Price and volume indices: Laspeyres, Paasche, Fisher, chain indices, deflator and the 4-5-4 calendar
How to separate revenue growth “in roubles” from growth in units: compute a price index for a basket (Laspeyres, Paasche, Fisher), link periods by a chain, restate revenue in comparable prices with a deflator, isolate the assortment mix effect and not be fooled by the calendar (different weekdays in months, the 53rd week). The decision: whether the business really grows or just gets more expensive.
Why this document
The finance and commercial directors read this when revenue has grown and you need to know why: price, volume or a shift in the sales mix. In inflation, revenue growth in roubles proves nothing by itself, so it is restated into comparable prices with a deflator. The Laspeyres, Paasche and Fisher indices give different estimates because they use different baskets, and the gap between them is itself informative. The 4-5-4 calendar lets you compare weeks and months without distortion from holidays and the number of weekends.
At a glance
- What is this document and why is it needed?
- How to separate revenue growth “in roubles” from growth in units: compute a price index for a basket (Laspeyres, Paasche, Fisher), link periods by a chain, restate revenue in comparable prices with a deflator, isolate the assortment mix effect and not be fooled by the calendar (different weekdays in months, the 53rd week). The decision: whether the business really grows or just gets more expensive.
- Who owns the document and how often is it rebuilt?
- A financial or sales analyst; indices are computed monthly and quarterly against plan and the previous year; read by the owner, the commercial and financial directors.
- What mistakes are the most common?
- Revenue divided by units is called price → the mix effect is taken for price policy → compute a price index on a fixed basket. Revenue growth is compared with national average inflation → real growth is distorted → a deflator from the own basket. Calendar months are compared year on year → weekdays and holidays create ±1-2% false growth → the 4-5-4 calendar or a correction for the trading mass of days.
For whom
- Brand
- A brand splits revenue growth into price, volume and mix by channel and collection and compares price growth with growth in cost and the exchange rate.
- Retail
- A retailer splits growth into traffic, average receipt, price and category mix and compares comparable periods allowing for the 4-5-4 calendar.
What it consists of
Why indices
- Revenue rose 3.2% over the year, but this is not business growth: prices in a comparable basket rose 7.6% (the Fisher index) and the number of units sold fell 3.4%. Without indices the two movements cannot be told apart, yet purchasing, staffing and space decisions are taken on nominal revenue.
- An index is the ratio of the cost of a fixed basket at current prices to its cost at base prices. The basket (a set of categories or SKUs with quantities) must be the same in numerator and denominator; otherwise what changes is composition, not price.
- The price index times the volume index gives the revenue index: I_revenue = I_price × I_volume. The identity is handy for checking: the product must equal revenue growth.
Laspeyres, Paasche, Fisher
- Laspeyres weights prices by base-period quantities, Paasche by current ones. Laspeyres is usually higher: buyers move from what rose in price to what rose less, and base weights do not see it (the substitution effect). Paasche is usually lower; the Fisher index, the geometric mean of the two, lies between them and is time-reversible.
- In the example over four quarters year on year: Laspeyres 7.71, 7.19, 7.65, 7.8%, Paasche 7.59, 7.12, 7.52, 7.71%, Fisher 7.65, 7.16, 7.58, 7.76%. The gap is small (up to 0.13 pp) because the mix changes moderately within a year; over a long interval and with a strong mix change it grows.
- Choice: for public reporting and comparison with official price statistics use Laspeyres; for internal analysis and growth decomposition Fisher is better because it does not depend on which period is called the base.
Chain indices and the deflator
- A chain index multiplies indices of adjacent periods: the basket is updated every period, so it copes with fast collection changes. In the example the chain Fisher price index rose to 113.7 over 7 quarters from a base of 100. The chain drawback: rounding error accumulates and for seasonal goods a quarter-on-quarter chain jumps.
- A deflator is the price index by which a nominal quantity is divided to get a real one: real revenue = nominal ÷ deflator. In the example by the last quarter nominal revenue is 120 on a base of 100 and real revenue 106: growth in roubles is not growth in comparable prices.
- For comparing quarters within a year seasonality is stronger than prices: compare a quarter with the same quarter of the previous year and isolate seasonality by a separate decomposition (the Holt-Winters and seasonal index document).
