary policy closer to the ground. It is not perfect, no single number can capture every household’s experience, but it is a better measure of the erosion of purchasing power than WPI ever was.

On 12 February 2026, MoSPI released the first inflation figures from a completely rebuilt CPI series, with base year 2024 replacing 2012. The previous basket had become steadily less representative as spending patterns shifted. Engel’s law was at work: the share of household budgets going to food had fallen from 45.86% in 2011-12 to 36.75% in 2023-24, while shares of housing, transport, health, and communication rose. If the old basket had persisted, food price swings, like the ongoing deflation, would have received disproportionate weight, distorting policy signals.

The new basket grew from 299 to 358 items, adding e-commerce prices, telecom data plans, and OTT subscriptions. It adopted the international COICOP classification with 12 divisions, replacing the old six-group layout. This is not a cosmetic change: it allows more granular tracking of services and aligns India with global standards.

The first print under the new series: 2.75% for January 2026. WPI, however, remains anchored to a 2011-12 base, a full 13 years behind. An out-of-date basket misweights what the economy actually produces and consumes, making the index less reliable as a policy guide.

CPI is the inflation of your household budget. If you want to know how much the purchasing power of your salary has eroded, look at CPI. WPI is the inflation of the supply chain. If you are a manufacturer worried about input costs, or a policymaker trying to spot pipeline pressures, WPI gives you an early warning that CPI may not yet show.

But neither tells the whole truth. CPI is an average, and your personal inflation depends on what you buy. A daily-wage labourer in rural Bihar spends a far larger share on cereals than the national average; a software engineer in Bengaluru spends more on OTT and dining out. WPI misses services entirely, so an economy where services are 55% of output is half invisible to it.

Read both. Know what each leaves out. And remember that the basket matters as much as the number: an index built on ten-year-old spending patterns is a map of yesterday’s economy. The new CPI 2024 series brings the map closer to today, but even it will age. The next revision is already inevitable.