US IRS volumes - SEF MAT cliff or steady decline?


A couple of recent pieces suggested a dramatic downturn in SEF IRS trading in week 1 of MAT.   Such short-term effects may be shown significant by more weeks of data but my analysis suggests there is a steady decline since the beginning of the year across both on and off SEF IRS implying that Basel III and Dodd-Frank ET rules are more the causes.

The pieces

Risk published an article basedon figures announcing cleared US dollar swap volumes drop 80% on the first day of the SEF regime (subscription required).  This was subsequently qualified / tamped down by contributor Amir at Clarus.

Kevin McPartland posted on his blog Kevinonthestreet that SEF volumes died last week dropping 64% from the 2014 weekly average SEF traded volume.  The short week exaggerates this drop but looking at SDRview I can only get to about a 40% drop on IRS fixed float volumes.  I am investigating and will update this post as I track down the discrepancy between SDRView and SEFView.  

The larger point I discovered in looking at the data which neither post covers is that 2014 has seen a steady decline in IRS volumes rather than a sudden drop at the MAT date.  Also off SEF volumes declined along with on SEF volumes rather than shifting on SEF.

Useful US reported data sources

ISDA's new service shows a clear steady-ish decline in cleared CDS and cleared IRS from since the beginning of the year.   Try selecting transaction data and then pick CDS and IRS in turn, select all currencies, products and tenors and select cleared or uncleared or both as you wish.  Similar downward volume trends are evident across both asset classes and across cleared and uncleared swaps.

Clarus's SDRView (subscription or trial required) enables us to look at weekly data.  Try selecting weekly IRS or CDS data since the beginning of January pick all products and G4 currencies and you'll get a similar story whether it be on or off SEF.  

Clarus's SEFView (subscription or trial required) enables only SEF volumes to be viewed.  Given imperfect data from SEFs (and possibly from SDRs) there are some differences between SDRview on SEF volume and SEFview volumes.

So why the decline?

Several possible effects are coincidental and hard to disaggregate:

1.  SEF MAT?  It's possible the MAT deadline encouraged people to suddenly shift off SEF.  Even after adjusting crudely for President's day holiday there's still a week on week decline but hard to know whether that was because of week long vacations or actually due to MAT.  Most likely people have been making their decisions and moving on or off SEF over a longer period.  Certainly on SEF figures didn't decline hugely differently from off SEF based on the above figures.  So far it seems like SEF MAT didn't cause the big picture trend.

2.  Dodd-Frank extra-territorial (ET) rules?  It's possible that US reported swaps declined and shifted volume to Europe.  My post on ISDA's analysis confirmed that some of this is going on in the D2D market at least.

3.  Basel III bank balance sheet pressure?  CRD IV risk capital went live Jan 1, hard leverage ratio and liquidity coverage ratio are ratcheting up over time - all encouraging either reduction in trading or more capital efficient trading models (e.g. cleared riskless principal, agency execution).

4.  Substitution by swapfutures?  Not yet.  ERIS and CME DSF volumes are steadily advancing and combined they are maybe $18 billion open interest currently with ERIS recently breaking through $10 billion.  Given the position netting involved this represents considerably more notional executed than $18 billion.  Given I'm seeing an average daily decline of several billion in USD fixed float IRS since the beginning of the year it would take a large execution to open interest ratio to make swapfutures a signficant contributory factor.

5.  IR market conditions?  Could be but I don't have easy access to a market view of e.g. Fed taper or anything else on IR levels / volatilities.  Views gratefully received.
6.  Full weeks holiday in Vermont.  Amir's post (link above) on Clarus suggests not.

Further analysis of US reported volumes

If you look at the graph in the summary "brick" for this article (hit home and scroll down until you see the brick) you can see that US reported (i.e. DTCC SDR) volumes both on and off SEF declined from the beginning of the year.  This suggests either curtailment or shift offshore of US trading.

Certainly the SDR reported figures for on SEF volume below suggest the 2014 YTD decline in on SEF volume is bigger than the one week impact before and after the MAT deadline.  This runs counter to the Risk and Kevinonthestreet posts which imply a cliff effect.

G4 IRS decline vs January averagea
  w/o Feb 10b w/o Feb 17c
On SEF -19% -25%d
Total -16% -24%d

Source: Clarus SDRView.  Analysis:


a.  "January average" is a simple arithmetic mean of the four weeks from January 6th  

b.  w/o Feb 10th is the week before MAT go live

c.  w/o Feb 17th is the week after MAT go live

d.  Ajusted for the short week (President's day) by mutliplying weekly G4 IRS fixed float notionals by 5/4

My take

To judge relative impact of 1, 2 and 3 above we need non-US reported volumes to assess the degree to which this decline is regional (i.e. SEF MAT / ET driven) or global (i.e. bank Treasury / Finance groups pushing trading businesses to optimize Basel III capital).

To the extent this is SEF MAT / ET driven, another few weeks data will start to tell us whether the US decline is a temporary implementation hiccup or a permanent curtailment / offshoring of US trading.