Statistics South Africa (Stats SA), the country’s official provider of national statistics, is tasked with producing reliable data on a regular basis. The COVID-19 pandemic, and the related lockdown measures, have caused a severe economic contraction in 2020, but have also resulted in confusion over the degree of the pain felt by the real economy over the last few quarters. In 2020’s second quarter, Stats SA reported a headline contraction of South Africa’s GDP by 51.0%[1]and (mystifyingly) an improvement in the unemployment level to 23.1%. In 2020 Q3, unemployment levels deteriorated to a 25-year low of 30.8%. This note briefly unpacks the method employed by Stats SA to estimate the reported changes in GDP and unemployment to better understand how the statistics office has arrived at these GDP and employment numbers.
Explaining GDP changes
In the second quarter of 2020, South Africa’s GDP contracted by 51.0% based on the seasonally adjusted and annualised rate (SAAR). The SAAR method projects the annual growth rate assuming the same quarter-on-quarter growth rates over four quarters each year. In other words, if the same demand and supply shocks experienced between March and June 2020 persist throughout the remainder of 2020, South Africa’s GDP is expected to contract by an astronomical 51.0% in 2020, eliminating half of the economy. While the SAAR method is the standard method used by Stats SA to compare GDP movements, this method is misleading, and uninformative during volatile periods.
Rather than employing the SAAR method, the real performance of 2020 Q2 can be assessed using two other methods: comparing to the preceding quarter (quarter-on-quarter change) or comparing to the same quarter in the previous year (year-on-year change); with the latter method also applied by Stats SA. Figure 1 shows that these three approaches yield different GDP growth rates for 2020 Q2. With the quarter-on-quarter changes unable to resolve seasonality effects, looking at year-on-year changes provides a more prudent measure to assess how the economy has performed under the pandemic.
The year-on-year changes indicate that South Africa’s real GDP declined by 17.1% in 2020 Q2. Such a contraction is chiefly explained by the supply shock induced by the restrictions in the sale, production and distribution of non-essential goods and services such as alcoholic beverages, tobacco products, the wholesale and retail of electronic products, along with the cessation of non-essential manufacturing, mining and construction activity. Even though this figure is less acute than the SAAR (the earlier mentioned 51.0%), such a reduction is South Africa’s worst recorded performance in close to a century.[2]
Figure1: Approaches to calculating changed in real GDP
Source: Statistics South Africa
Explaining the unemployment changes
In 2020 Q2, South Africa’s official unemployment rate improved by 6.9%, improving from 30.1% in 2020 Q1 to 23.3% in 2020 Q2. By 2020 Q3, the official unemployment level fell to 30.8%, a 30-year all-time low. Even though the number of employed individuals increased by 543,654; over 2.2 million individuals lost their jobs (Table 1).
Table 1 : Analysis of changes in the QLFS key indicators
Source: Stats SA QLFS
The technical improvement in the unemployment rate in 2020 Q2 was primarily explained by the ‘not economically active’ cohort of the labour force. In considering the unemployment level, Stats SA looks at three key aspects: the proportion of individuals employed; unemployed; and not economically active within the population group of 15-64 years of age. By definition, the ‘economically active’ population includes those who are employed, and all individuals that were (i) not employed in the reference week, and (ii) had actively looked for work or attempted to set up a business, and (iii) would have been able to start work in the reference week, or (iv) had not actively looked for employment but had secured employment in the near future.
Given the lockdown restrictions, a significant proportion of the respondents had not engaged in these activities. This resulted in close to 5 million individuals, being deemed to be ‘not economically active’ in 2020 Q2. However, the easing of the lockdown regulations resulted in more individuals seeking job opportunities, and therefore being reclassified as economically active but unemployed. Added to company retrenchments, the unemployment level reached 30.8%. Unless the economy rebounds soon, unemployment is likely to increase further over the remainder of 2020.
The COVID pandemic has undoubtedly posed new challenges for our official statisticians, both in the collection and reporting of critical data. Stats SA was faced with the novel task of gathering economic data during a time when traditional methods of data collection were constrained by health and safety protocols. Instead of conducting face-to-face surveys with a representative sample of economic actors, Stats SA relied on computer-assisted telephonic interviews, a common survey method. The lack of up-to-date contact information and forced firm closures resulted in lower response rates. While low response rates undoubtedly affect the robustness of the official statistics, the national office noted that additional measures were employed to ensure that the economic indicators published in the first two quarters of 2020 are reliable.
Sound policy making requires access to timely and reliable data. As such, it is important to ensure that state agencies are adequately capacitated and resourced to collect, analyse, and report accurate official statistics. Importantly, Stats SA should use this as an opportunity to make better use of technological advances and novel methods of data collection and gathering. Lastly, it is recommended that Stats SA engages in outreach and advocacy initiatives to ensure that citizens are equipped to understand and interpret movements in key economic indicators.
[1]The GDP data for 2020 Q3 will be released on Tuesday 8 December at 11:30.