- Brand: the chain price index by collection is built on models present in both seasons; new models enter the basket from the second period so that price and composition are not mixed.
- Retailer: a store chain index is computed by week with a holiday correction, and real sales per square meter and per employee use the store basket deflator rather than the general consumer price index.
Mix, price and volume: the revenue bridge
- Revenue growth decomposes into unit volume, assortment mix and price. In the example: revenue 2,337 → RUB 2,412 million; volume −RUB 80 million, mix -15, price 169. The mix is negative: sales shifted towards cheaper categories.
- The average receipt and price are not the same: the average unit price rose 6.8% while basket prices rose 7.6%; the difference is the mix effect. If “price” in a report is simply revenue ÷ units, the conclusion about price policy is wrong.
- The mix effect is computed at one category level: groups that are too large hide the shift inside, those that are too small (SKU) give noise and new items without a base period; for new items the base-period price is taken from an analog.
- Brand: the mix is shown by channel: a shift of sales to marketplaces with a lower price for the brand lowers the average price even as the price list rises.
- Retailer: the mix is shown by category and by the share of sale items: a growing share of discounted goods explains part of the fall in the average receipt.
Calendar: month against 4-5-4
- Calendar months have different weekday compositions: with weekday weights (Saturday and Sunday heavier than weekdays) the trading mass of a 2026 month differs from the same month of 2025 by -1.7...+1.5%. This is a pure distortion of the year-on-year comparison unrelated to demand.
- The 4-5-4 calendar divides the year into 52 weeks (quarters of 13 weeks: months of 4, 5 and 4 weeks) and every month holds a whole number of identical weeks: the trading mass of a month is the same year on year. The price: 52 weeks is 364 days and the year drifts, so a 53rd week is added periodically (under the “last Sunday of January” rule in 2021 and 2027).
- In a year with a 53rd week comparable 52 weeks are compared (without the extra one) and the effect of the extra week is shown separately; plan targets and bonuses are tied to the calendar, not to the “month”.
Brand and retailer: what is computed in practice
- Brand: a price index by channel (wholesale, D2C, marketplaces) and by collection; price growth is compared with growth in cost and the exchange rate to see whether the cost rise passes into price.
- Retailer: a price index by store basket category; LFL revenue growth is divided into traffic, receipt (price and mix) and conversion; comparable stores are taken by the 4-5-4 calendar.
- Brand: wholesale and retail prices are compared in different indices: the brand wholesale price and the partner retail price move with a lag and different discount depth.
- Retailer: the index of price after discounts (the actual selling price) separately from the list price index; their gap is promo depth and frequency.
- Brand: real revenue (in comparable prices) goes into brand growth assessment and bonuses; nominal revenue without a deflator overstates growth in high inflation.
- Retailer: real revenue per square meter and per employee allows comparing stores and years; nominal measures rise with inflation even when traffic falls.
Where the data comes from
| Data source | Fields |
|---|---|
| Sales by category or SKU | Revenue and quantity by category (better by SKU) for base and current periods; actual selling prices after discounts. |
| Price list | List prices by date, effective dates and price reductions; for new items the base-period price of an analog. |
| Calendar | The 4-5-4 calendar or calendar days by month, weekday weights from sales, holidays. |
How to build it step by step
- Step 1. Define the comparison base (the same period last year, the previous period, plan) and the basket level (categories or SKUs).
- Brand: Step 2. Build the basket by channel and collection; for new collections take the base-period price from an analog of the previous season.
- Retailer: Step 3. Build the basket by store category; exclude closed and opened stores (a comparable base) and drop weeks of refit.
- Step 4. Compute Laspeyres and Paasche, then Fisher; check that Fisher lies between them.
- Step 5. Multiply the price index by the volume index and reconcile with revenue growth: a gap above rounding means a basket or data error.
- Step 6. Build the revenue bridge: volume, mix, price; write out each category contribution.
- Brand: Step 7. Compare the price index with growth in cost in roubles and the exchange rate; prepare a conclusion on how much price compensated the cost rise.
- Retailer: Step 8. Split LFL growth into traffic, conversion and average receipt and the receipt into price and mix; take comparable weeks by the 4-5-4 calendar.
