CPI Measurement Problems The Case of Malawi By Charles Machinjili National Statistical Office Malawi Introduction • • • • • • • Malawi Small landlocked sub-Saharan Africa Area around 118,000 sq kms Population 12 million mid yr 2005 Economy predominantly agricultural Accounting for 40 % of GDP Household Expenditure Surveys have been done every 10 yrs since 1964 • The data has been used for constructing CPI Methodology • Two Strata: Urban and Rural • Fresh produce actually bought since sold in heaps and bundles • Other commodities use price quotations • Prices collected through out country using Regional Statistical Offices • Index calculated using Excel spreadsheets • Use base-weighted Laspeyres formula Measurement Problems • Derivation of weights: Over 85 % population is in rural areas without monetized expenditure • Engage in own-consumption: should this be included or left out • Current CPI includes own-consumption of foodstuffs leading high weight for Food Group in index • Collection of Prices: Timing is important since it determines the level of prices • In Malawi certain days are designated as market days and it is necessary to collect prices on such days • However it may be necessary to get an average price for the month • Purchasing and Weighing of Items: Items from produce markets, usually sold in heaps and bundles, are bought and later weighed to get unit prices • Problems of cost, getting accurate readings from weighing equipment, disposal of commodities • Missing Items: Where no prices are available what should one do: assume last month’s price which is the case in my country----leading to an underestimate of the price • Quality Changes: When observed differences are negligible price collector substitutes new for the old product • Non-comparable substitution maybe used • Non-comparable substitution uses overlap pricing. This requires that the prices of both the old and new product are available at least one time period • Choice of formulae: Use of Laspeyres formula with its fixed base takes no account of people’s ability to substitute to achieve the same utility when commodities become more expensive over time Improvements to Measurement of CPI • Commodity Weights: Right methodology for collecting expenditure has to be determined from outset: Is it diary and for what duration or is it a one time question with short recall period? • Monetary versus non-monetary expenditure: inclusion of one or both requires consideration • Timing of Price Collection: Special “market days”, common in developing countries especially in rural areas should be taken into account • Weighing of Commodities: As far as possible electronic/digital scales should be used • Missing items: Explicit imputation is recommended • Quality Change: Estimate price trend of similar products to represent price trend of discontinued product without direct comparison between price of old and new product. Difference in price of imputed price (old product) and new product is the implicit quality adjustment • Substitution bias: Use formulae that allow for substitution effects e. g the Fisher Ideal Price Index • THANK YOU