How Much is a Kilo?

By Tomas Kabina, WageIndicator / Celsi salary and cost of living survey manager

Testing the Cost of Living Survey

The value of our statistics, of all statistics for that matter, is only as good as the quality of the basic input allows them to be. Therefore checking at the source, where the figures are given/collected, remains of prime importance. Recently WageIndicator has tested its Cost of Living survey, which is essential for the calculation of living wages. In this particular instance the test was carried out earlier in 2015 in South Africa. The results of the test were presented at a special seminar during one of the regularly held Workers Colleges in the capital of KwaZula Natal province, Durban. The Workers’ Çollege is a private initiative offering upgrading courses for activists, both from trade unions and from community work throughout South Africa and beyond. WageIndicator was present in the person of Paulien Osse, who introduced WageIndicator and the living wages as part of Wages in Context and by Tomas Kabina, from WageIndicator and CELSI, who explained the experiment and its results to some 70 South African students. Many of those present had previously contributed their offline collection of prices to the test. What follows is an extract from Tomas’ presentation. 

The Online-offline Experiment

You, the students of the Workers College, have collected a total of 28,731 prices from the province of KwaZulu-Natal. WageIndicator has collected online 8,395 prices from KwaZulu-Natal in 18 months (2014 to June 2015).

Through the face-to-face surveys you conducted, you have collected 3.4 times more prices than we have collected online, which represent 77.3% of all data collected in KwaZulu-Natal. Well done!

In the online survey, we collect prices for 80 different items:

  • 60 different food and drink items
  • 3 transport costs items (one-way ticket, monthly pass, 1 liter of gasoline)
  • 6 housing costs items (rent of 1-bedroom/3-bedroom apartments in/outside of the city centre, energy & utilities, 1 min. of mobile tariff)
  • 11 average monthly items: net salary, expenses on various needs (food, housing, transport, education, clothing, medical, culture, drinking water) + living expenditure (wage) according to the respondent’s opinion (both for individual and for family).

In the face-to-face (offline) surveys, you have been collecting the same set of 80 important item prices (that was the intention).

The prices we collected were subsequently compared with real prices from the South African Woolworths supermarket chain.

Combined, these prices are used to calculate the living wages. 

Consumers do Not Buy Kilos but Packs

In the survey, we usually ask people on food prices for the standard metric units (1 kg, 1 liter …)

Now our very important finding: people know quite well the prices of the food products in the packs/baskets they are usually sold in, but they don’t have a clue about sizes/weights in which they buy them! They seem to always answer the price of the usual pack/basket. This important research finding (to which you have contributed a lot) has prompted us to change the sizes/weights in the survey. In the future, the respondents will have to be able to choose the size/weight from several options. Then, the prices will be rescaled to standard units to allow international comparison.

This problem doesn’t occur for drinks by the way, as those prices are usually asked for the volume they are sold by (1 liter of milk, 1.5 liter of bottled water…).

In Conclusion

We have collected a lot of offline (28,731) and online (8,395) item prices from KwaZulu-Natal, and compared with real prices in the Woolworth supermarket chain:

  • You have done a great job, collected many prices!
  • People don’t know the sizes/weights of food products they buy, they answer the packs/baskets how it’s usually sold.


And one more important finding: online respondents answer a bit higher food prices than offline respondents, but a lot higher housing/energy prices and their own monthly expenditure. Presently we explain this discrepancy by assuming that since the online e respondents have answered higher monthly expenditure, it seems they must have higher salaries! Respondents answering online surveys usually might have good education à their salaries might be higher à they probably can afford more expensive housing and other goods à therefore they report higher prices.