- Step 9. Obtain real revenue by dividing by the chain deflator and compare nominal and real growth rates; for seasonal data compare year on year.
- Step 10. Record the method (basket, level, weights, calendar) in a document and do not change it without recomputing history, otherwise the series are not comparable.
Formulas
| Measure | Formulas | Note |
|---|---|---|
| Laspeyres index | L = Σ p_t·q_0 / Σ p_0·q_0 | Base-period weights; sums over basket categories. |
| Paasche index | P = Σ p_t·q_t / Σ p_0·q_t | Current-period weights. |
| Fisher index | F = √(L·P); F_price × F_volume = V_t / V_0 | V is revenue; the identity holds exactly if the volume index is also Fisher. |
| Chain index | I_(0,T) = Π_(t=1..T) F_(t−1,t) | The basket is updated every period; for seasonal data use year on year. |
| Deflator and real quantity | Real_t = Nominal_t / (I_(0,t) / 100) | A base of 100 in the base period. |
| Revenue bridge | V_1 − V_0 = V_0·(Q−1) + [Σp_0(q_1−q_0) − V_0·(Q−1)] + Σ(p_1−p_0)·q_1, Q = Σq_1/Σq_0 | Three terms: unit volume, mix, price; the order of decomposition affects the split. |
| Trading mass of a month | W_m = Σ_(d ∈ m) w_(weekday of d); distortion = W_m(y) / W_m(y−1) − 1 | Weights w_d are taken from sales by weekday; in the 4-5-4 calendar the distortion is zero. |
Formulas in notation
- Fisher
F = √(L · P)
The geometric mean of the two indices.- Identity
V_t / V_0 = F_price · F_volume
Revenue growth equals the product of the price and volume indices.- Real quantity
Real_t = Nominal_t / (I_{0,t} / 100)
Division by the deflator.- Calendar distortion
δ_m = W_m(y) / W_m(y − 1) − 1, W_m = Σ_{d∈m} w_{weekday(d)}
Trading mass of the days of a month; zero in the 4-5-4 calendar.
Worked example
Quarterly year-on-year indices: price by Laspeyres, Paasche, Fisher and volume
| Quarter 2026 | Revenue, % | Laspeyres, % | Paasche, % | Fisher, % | Volume, % |
|---|---|---|---|---|---|
| Q1 | 0.78 | 7.71 | 7.59 | 7.65 | −6.38 |
| Q2 | 4.6 | 7.19 | 7.12 | 7.16 | −2.39 |
| Q3 | 4.19 | 7.65 | 7.52 | 7.58 | −3.15 |
| Q4 | 3.48 | 7.8 | 7.71 | 7.76 | −3.97 |
Six categories (clothing, footwear, accessories) over 8 quarters are generated in code (fixed seed): prices rise 4-11.5% a year, volumes react to price with an elasticity of −1.3, seasonality is set. All values are changes against the same quarter of 2025. Check: (1 + revenue) = (1 + Fisher) × (1 + volume). The data is illustrative.
Figures
- Nominal revenue
- Real revenue
- Chain price index
Data as a table
| Measure | Nominal revenue | Real revenue | Chain price index |
|---|---|---|---|
| 25 Q1 | 100 | 100 | 100 |
| 25 Q2 | 85.9 | 83.9 | 102 |
| 25 Q3 | 90.7 | 87.3 | 104 |
| 25 Q4 | 116 | 110 | 106 |
| 26 Q1 | 101 | 93.8 | 108 |
| 26 Q2 | 89.9 | 81.9 | 110 |
| 26 Q3 | 94.5 | 84.5 | 112 |
| 26 Q4 | 120 | 106 | 114 |
Nominal and real revenue and the chain price index. By the last quarter prices on the chain index rose to 113.7, nominal revenue 120, real 106; seasonal swings are stronger than price ones, so quarters are compared year on year. Axis: quarter.
- Laspeyres
- Paasche
- Fisher
Data as a table
| Measure | Laspeyres | Paasche | Fisher |
|---|---|---|---|
| Q1 | 7.71 | 7.59 | 7.65 |
| Q2 | 7.19 | 7.12 | 7.16 |
| Q3 | 7.65 | 7.52 | 7.58 |
| Q4 | 7.8 | 7.71 | 7.76 |
Laspeyres, Paasche and Fisher: price growth year on year. Laspeyres is not below Paasche in any quarter; Fisher lies between them. The gap is up to 0.13 pp: within a year the mix changes moderately. Change against the same quarter of 2025.
