10 Time Series & Index Numbers · 10.1 Time series and index numbers

10.1.2 Index numbers: types and methods (Laspeyres, Paasche, Fisher, Marshall-Edgeworth), tests of adequacy

An index number measures the relative change in a group of related variables (prices, quantities, values) between a base period (= 100) and a current period; the weighted formulas of Laspeyres (base-year quantities), Paasche (current-year quantities) and Fisher (their geometric mean) are judged by the unit, time reversal, factor reversal and circular tests.

  • APPSC Paper I
  • Advanced level
  • 10 min read
  • Free

Free PDF

Download this PDF

Enter your email once and the PDF is yours. We send exam updates and free material; you can stop them any time.

In the print window, choose Save as PDF.

Core facts

  • Index numbers are called economic barometers. Base period index = 100.
  • Notation: p₀, q₀ = base-year price and quantity; p₁, q₁ = current-year price and quantity. P₀₁ = price index for year 1 on base 0.
  • Types: price index, quantity (volume) index, value index (Σp₁q₁ ÷ Σp₀q₀ × 100).
  • Simple aggregative: P₀₁ = Σp₁ ÷ Σp₀ × 100.
  • Price relative: R = p₁ ÷ p₀ × 100. Simple average of relatives: ΣR ÷ n (AM) or antilog (Σlog R ÷ n) (GM).
  • Laspeyres: P₀₁ = Σp₁q₀ ÷ Σp₀q₀ × 100 (weights = base-year quantities).
  • Paasche: P₀₁ = Σp₁q₁ ÷ Σp₀q₁ × 100 (weights = current-year quantities).
  • Fisher: P₀₁ = √(Laspeyres × Paasche) — the geometric mean of L and P; called the ideal index.
  • Marshall-Edgeworth: P₀₁ = Σp₁(q₀ + q₁) ÷ Σp₀(q₀ + q₁) × 100 (weights = sum, i.e. average, of both years' quantities).
  • Dorbish-Bowley: (L + P) ÷ 2 — the arithmetic mean of L and P.
  • Kelly (fixed weight): Σp₁q ÷ Σp₀q × 100, q = fixed quantities of some chosen period.
  • Quantity indices: swap p and q, e.g. Laspeyres Q₀₁ = Σq₁p₀ ÷ Σq₀p₀ × 100.
  • Tests of adequacy (Fisher):
    1. Unit test: index unaffected by units of measurement. Failed only by the simple aggregative index.
    2. Time reversal test: P₀₁ × P₁₀ = 1. Passed by Fisher, Marshall-Edgeworth, simple aggregative, simple GM of relatives, Kelly. Failed by Laspeyres, Paasche, simple AM of relatives.
    3. Factor reversal test: P₀₁ × Q₀₁ = Σp₁q₁ ÷ Σp₀q₀ (value ratio). Passed only by Fisher.
    4. Circular test: P₀₁ × P₁₂ × P₂₀ = 1. Passed by the simple GM of relatives, the simple aggregative and Kelly's fixed-weight index; not by Fisher.
  • Fisher is "ideal" because it passes the time reversal and factor reversal tests (not because it passes all tests).
  • Geometric mean is theoretically the best average for index numbers.

Example 1. Laspeyres index = 120 and Paasche index = 118.75. Find Fisher's index. Solution. √(120 × 118.75) = √14,250. Answer: 119.37 (about)

Example 2. Which index satisfies the factor reversal test? (a) Laspeyres (b) Paasche (c) Marshall-Edgeworth (d) Fisher Solution. Only Fisher's formula gives P₀₁ × Q₀₁ = value ratio. Answer: (d)

Explained

1. Problems in constructing an index number.

  • Purpose: decides items, prices and weights (a CPI for workers uses retail prices; WPI uses wholesale prices).
  • Base period: a normal year, free from wars, floods, booms; not too distant. Fixed base or chain base.
  • Selection of items: representative, standard quality, enough in number.
  • Price quotations: wholesale or retail, from representative markets.
  • Choice of average: AM is easy; GM gives equal weight to equal ratios and passes the time reversal test.
  • Weights: quantity weights (aggregative) or value weights (relatives); weighting is needed because items differ in importance.

