The Fed manipulated SOFR
It is now official, SOFR is manipulated by the Federal Reserve Bank of New York. To avoid any misunderstanding, I have to repeat the first meaning of manipulate, which, according to the Oxford dictionary, is "handle or control in a skillful manner". On 31 May, the Federal Reserve Bank of New York has decided not to use the actual trade data to compute SOFR but the "rate was calculated using survey data" as indicated on the Fed SOFR's site reproduced below. This certainly proves that the Fed "controls" the figure as it can decide how it is computed from its own authority. And who I'm to think of disputing that the Fed is skillful. Their actions are perfectly in line with the definition of manipulation provided above: control in a skillful manner.
Update 7-Jun-2019: An article was published on 6-Jun-2019 in risk magazine about the issue: A mystery: why did the NY Fed use a survey to get SOFR? (subscription required).
The only indication on the why, is that there would have been a feed issue. But the only reason for the publication of the rate set in T only in T+1 is to be able to solve data issues. The SOFR print is published in T+1 around 8:00 am and is related to data on transactions done in T before the close of the Fed. The underlying is an overnight rate, so we are certain that the near leg settles before the Fed closes. There is not even the excuse of "late trades". The T+1 publication (I ranted about it in a Speed of information is decreasing blog post) create problems in operational terms and potential asymmetrical information availability (on top of measurability questions). The only justification for this is that there is a strong contingency arrangement that guarantees that the rate is always published according to the rules by 8 a.m., including in case of technical problem. There is more than 12 hours between the end of trading and the publication. If there is a "feed issue", you have time to solve the problem, by sending a USB key with the data by bike or carrier pigeon if needed.
Maybe the feed issue indicates that the Fed's pigeons were not fed (pun intended) properly and went on strike! That would be a nice (finance fiction) title: The Fed's pigeons were not fed, the (food) feed issue created a data feed issue.
Update 7-Jun-2019: An article was published on 6-Jun-2019 in risk magazine about the issue: A mystery: why did the NY Fed use a survey to get SOFR? (subscription required).
The only indication on the why, is that there would have been a feed issue. But the only reason for the publication of the rate set in T only in T+1 is to be able to solve data issues. The SOFR print is published in T+1 around 8:00 am and is related to data on transactions done in T before the close of the Fed. The underlying is an overnight rate, so we are certain that the near leg settles before the Fed closes. There is not even the excuse of "late trades". The T+1 publication (I ranted about it in a Speed of information is decreasing blog post) create problems in operational terms and potential asymmetrical information availability (on top of measurability questions). The only justification for this is that there is a strong contingency arrangement that guarantees that the rate is always published according to the rules by 8 a.m., including in case of technical problem. There is more than 12 hours between the end of trading and the publication. If there is a "feed issue", you have time to solve the problem, by sending a USB key with the data by bike or carrier pigeon if needed.
Maybe the feed issue indicates that the Fed's pigeons were not fed (pun intended) properly and went on strike! That would be a nice (finance fiction) title: The Fed's pigeons were not fed, the (food) feed issue created a data feed issue.
Comments
Post a Comment