Skip to content

The South African CPI Sources and Methods Manual, Release v.1

Literature › Works

Statistics South Africa (2009). The South African CPI Sources and Methods Manual, Release v.1. Statistics South Africa (Stats SA).
Type report
Scope Comparative and methods
Method policy analysis
Period studied 1917–2009
Regions South Africa, urban areas by province
Themes methods data sources, macro fiscal finance debt
Citation key africa2009

Argument

This is an official technical manual documenting how Stats SA compiles the post-2008 consumer price index. It sets out the CPI's two objectives (inflation measurement and cost-of-living indexation), the COICOP classification, the 2008 weights drawn from the 2005/6 Income and Expenditure Survey (IES), basket selection, geography, price collection, the treatment of housing, and the Jevons/Young index calculation. It argues that the methods follow ILO and international best practice. It also explains how the new series is linked to the old one. It is not a historical or interpretive study of apartheid; its value to the book is as a methods reference for deflating nominal series and for judging the continuity of the CPI.

Evidence and method

Stats SA documentation: the IES 2005/6 (24 000 dwelling units), the 2005 retail trade large sample survey, the food and beverages industry survey, excise duty data, the 2001 Census, Urban Function Index, rental survey of letting agents, plus appendices listing indicator products and collection methods. It is not based on archival or historical data.

Findings

  • The CPI dates from 1917 and initially covered only large urban areas; coverage was widened later. (p. 4)

    The South African CPI originated in 1917, covering large urban areas only. Since 1997, smaller urban areas were included.

  • The CPI has two objectives: macroeconomic inflation measurement and cost-of-living measurement for indexation of wages, grants and contracts. (p. 4)

    To measure changes in the cost of living of South African households to ensure equity in the measures taken to adjust wages, grants, service agreements and contracts.

  • No prices are collected in rural areas; coverage is urban only, a limit for long-run poverty or deflator work. (p. 19)

    The CPI generally collects and publishes for all areas in which prices are collected except rural areas. No price collection is done in rural areas.

  • Before 2006 prices were collected from head office, mostly by post, and direct field collection replaced this from 2006, a break in collection method. (p. 4)

    Prior to January 2006, all prices of goods and services were collected from the head office of Statistics South Africa (Stats SA) mainly using the post.

  • Food expenditure in the IES was substantially under-reported relative to retail sales data, so the weights were adjusted upward. (p. 16)

    Stats SA arrived at a total value of sales of food and non-alcoholic beverages to households of R143,402 billion compared to the R100,971 billion derived from the IES.

  • Alcohol and tobacco were heavily under-reported in the IES; excise-based data raised the estimate more than fivefold. (p. 17)

    This exercise resulted in an estimated expenditure level of R43,6 billion compared with R8 billion from the IES.

  • Housing measurement changed from mortgage interest rates to owners' equivalent rent, which is the largest single CPI component. (p. 23)

    Owner-occupier housing costs represent the largest single component of the Consumer Price Index. Previously these costs were measured by interest rates on mortgage bonds.

  • Elementary indices use the Jevons formula and higher-level aggregation uses a Young index with fixed weights. (p. 26)

    an elementary index is calculated using the Jevons index number formula. The Jevons index is

  • The classification moved from the International Trade Classification to COICOP, which reshuffles categories (e.g. insurance, medical aid, restaurants) and affects comparability with older series. (p. 11)

    In COICOP all insurance is classified under Miscellaneous goods and services.

  • Weights are updated only about every five years from the IES, and linking new to old series produces non-additivity. (p. 30)

    Linking leads to non-additivity. When the new series is chained onto the old, the higher-level indices after the link cannot be obtained as weighted arithmetic averages