2017-03-15

Change of benchmark overnight index is a difficult task: interview

An interesting follow-up article in Risk magazine about Swiss rate reform: Swiss rate reform in race against the clock (subscription required). With respect to the previous Risk article on the same subject, it presents a more balanced view and not only the point of view of the WG on benchmark reform. At lot of the opinions expressed are closer to what could be found in my first blog on the subject.

Chris Davis, the author of the Risk article, found my blog on the subject and contacted me. We had a roughly 30 minutes discussion. He used a lot of the background information from the discussion and from my blog for the article, even if only two sentences from the discussion were quoted.

I was surprised that my blog was not explicitly cited in the references, as would be customary when the blog presented original research and analysis and was used for background information. On my side I will continue to reference Risk explicitly when I read something of interest in it.

On top of the subjects I introduced in my previous blogs, one element that I did not forecast was presented. This is that the current TOIS market already prices the change from TOIS to SARON. In some sense, this means that the TOIS market is not the TOIS market anymore; it is a blending of TOIS and expectation of conversion to SARON.

This is a quite interesting point of view from a pricing perspective. The actual market includes the "legal jump risk" from one index to the other without compensation. The pricing of the benchmark change is now impossible, the market has already been distorted by evoking the possibility of the change to a new benchmark without compensation. That creates more uncertainty for potential disputes. In some sense, the risk is becoming impossible to hedge. If you want to price or hedge pure TOIS, there is no market for it. The fact that there is a legal jump risk in the actual instrument probably make the market incomplete. Certainly it can not be treated with a diffusion process; you have to include a one-off jump.