Stats SA stands by its GDP numbers

The following has been issued by Risenga Maluleke, statistician-general and head of Statistics South Africa, in response to the article, Pretoria, we have a counting problem.

According to Stats SA’s Quarterly Employment Statistics (QES) survey released last week, the formal sector of the South African economy employed approximately 10.1 million people who earned R725 billion in the last quarter of 2018. This is an increase of 1.6% in employment and 7% in earnings from the last quarter of 2017.

The QES results are obtained from a survey of payroll reports from businesses, government and other employers. The survey sample is based on a range of administrative data sources including PAYE, UIF and company registrations. A key feature of the survey is that its sample is refreshed each year and adjustments are made to the historically published numbers to provide a continuous time series. The adjustments maintain the growth trend, which is vital to properly understand economic dynamics.

Moneyweb last week published two articles penned by economist Mike Schüssler (Pretoria, we have a counting problem and SA underestimates everything), which cast doubt on the validity of the employment numbers and the measurement of the country’s GDP. Schüssler tracked changes in the historically published figures and concluded that Stats SA had overestimated the growth in employment. Because he has not accounted for the most recent adjustments to the historical time series, Schüssler reports “the wonderful news that employment in our formal sector grew by 16.4% between September 2014 and September 2018”. In fact, the current edition of the QES shows only a 9.3% increase over this period. The growth in formal sector employment indicated by the QES is in line with other indicators, including Sars tax data.

Through some back-of-the-envelope calculations, Schüssler concludes that South Africa’s GDP is understated as a result of not incorporating revised QES figures.

To understand why it is not that simple, one has to get a bit technical.

The GDP is compiled using a wide variety of different sources to track different economic activities and these must all balance to obtain a total value of the economy. A simple example is a factory that has manufactured 100 widgets. These 100 widgets must either be sold to households, government, other businesses, exported or held in storage. They could further be used to make other goods or sold onwards. Information about what has happened to the widgets is available at different times from different sources. Within this accounting system, the GDP is calculated each quarter with an annual total published a year later and revisions made to previously published numbers. 

Let us go back to the earnings data. The changes in earnings reported by the QES are used to derive the growth (or trend) in the GDP each quarter. However, as the QES only measures employment and earnings, it is not designed to balance with any other economic data. The GDP rather relies on data from company annual financial statements for the level of employee expenses, as the accounting system requires this item to balance with other financial metrics.

To further ensure its integrity, the GDP figures are continuously revised as fresh data becomes available and, in line with best practice, even more comprehensive revisions are made every five years when the entire set of national accounts are benchmarked.

Transparency of our methods and responsiveness to the needs of the user community are important principles for Stats SA. This is equally the case for national statistics offices around the world, which follow the United Nations fundamental principles of official statistics. We will continue to use the best methodologies to ensure the accurate and timely provision of economic and social data to all our users within the resources provided to us. Furthermore, we welcome comments from our users, including sharing their views publicly because statistics is a public good.

Risenga Maluleke is South Africa’s statistician-general and head of Statistics South Africa.