Building blocks of CPI tap into our evolving lifestyles
Statistics SA has announced that the inflation rate for January was 6.6% — a drop from 6.8% in December. The January figure is the first of a new four-year consumer price index (CPI) series based on an updated product basket and weights. These changes not only determine the measurement of inflation but provide insight into the changing spending patterns of South African households.
The CPI basket is the list of products for which prices are surveyed and which are used for calculating the index. In order for the CPI properly to measure the inflation experience of households, these goods and services must reflect the purchases of consumers as closely as possible.
Of course, it is not possible for the CPI to mirror each individual’s purchases. Rather, the weights and basket are designed to represent the total expenditure of all households, based primarily on a detailed expenditure survey.
The timing of this update is well within the five-year limit set out in international standards and has not been accompanied by any major changes in the methods used to compute the CPI.
Changes in the basket show the evolving character of South Africans’ lifestyles. Convenience foods such as instant noodles, ready-mix flour and frozen pies and pizzas feed fast-paced lives.
Increasingly, middle-class families live in sectional title estates. Almost 5% of Gauteng households live in sectional title estates, compared with less than 2% in 2002, according to Stats SA’s General Household Survey. Levies for these developments are now in the basket, but stand-alone freezers have disappeared — a casualty of smaller homes.
The increasing dominance of the internet and electronic-based services has seen the demise of postage stamps, DVDs and blank CDs. Video games have entered the basket for the first time.
These lifestyle changes are characteristic of the middle class, which represents the bulk of consumer spending but the minority of the population. The wealthiest 10% contributes 48% of the total CPI weight; the poorest 10% contributes 0.5%.
To aid understanding of the differential effect of inflation on different strata of society, Stats SA will publish inflation rates for 10 economic segments from the very poor to the most affluent 10%. This talks to the significant inequalities found in SA.
Basket inclusions such as chicken giblets and beef and mutton offal products aim to better track the consumption patterns of poorer households.
Stark differences in spending power differentiate provinces. Gauteng has long been the biggest spender, with 36% of national expenditure. At the previous reweighting KwaZulu-Natal and the Western Cape both had a weight of about 15%. The Western Cape now has a share of 17%, significantly higher than the 13% of KwaZulu-Natal. The Northern Cape has the smallest contribution at just 2%.
Each province has its own set of weights. For example, the provinces with the highest and lowest weights for food are Limpopo (22.5%) and Free State (14.9%).
Each product in the CPI basket has a value or weight attached to it, which is proportional to the total amount spent on that product by all households in the country in a 12-month period.
These weights and the selection of products in the basket, are calculated using a number of data sources.
Most important is a survey of households’ expenditure conducted over a full 12 months in which each sampled family is expected to record every single item they pay for.
Fortunately for the approximately 25,000 selected households, they need to give this information only for two weeks, after which the interviewers rotate to the next sampled household.
Despite the best efforts of Stats SA’s interviewers, there is always underreporting of spending. This happens for several reasons.
First, it happens with products that are bought frequently and by different family members — food is the most affected group. Underreporting also happens when respondents may be embarrassed to disclose the full value of their purchases of alcohol or cigarettes.
Difficulties in interviewing residents of security complexes and well-off areas lead to under-reporting of spending on items such as education, health and entertainment.
To compensate for these and other technical factors, Stats SA consults a wide set of additional data sources. These include surveys of retail businesses; information compiled by government agencies, regulators and industry associations and, the financial results of service and retail businesses.
All values are calculated for 2015 and the final weights are obtained by inflating them with an appropriate price index to December 2016. This month now forms the base period for the CPI, where all indices are set to 100.
The biggest shifts in weights from January are an increase in the importance of food and a decrease in transport. Housing continues to attract the largest share of household expenditure (25% of headline CPI).
As food is a basic need, the amount that people buy remains relatively stable. Expenditure therefore rises when prices have increased significantly, as they have over the past year. Food now attracts the second-largest proportion of expenditure — up to 17% from 15%. The weights of meat and dairy products have increased more than most other foods.
The transport category has dropped in weight, largely due to the sustained moderation in the petrol price, which has increased only 7% since January 2013.
Patterns of consumer spending do not shift markedly and this is illustrated by a technical calculation using the new weights and applying them to the pricing information from the previous period. This is known as a Paasche index.
Index theory proposes that an average of the original indices and the Paasche index — known as a Fischer index — is the ideal inflation index.
Calculations by Stats SA show that in December 2016, the annual inflation rate would have been 6.1% using the Paasche index and 6.5% using the Fischer. This finding confirms the theory that consumers shift their spending from products with higher inflation rates to those with lower rates, therefore, leading to a Paasche index being lower than that published.
Stats SA aims to run an expenditure survey in 2019. This will form the basis for re-evaluating and updating the basket and weights to reflect how consumers have changed their buying behaviour.