Swap Data Repository, Block Trade Rule, the Bad News

Under the Dodd-Frank Act, the CFTC issued Part 43 of its regulations to implement the real-time reporting mandate.

This has a requirement for “real-time public reporting”, which is defined as “to report data relating to a swap transaction, including price and volume, as soon as technologically practicable after the time at which the swap transaction has been executed.”

This “public dissemination” is operational in DTCC DDR and may be seen in SDR View.

This article discusses the impact of the Block Trade Rule, which will become effective on 30 July, 2013.

First a definition

Block trades are most commonly known in the Equity markets as large share sales, much larger than market size and above an exchange specified minimum that are executed away from the market. The same concept exists in the Swap markets. After execution block trades are allocated (divided into chunks) to distinct legal entities (funds) of the asset manager or hedge fund group. For the purposes of the Block Trade rule, large notional trades are treated in a similar manner, though these would not typically be allocated e.g. a firm issuing a bond in the primary market, would swap the coupon into floating and for a large bond issue, this would be a large swap trade.

Second a little history

In the proposal stage of the CFTC rule, market participants gave the feedback that block trades should not be publicly disseminated in real-time as that would give an opportunity for others to take an opposite position and so increase the risk and cost of hedging these block trades, which would result in less liquidity in the market. An entirely sensible position.

It was agreed that a public dissemination time-delay for block trades was appropriate, pending a final rule on minimum block sizes.

An interim time delay of 30 minutes and an interim notional cap size was agreed for all trades. Meaning that for block trades the actual notional would not have to be disclosed publicly.

Where we are now – the Interim period

We are currently in the interim period, meaning that the trades you see in SDR View sourced from DTCC DDR have a time delay of 30 minutes or 1 hour and capped notionals that vary depending on asset class and maturity. So for USD IRS trades, less than 2 years the cap is $250 million, between 2 and 10 years it is $100m and above 10 years it is $75m.

Where we will be from July 31 – the Initial period

Only block trades will be subject to a time delay, so all other trades will be as real-time as technologically practicable; which I assume to be within 60 seconds.

The capped notional sizes will be increased according to a 50% methodology rule. So for USD IRS Trades, a 2Y Swap cap will be $460m, a 5Y Swap will be $240m, a 10Y $170m and a 30Y $120m. So significant increases.

Where we will be from Q2 2014 – the Post Initial period

The time delay for dealer block trades will be reduced to 15 mins, except for off-facility not-subject to clearing which will be 30 mins.

The interim capped notional sizes, rather than being dispensed with altogether (as was the original intention)  will be increased to a 67% methodology rule. This means that while the actual size of these trades will be reported to regulators, it will not be publicly disseminated.

The Bad News

I believe capped notionals are bad for transparency  and un-necessary with the time delays in-place to take care of the hedging liquidity concern outlined earlier.

The other argument put forward that disclosing a large block size will somehow not preserve anonymity, does not resonate with me. I can see that rounded notionals for odd size trades is plausible in preserving anonymity; but the same argument for large trades leads to an acceptable loss in transparency as these trades will never be disclosed to the public.

More Bad News

When aggregating gross notionals, as we do in SDR View, to show the volume of activity on a given day or week; the numbers significantly under-estimate the true volumes traded.

The effect is particularly pernicious in the current interim period and much larger than I expected.

Using 31 business days of date from May 1 to June 14, with a trade population of 29,000 USD Swaps:FixedFloat, I found that the average percentage of capped trades was 30% in trade count terms, so less than 1/3 of trades  were larger than the notional cap. However the same percentage in notional terms was 57%. So put another way just over half the notional figure is under-stated from its true figure! Much higher than I would have expected and I believe the Cap sizes have been set too low.

The question is can we make a guess at the size of the understatement?

Well we can try and put a range on the number. So if we assumed that block trades are between 50% to 100% higher than the capped notional, we would say that the publicly disseminated total notional was being understated by between 30-55% or between $10-20billion a day.

The Good News (What there is some?)

Absolute numbers may be lower, but we can still count on the trend in numbers being correct.

Why? Because this is true as long as the understatement of notionals is consistent over our period of Jan to Jun 2013; which amounts to answering yes to the question “are roughly the same proportion of block trades done every day or week?”

Even better news.

July 31st is less than 6 weeks away, so with the rise  in capped notionals detailed above we should see a a large reduction in the notional understatement.

This will unfortunately make comparing volumes pre July 31st to post July 31st a little tricky. But that is a small price to pay for better coverage of the transaction volumes.

In the meantime we will add a feature in SDR View to separate out the capped notional portion from the total amount.

And from April 2014, when CFTC publish higher capped notionals for the post-initial period, the picture becomes better.

And the point of this article is?

Apart from a journey into the depth of arcane detail?

We believe that a time-delay for block trades should be sufficient by itself.

Capped notionals reduce transparency in the Swap market.

Capped notionals understate the liquidity of the Swap market as compared to other markets.

And should be dispensed with altogether.

Views on a postcard please.

Or better still as a comment where you read this article.

References

For those of you interested in reference materials and with a few hours to spare, I can recommend the following nighttime reading:

Davis Polk write-up on Block Sizes

Gibson Dunn write-up on Real-time public reporting of Swap data

CFTC Federal Register Vol 78, No 105 (all 80 pages!)