FCA short data looking glitchy14 Jul 2018 17:12
Looking in some detail at the FCA's spreadsheet of short positions leads to a strong suspicion that a substantial part of the data is erroneous.
(btw the site which most people use, shorttracker.co.uk, is a front end for the FCA data)
For Sirius, the current data is Citadel 0.64%, JPM 0.69%, Och-Ziff 1.21%, Highbridge 1.63% ... total 4.17%
I'm going to argue that the Citadel and JPM numbers probably are not current, and the correct total is probably 2.84%
***but*** before anyone gets too excited, please note that the data *anyway* includes only positions above 0.50%, so a reduction (even a substantial reduction) in the total reported by the FCA does not have a clear implication for the likely current actual total, counting all participants. Which could actually be very large, since a CB holder can short up to 35 bonds-worth of shares without anyone except the FCA knowing.
Only city insiders would have a good grasp of the actual total short position, based on gossip and intuition concerning what proportion of short positions less than 0.50% there are, which is completely unknown to us. (Though I do sometimes wish that a city insider would post here from time to time).
Anyway, here are the details. The rules say that changes in short positions must be publicly notified at every 0.1% increment or decrement when above 0.5%, and when dropping below 0.5% having been above.
This gives rise to a typical behaviour, where an institution makes an initial notification above 0.5%, followed by a sequence of changes over a few months, typically slowly increasing and then slowly decreasing, followed by a final notification slightly less than 0.5%. When the FCA receives this final notification, it removes the institution from list of current shorters.
Looking at the historical FCA data, one can see this pattern, with variations, over and over again.
Now, what would happen, you may ask, if the final notification is omitted, or gets lost in the system?
What happens is the institution's reported short position gets stuck on the last number which was reported before it dropped below 0.50%. i.e. typically it will be a number slightly above 0.50.
It is essentially a bug in the FCA algorithm that a single lost notification can cause this to happen. If an intermediate notification gets lost, it doesn't much matter, because there will probably be a further notification coming along in a few weeks which will correct the situation. But if the final notification gets lost, there is no way for the situation to be corrected by subsequent notifications.
By viewing the spreadsheet in order of ascending date, one can see that this (very likely) has happened numerous times. According to the FCA, Teton Capital Partners have a short position of 0.67% in Nanoco, unchanged since November 2012. By the time we get to 2017, there are several dozen almost certainly stuck short positions. Reported positions unchanged since the early months of 2018
Sign In
Follow the stocks
