This post evaluates the implications of yesterday’s South African GDP release for potential growth estimation. The chart compares SARB estimates and the IMF’s implied estimates to those of a simple Hodrick-Prescott (HP) filter. Updated HP-filter based estimates of potential look through the COVID-19 fall in GDP, attributing more of the decline in output to a larger output gap than alternative approaches. SARB has progressively lowered its estimates of potential growth since it first began publishing estimates, and has revised its estimates over recent MPCs to account for the base effects from the COVID-19 lockdowns. The IMF has also lowered its projections of South African potential growth over time (here backed out in the chart using their real time output gap and real GDP estimates). The IMF’s projections imply a slightly lower profile of potential growth over the medium-term than SARB’s.
Statistical filters can be sensitive to large shocks such as the pandemic represented. In a future post, I will present real-time estimates of several statistical filters from our output gap suite to demonstrate which techniques have provided the most stable estimates of potential growth.