Methods of Construction of Index Numbers

Index numbers are statistical devices designed to measure the relative changes in the magnitude of a variable or a group of related variables over time, space, or between different categories. They are indispensable tools in economics and business, used primarily to measure changes in the price level (inflation), the volume of production, or the volume of trade. Common examples include the Consumer Price Index (CPI), the Wholesale Price Index (WPI), and the Index of Industrial Production (IIP).

Constructing an index number is a systematic process that involves several steps, from defining the purpose to applying a specific mathematical formula. The choice of formula determines the characteristics of the index and its suitability for specific analytical purposes. Broadly, the methods of construction of index numbers can be classified into two categories: Unweighted Index Numbers and Weighted Index Numbers.

General Steps in Construction

Before diving into the specific formulas, it is essential to understand the preliminary steps involved in the construction of any index number:

1. Purpose Definition: The specific objective of the index must be clearly stated. For instance, is the index intended to measure the cost of living for urban workers or the fluctuations in the prices of raw materials for manufacturers?
2. Selection of Commodities: It is impossible to include all commodities. A representative sample of items must be selected based on their importance and relevance to the target population.
3. Selection of Sources: Reliable sources of data must be chosen to collect price or quantity data.
4. Selection of Base Period: The base period is the period against which changes are measured. It should be a period of relative economic stability and normal conditions, not a year of depression, boom, or war.
5. Selection of Averages: A suitable method of averaging (such as Arithmetic Mean, Geometric Mean, or Median) must be chosen to aggregate the data.

Methods of Construction: Unweighted Index Numbers

Unweighted index numbers, also known as Simple Index Numbers, are constructed where all commodities included in the index are assigned equal importance. This implies that no specific weight is attached to any commodity based on its quantity consumed or produced. These are rarely used in practical scenarios because they often give misleading results due to the differing importance of various items in a typical budget. However, they form the theoretical basis for understanding weighted indices.

1. Simple Aggregative Method

This is the most elementary method of constructing an index number. In this method, the total of the current year prices for all commodities is divided by the total of the base year prices.

Where:
- $P_{01}$ = Price Index of the current year with respect to the base year.
- $\Sigma P_1$ = Sum of prices of the current year.
- $\Sigma P_0$ = Sum of prices of the base year.

Limitation: This method suffers from a major defect: it is influenced by the units of measurement. If the unit of measurement for a commodity is changed (e.g., from kilograms to grams), its price changes drastically, thereby altering the total sum and subsequently the index number, even if the actual price for the consumer remains unchanged.

2. Simple Average of Price Relatives Method

To overcome the unit-measurement issue of the aggregative method, the Price Relative approach is used. A price relative is the ratio of the current year price to the base year price for a commodity, expressed as a percentage.

In this method, price relatives are calculated for all commodities, and then an average (Arithmetic Mean, Geometric Mean, or Median) is taken. If Arithmetic Mean is used:

Where $N$ is the number of commodities. Since price relatives are pure numbers (percentages), they are not affected by the units of measurement. However, this method still fails to account for the relative importance (quantities) of different commodities.

Methods of Construction: Weighted Index Numbers

In reality, different commodities have different levels of importance in the consumption pattern or production process. For example, a rise in the price of salt has a negligible impact on the cost of living compared to a rise in the price of rice or fuel. Weighted index numbers assign specific "weights" to each commodity to reflect their relative importance. Weights can be quantities ($q_0$ or $q_1$) or values (expenditure).

Weighted index numbers are generally classified into two groups:

1. Weighted Aggregative Method

In this method, prices are multiplied by quantities (weights) to obtain values, and these values are then aggregated. There are several well-known formulas under this category:

Laspeyres Method

Proposed by the German economist Laspeyres, this method uses base year quantities ($q_0$) as weights. This is the most widely used method because base year quantities are fixed and easily available.

Advantage: It facilitates easy comparison between different time periods as the weights remain constant.
Disadvantage: It tends to overstate price increases (or understate decreases) because it does not account for consumers substituting away from goods that have become expensive in the current year.

Paasches Method

Proposed by the German statistician Paasche, this method uses current year quantities ($q_1$) as weights.

Advantage: It reflects current consumption habits.
Disadvantage: It is difficult to compare indices across different years because the weights change every year. Furthermore, current year data for quantities is often not readily available. It tends to understate price increases (or overstate decreases) because of the substitution effect.

Marshall-Edgeworth Method

This method attempts to strike a balance between Laspeyres and Paasche by using the average of base year and current year quantities as weights.

While this appears symmetric, it lacks a direct economic interpretation regarding which basket of goods is actually being represented.

Fishers Ideal Index

Irving Fisher proposed this index, calling it "ideal" because it satisfies two crucial statistical tests: the Time Reversal Test and the Factor Reversal Test. It is the geometric mean of the Laspeyres and Paasche indices.

Despite its statistical properties, it is computationally complex and requires data on both current and base year quantities, making it less popular for routine official indices like the CPI.

2. Weighted Average of Price Relatives

In this method, instead of aggregating values, we calculate the weighted average of the price relatives. The weights used are usually the values (expenditure) of the commodities.

Where $w$ represents the weight. Typically, base year values ($p_0 q_0$) are used as weights. When base year values are used, the weighted average of price relatives yields the same result as the Laspeyres index.

Tests of Adequacy of Index Numbers

To determine the accuracy and reliability of an index number formula, several tests are applied:

• Unit Test: This requires that the index number formula should be independent of the units of measurement. All formulas based on price relatives satisfy this test, while the simple aggregative method does not.
• Time Reversal Test: This test purports that if the time subscripts (0 and 1) are interchanged in the formula, the resulting index should be the reciprocal of the original index. In simpler terms, if the index for period 1 based on period 0 is 200, then the index for period 0 based on period 1 should be 50. Fishers Ideal Index satisfies this test.
• Factor Reversal Test: This test maintains that the product of a price index and a quantity index should equal the true value ratio (the ratio of total value in the current year to the total value in the base year). This ensures consistency between price and quantity indices. Fishers Ideal Index is one of the few that satisfies this.
• Circular Test: This test requires that if an index is calculated for periods 0, 1, and 2, then the index from 0 to 1 multiplied by the index from 1 to 2 should equal the index from 0 to 2. This is useful for chain base indices. The Geometric Mean of price relatives satisfies this test.

Conclusion

The construction of index numbers is not merely a mathematical exercise but a careful process of selection and weighting that reflects economic reality. While simple methods provide a basic understanding, weighted methods like Laspeyres and Paasche are essential for capturing the true impact of price changes on the economy. Fishers Ideal Index stands out for its statistical elegance, though practical constraints often favor the Laspeyres method for its feasibility and consistency. The choice of method ultimately depends on the purpose of the index, the availability of data, and the specific economic characteristics one wishes to measure.