The Financial Services Board (FSB) said on Friday it is looking into a programming error that caused the JSE to miscalculate inflows into SA equities by R98bn.
The JSE reported on Sunday that instead of foreigners investing a net R98,1bn in their South African portfolios in May, June and until 20 July, they were net sellers of an accumulative R36,4bn of South African equities in May and June and buyers of just R50m in July.
The bourse blamed the error on a computer programming problem which impacted the manner in which data was extracted from core transactional systems.
But the JSE corrected the glitch on 21 July and has apologised for the error.
“This is unfortunate and as soon as we found out that we had an error, we addressed it immediately and corrected it,” Leanne Parsons, director of information services at the JSE, said on Monday.
The FSB, however, moved on Friday to reassure the market that it was looking into the circumstances of the lead up and causes of the error, and said it will be taking appropriate steps in this regard.
Parson said the error only related to those “particular statistics” of non-residents and that the incorrect data came about due to a system update.
“It was correctly programmed initially but there was an improvement made to help performance and unfortunately, inadvertently, it didn’t translate into exactly the same program being implemented.”
She said it was the first time that the JSE had any error in its non-resident equity stats.
Meanwhile, Sasfin Securities’ deputy chairman, David Shapiro, said investors and traders should not be alarmed by the JSE’s computer glitch.
Shapiro said the real cash that flows would reflect in prices of shares, the rand and bond yields. “That’s already there,” said Shapiro.
However, the FSB said it was important that the reliability of information released to the public by a market infrastructure, in this case an exchange, was assured and maintained.
“The error has raised concerns in the market and among fellow regulators,” the FSB said in a statement.