2. Simple (unweighted) indices and the weighted formulas: one data set.

Itemp₀, q₀p₁, q₁Relative p₁/p₀
A10, 412, 5120
B8, 1010, 8125
C5, 66, 10120
D20, 222, 3110

Products:

Itemp₀q₀p₁q₀p₀q₁, p₁q₁
A404850, 60
B8010064, 80
C303650, 60
D404460, 66
Total190228224, 266

Example 3. Find the simple aggregative index and the simple AM of relatives. Solution. Σp₁ = 12 + 10 + 6 + 22 = 50; Σp₀ = 10 + 8 + 5 + 20 = 43. Simple aggregative = 50 ÷ 43 × 100 = 116.28. AM of relatives = (120 + 125 + 120 + 110) ÷ 4 = 475 ÷ 4 = 118.75. Answer: 116.28 and 118.75

Example 4. Find the Laspeyres, Paasche, Fisher, Marshall-Edgeworth and Dorbish-Bowley price indices. Solution.

  • Laspeyres = 228 ÷ 190 × 100 = 120.00
  • Paasche = 266 ÷ 224 × 100 = 118.75
  • Fisher = √(120 × 118.75) = 119.37
  • Marshall-Edgeworth = (228 + 266) ÷ (190 + 224) × 100 = 494 ÷ 414 × 100 = 119.32
  • Dorbish-Bowley = (120 + 118.75) ÷ 2 = 119.375

Answer: L 120, P 118.75, F 119.37, M-E 119.32, D-B 119.38

3. Weighted average of relatives. P₀₁ = Σ(R × W) ÷ ΣW. With value weights W = p₀q₀ it gives exactly the Laspeyres index: (120 × 40 + 125 × 80 + 120 × 30 + 110 × 40) ÷ 190 = 22,800 ÷ 190 = 120.

4. Fixed base and chain base.

  • Fixed base: every year compared with one base year.
  • Chain base: each year compared with the previous year. Link relative = current value ÷ previous year's value × 100; chain index = link relative × previous chain index ÷ 100.

Example 5. Link relatives of three years are 110, 120 and 90 (first year's base = 100). Find the chain indices. Solution. Year 1: 110. Year 2: 120 × 110 ÷ 100 = 132. Year 3: 90 × 132 ÷ 100 = 118.8. Answer: 110, 132, 118.8

PointFixed baseChain base
Compares withone base yearprevious year
New items, weightshard to addeasy to add
Long-run comparisondirectvia chaining
Errordoes not build upcan accumulate

5. Uses of index numbers. Measure changes in price level and the value of money; guide policy (wages, dearness allowance, monetary policy); deflate money values into real values; show trends in production, trade and business; compare places and periods.

Deeper

1. Checking the tests with the data of Example 4.

  • Time reversal for Laspeyres: P₀₁ = 228 ÷ 190 = 1.2; P₁₀ (year 1 as base, its quantities q₁ as weights) = Σp₀q₁ ÷ Σp₁q₁ = 224 ÷ 266. Product = 1.2 × 224 ÷ 266 = 1.0105 ≠ 1. Fails.
  • Time reversal for Fisher: P₀₁ × P₁₀ = √(228/190 × 266/224) × √(224/266 × 190/228) = 1. Passes.
  • Factor reversal for Fisher: quantity indices are Laspeyres Q = 224 ÷ 190 = 117.89, Paasche Q = 266 ÷ 228 = 116.67, Fisher Q = √(117.89 × 116.67) = 117.28. P₀₁ × Q₀₁ = 1.1937 × 1.1728 = 1.40 = 266 ÷ 190. Passes.

