|
Related Topics |
|
Today's blog takes a look at a massive blunder made by the Johannesburg Stock Exchange (JSE). They initially recorded net purchases of shares by foreigners to the tune of around R70billion in May and June 2016, and then later corrected the figure and stated it was in fact net sales of around R36.4billion. Wow! Thats a reversal of almost R106billion. Instead of R70billion being invested by foreigners in SA shares, they actually sold R36.4billion's worth.
|
JSE, The Reserve Bank (SARB) and Statistics South Africa (Stats SA)
Now the JSE stated that immediate steps where being taken to avoid such an error reoccurring. Wow that is reassuring. Wonder how they picked up on the error? Did an outside party such as SARB question their figures? Or did someone within the JSE pick up on the error and notified their bosses?
The number that was misrepresented by the JSE can have a significant impact on policy setting as it shows the level of cash flowing into or out of South African liquid assets such as stocks and bonds. Massive outflows might have convinced SARB's monetary policy committee to increase interest rates to lure foreign investors. Someone should ask SARB's MPC members if their vote would have been different if they knew the correct figure of the net sales (instead of buying of shares) for May and June 2016.
Now imagine South Africa's statistical office made this blunder. It would have been headline news in all papers, not just the business papers. And immediate action against various individuals would have been demanded by a whole host of groupings, be it political parties, private institutions such as investment managers or the public at large. Not to mention Statistics South Africa's credibility would have been shot to pieces if they made a mistake like this.
But because this mistake was made by a private firm, it seems to be ok. From what we can tell it is business as usual at the JSE, sure they a little embarrassed, but it will blow over and in a few weeks all will be forgotten. JSE has not been very open on what exactly caused the error (other than saying its programming error), and how it will be rectified. Will someone be fired for this mistake or just a small reprimand, if any action is taken against anyone at all.
Has the Reserve Bank ever discussed data quality with the JSE? Or do they just take the data received from the JSE and use it without fully understanding how it was compiled or obtained? Seems a little risky, especially if such data can have a significant impact on monetary policy.
People still remember mistakes made by Stats SA many many years after the fact, and if they made a mistake like this, demands for heads to roll would be streaming in. Considering the vast number of data sets Stats SA publish, compared to the JSE, one should take a step back and commend the work Stats SA is doing.
Perhaps the JSE should send a few staff members to Stats SA for training in data validation and checking before publishing any more numbers. Question now is can any of the data sets published by the JSE be trusted. Was this big blunder the tip of the ice berg or a small blip on their otherwise pretty clean record for the JSE?
Has the Reserve Bank ever discussed data quality with the JSE? Or do they just take the data received from the JSE and use it without fully understanding how it was compiled or obtained? Seems a little risky, especially if such data can have a significant impact on monetary policy.
People still remember mistakes made by Stats SA many many years after the fact, and if they made a mistake like this, demands for heads to roll would be streaming in. Considering the vast number of data sets Stats SA publish, compared to the JSE, one should take a step back and commend the work Stats SA is doing.
Perhaps the JSE should send a few staff members to Stats SA for training in data validation and checking before publishing any more numbers. Question now is can any of the data sets published by the JSE be trusted. Was this big blunder the tip of the ice berg or a small blip on their otherwise pretty clean record for the JSE?