A PrimeX primer, also featuring ABX | FT Alphaville

A PrimeX primer, also featuring ABX

Last week, Fitch completed a review of U.S. Prime RMBS looking at transactions involving thousands of bonds. As a result of their review, 42 per cent of the portfolio was downgraded. Since then, a tradeable index linked to a subset of prime mortgages has been falling in value and was last seeing travelling towards par.

For an overview of the situation, along with an explanation for your feeling of déjà vu (re: ABX), we give you this video starring former Alphaviller Tracy Alloway, recorded on Monday – click to view.

So what is the Markit PrimeX exactly? And why on earth does it exist? To discuss, let’s take a look at the family which the index belongs to.

The Synthetic RMBS Family

There are a number of tradeable indices that reference pools of bonds where the pools of bonds themselves reference a pool of mortgages. For more on how the process of securitisation works, FT Alphaville refers you to this classic slideshow primer that happens to use subprime mortgages as an example out of all the pools of mortgages that could have been selected.


ABX – the classic, the one and only. This family of indices reference 20 of the largest subprime deals out there that met certain criteria (including the “requirement” of having a FICO score under 660). From those 20 selected deals, five indices were created, one for each of these rating categories: AAA, AA, A, BBB, and BBB-.

The five indices were born at the height of the subprime craze. They only “rolled” four times over 2006 and 2007, meaning that there are only four versions of each category of index, i.e. four for AAA, four for AA, and so on.

After that, issuance was so poor that it wasn’t possible to create a new “version” of the indices (i.e. that reference new deals). Series 08-1 had only five deals that met the criteria for inclusion hence the roll was delayed. Infinitely, as it turns out.

And in case you are wondering who traded this errr.. stuff, here’s a handy table from the marketing presentation.


There are also TABX indices that provide the “opportunity for more granular exposure to the BBB and BBB- rated tranches of the ABX.HE.” We are guessing that those didn’t go so well..

Those two index families were key to getting synthetic exposure to non-agency subprime RMBS during the crisis. They enabled more positions to be placed on subprime that mere cash markets would allow. But let’s not get all emotional about it. [Sniffle].

An important distinction between how these indices operate versus credit indices like the Markit iTraxx and CDX, is that these are “pay-as-you-go”. Under that model, which ISDA had the courtesy to standardise in order to allow an incredible amount of leveraged subprime exposure to build-up in the system alleviate otherwise overly burdensome operational constraints, “soft” events trigger payouts too, as this diagram from a Bank of America primer illustrates.

In the ill-fated words of Robert Pickel, Executive Director and Chief Executive Officer of ISDA at the time: “The revisions [to the PAUG documentation] respond to evolutions in the CDS on ABS product area and will help these instruments continue to thrive. With each of these forms, ISDA is assisting continued innovation in the application of credit risk mitigation tools.”

(OK, quick poll: when was the last time you heard the word “innovation” used in the context of derivatives and you didn’t flinch?)


PrimeX – What do you do then if new deals are not forthcoming and the old deals have blown up in your face? The answer is, you go back and make indices out of old deals that you didn’t bother making synthetic products out of the first time! It helps if they haven’t completely blown up.

Among the mortgage securitisations that are still breathing, there are some agency RMBS (more on those later) and non-agency securities that reference prime mortgages.

Out of the non-agency primes, there are two major classes – one that ultimately references hybrid adjustable rate mortgages (ARMs) and another that references fixed rate mortgages. To be included in the indices, mortgage pools for the first type require a weighted-average FICO score of above 730 and the later a score above 735. The deals referenced were put together between 2005 and 2007, but here’s how the indices referencing those deals have performed more recently:

It’s that more recent dip that’s getting everyone worked up. And given the deteriorating fundamentals highlighted in the Fitch report, it shouldn’t be terribly surprising.

To the heart of the matter though – how big is the market for this stuff anyway? At the end of last week, all four indices together had $10.9bn gross notional outstanding, a net figure of $4.0bn and a total of just 711 contracts. Leaving FT Alphaville to wonder if you should even cough in the general direction of these. A year ago, the amounts were $6.7bn gross, $3.4bn net, and 444 contracts.

Amazing how adding a bit of volume data changes things, isn’t it?


MBX, IOS, PO (don’t go there) – these indices reference various bits of agency securities issued in 2008 and 2009, were launched over the course of 2010, and they get even more complicated in terms of flows of payments. This diagramme for the MBX helpfully makes it look a lot more simple than it actually is:


By now, you are probably wondering how much volume there is on these indices, versus ABX versus PrimeX. Well…

ABX – $84bn gross; $20bn net; 2,976 contracts
MBX/IOS/PO – $135bn; $36bn; 4,472 contracts
PrimeX – $11bn gross; $4bn net; 711 contracts

For the purposes of comparison, some more familiar single-name CDS…

Italy – $310bn gross; $22bn net; 9,830 contracts
China – $55bn gross; $8bn net; 5,820 contracts
Denmark – $20bn gross; $3bn net; 1,992 contracts


Before being dismissive of PrimeX though, we need to look at activity. After all, the indices do seem to be responding to fundamentals even if there are relatively few contracts outstanding.

Last week there were 28 trades involving a gross notional of $170m. The week before last, there were 57 trades involving a notional of $550m. And the week before, just 9 with a notional of $155m. That is one inactive canary, coal-mine or no.

That said, these are seriously operationally-involved indices because of how the payments work, so even with the low volume, they might, just might still be a reasonable indicator. Just don’t take your scepticism hat off.

Not that we think you ever would, of course…

Related link:
PrimeX – The Time For The Next “Subprime Trade” Has Come – Zero Hedge
Trade Information Warehouse – DTCC