2. Proof sketch of factor reversal for Fisher. P₀₁ × Q₀₁ = √(Σp₁q₀/Σp₀q₀ × Σp₁q₁/Σp₀q₁) × √(Σq₁p₀/Σq₀p₀ × Σq₁p₁/Σq₀p₁) = √((Σp₁q₁)² ÷ (Σp₀q₀)²) = Σp₁q₁ ÷ Σp₀q₀.

3. Bias. Laspeyres uses old quantities and ignores consumers shifting away from goods whose prices rise, so it tends to overstate price rises (upward bias); Paasche tends to understate them (downward bias). Fisher lies between them. This holds when prices and quantities move in opposite directions; it is not a mathematical law, and the order can reverse.

4. Fisher's limitations. Needs current-year quantities every year (costly; that is why official CPIs and WPIs use the Laspeyres type), is harder to compute and to explain, and fails the circular test.

IndexUnitTime reversalFactor reversal
Simple aggregativefailspassesfails
Laspeyres / Paaschepassesfailsfails
Marshall-Edgeworthpassespassesfails
Fisherpassespassespasses

Circular test: passed by the simple aggregative, simple GM of relatives and Kelly's fixed-weight index; failed by all four weighted indices above.

Example 6. If Σp₀q₀ = 200, Σp₁q₀ = 250, Σp₀q₁ = 240 and Σp₁q₁ = 300, find Laspeyres, Paasche and Fisher. Solution. L = 250 ÷ 200 × 100 = 125; P = 300 ÷ 240 × 100 = 125; F = √(125 × 125) = 125. Answer: all 125 (when L = P, Fisher equals both)

Common traps

  • ✗ "Laspeyres uses current-year quantities." ✓ Laspeyres = base-year (q₀); Paasche = current-year (q₁).
  • ✗ "Fisher's index is the arithmetic mean of L and P." ✓ It is the geometric mean; the AM is Dorbish-Bowley.
  • ✗ "Fisher's index passes every test." ✓ It fails the circular test; it is ideal for passing time reversal and factor reversal.
  • ✗ "Marshall-Edgeworth passes the factor reversal test." ✓ It passes time reversal only; factor reversal is Fisher alone.
  • ✗ "Time reversal test: P₀₁ + P₁₀ = 1." ✓ The product P₀₁ × P₁₀ = 1 (without the factor 100).
  • ✗ "Simple aggregative index passes the unit test." ✓ It fails it: changing kg to quintal changes the result.
  • ✗ "Base-year index = 1 or 0." ✓ Base-year index = 100.

Quick revision

  • Laspeyres = Σp₁q₀/Σp₀q₀ × 100 (base quantities).
  • Paasche = Σp₁q₁/Σp₀q₁ × 100 (current quantities).
  • Fisher = √(L × P), ideal index.
  • Marshall-Edgeworth = Σp₁(q₀ + q₁)/Σp₀(q₀ + q₁) × 100.
  • Dorbish-Bowley = (L + P)/2.
  • Unit test: fails only simple aggregative.
  • Time reversal: P₀₁ × P₁₀ = 1 (Fisher, M-E pass; L, P fail).
  • Factor reversal: P₀₁ × Q₀₁ = Σp₁q₁/Σp₀q₀ (Fisher only).
  • Circular test: simple aggregative, simple GM and Kelly pass; Fisher fails.
  • Chain index = link relative × previous chain index ÷ 100.

Sources: Telugu Akademi, Statistics (B.A. / B.Com / B.Sc.): index numbers; S.P. Gupta, Statistical Methods (Sultan Chand), index numbers; S.C. Gupta and V.K. Kapoor, Fundamentals of Applied Statistics (Sultan Chand), index numbers; NCERT Statistics for Economics, Class 11, index numbers

Last checked 2026-10-21.

Practise this topic. Questions on Index numbers: construction, tests & cost of living, with answers explained.

Medha AI writes its own notes and checks them against the sources listed. Medha AI is an independent preparation service. It is not affiliated with, endorsed by or connected to APPSC, TSPSC, MPSC, UPSC or any other commission or government body.