- positive
- negative
- total
Data as a table
| Step | Change | Level |
|---|---|---|
| Revenue 2025 | — | 2,337 |
| Unit volume | −80 | 2,257 |
| Assortment mix | −15 | 2,242 |
| Price | +169 | 2,411 |
| Revenue 2026 | — | 2,412 |
Revenue bridge 2025 → 2026: volume, mix, price. Price added RUB 169 million, unit volume and mix took away RUB 94 million: revenue rose by RUB 75 million but the business shrank in units.
- Price
- Unit volume
Data as a table
| Measure | Price | Unit volume |
|---|---|---|
| Outerwear | 8.9 | −7.8 |
| Knitwear | 3.5 | 3 |
| Trousers and jeans | 8 | −7.4 |
| Dresses | 6 | 0.2 |
| Footwear | 10.8 | −7.8 |
| Accessories | 8.6 | −4.3 |
Price and volume growth by category, 2026 against 2025. Where prices rose the most (footwear, outerwear) volume fell the most: elastic demand moves to cheaper categories, hence the negative mix.
- Calendar month
- 4-5-4 calendar
Data as a table
| Measure | Calendar month | 4-5-4 calendar |
|---|---|---|
| 1 | 1.49 | 0 |
| 2 | 0 | 0 |
| 3 | −1.62 | 0 |
| 4 | 0.31 | 0 |
| 5 | 0.74 | 0 |
| 6 | −1.08 | 0 |
| 7 | 0.75 | 0 |
| 8 | −0.73 | 0 |
| 9 | 0.16 | 0 |
| 10 | 1.49 | 0 |
| 11 | −1.66 | 0 |
| 12 | 0.3 | 0 |
Year-on-year distortion: calendar months and 4-5-4. Calendar months of 2026 against 2025 are distorted by -1.7...+1.5% only because of the weekday composition; the 4-5-4 calendar gives zero. A 53rd week in 2026-2027: 2027. Axis: month of the year; weekday weights are illustrative.
Specifics by category
- Apparel
- Fast collection changes: the basket is built by analog groups, not by SKU; a chain index is preferable to a fixed base, the season structure (outerwear in winter) is stronger than price.
- Footwear
- Prices rise faster (imported components, the exchange rate), volume falls harder: in the revenue bridge price gives a larger plus and volume a larger minus; an index by models with separate sizes is not needed.
- Accessories
- Many cheap items with fast turnover: an index by groups, the mix is set by gift peaks; the 4-5-4 calendar matters because of pre-holiday weeks.
- Kids
- Size groups and the school calendar: a basket by age; price rises are not always followed by a volume fall because the purchase is forced; indices are computed separately for school items.
- Men
- A stable mix: Laspeyres, Paasche and Fisher almost coincide, Fisher by categories is enough; the mix effect is small, price makes the main contribution to growth.
- Women
- Fast range turnover and a high share of new items: the base-period problem for new items; the base price is taken from an analog and the mix effect in the revenue bridge is visible and often negative because of markdowns.
Common mistakes
- Revenue divided by units is called price → the mix effect is taken for price policy → compute a price index on a fixed basket.
- Revenue growth is compared with national average inflation → real growth is distorted → a deflator from the own basket.
- Calendar months are compared year on year → weekdays and holidays create ±1-2% false growth → the 4-5-4 calendar or a correction for the trading mass of days.
- The basket level changes from period to period → indices are not comparable → one level and recomputed history when the method changes.
- LFL growth is computed ignoring the 53rd week → a year with an extra week looks like growth → compare 52 weeks and show the effect of the week separately.
Russian specifics
- Retailer: tax and accounting reporting uses calendar periods; the 4-5-4 calendar is a management axis and reconciliation with reporting requires a bridge table between calendar months and trading weeks.
- Brand: price indices for wholesale partners must not be used to coordinate retail prices between competitors; price information exchange is limited by competition protection requirements.