Repo On The African Plain

| About: PowerShares DB (UUP)

That the repo market, as noted yesterday, has been beset by a persistent collateral shortage is relatively uncontroversial. Where once large blocks of MBS tranches were central to interbank flow and funding, their absence is still a fact of operation though that repudiation was a very long time ago. Even with that backdrop, however, it doesn't explain a whole lot about the last few years, particularly why China looms over everything.

The issue was clouded by the latest hysteria, as much as absurdity seeking out the benign can be classified that way, where over this summer much was made out of the nothing of money market reform. Some of it was surely intentional, but whatever the ultimate goal, it became a repeated supposition (substituting for actual insight) that illiquidity was a product of 2a7 alone. Forget for a minute that the predicted disaster in commercial paper never materialized, it still never answered for all that came before, including the "great" events in 2015.

The slightest glance at the TED spread shows that there was "something" that had changed in the middle of 2015; the inflection so prominent that money market reform would have been a welcome distraction from the obvious truth of it for those choosing to think of QE favorably. That it occurred on August 10, 2015 leaves no doubt at all about its nature as well as where it has transited from - China and its "dollar" struggle. It was that date on which the Chinese shocked the globe with "devaluation" that wasn't devaluation, it was the recognition from the People's Bank of China, a central bank with enormous prestige, that they were fighting a losing battle with eurodollar funding.

Click to enlarge

It's not, however, as if TED is alone in that verdict. While the spread between 3-month LIBOR and the 3-month T-bill is one measure of liquidity risk derived from unsecured eurodollar interbank lending, repo fails to provide an entirely separate view of wholesale dollar funding from secured sources domestic as well as chains of rehypothecation stretching into eurodollar reaches. Though a much different perspective, we find still exactly the same thing - an obvious change, for the worse, starting the week of August 12, 2015. With such corroboration, there really is no doubt.

abook-nov-2016-repo-fails-8w

Like the chart showing the history of the TED spread, the one immediately above for repo fails in its own confirmation. Both of them indicate mildly growing risks from before last August, which were then turned into wild gyrations afterward. Again, since CNY was the prominent mode of monetary violence, there is no big intuitive leap here.

More recently, repo fails, as LIBOR and the sometimes backward nature of Chinese-style TED related to T-bills, have indicated big problems. Repo fails spiked in the middle of September, the week of the 14th, which was the week before all the "confusing" drama in offshore RMB (CNH HIBOR). Total fails that week were a reported $685 billion (both "to deliver" plus "to receive"), the highest by far since March. At that time, the PBOC was using whatever of its own wholesale means to peg the CNY exchange rate, suggesting, including RMB money markets, that there was a big increase in cost for Chinese "dollar" interventions in September.

Click to enlarge Click to enlarge

But fails surged again in October, just as CNY was set to count down to zero in its regular "ticking clock" descent. Repo fails actually didn't recede all that much in between, but by mid-October just as CNY was falling off its 3-month cliff, dealer reported fails jumped back up to $527 billion. That meant more than a month of constant irregularity, to put it mildly.

abook-nov-2016-repo-fails

In between those two big jumps in repo fails, the GC rates exploded not just at quarter end but also the days leading up to it and then a few days thereafter. But while at that time repo rates were, like fails, jumping with overall illiquidity, in the past week GC rates have now collapsed. As of today's fix, DTCC reports that the UST GC rate actually fell just below (GC MBS just barely above) the so-called rate floor meant to be comprehensively enforced by the Fed's various tools including its RRP.

abook-nov-2016-repo-gc-ust-dtcc

It makes a complete mockery of the Federal Reserve where it cannot impose either a ceiling or a floor, continuing an unbroken string of ineptitude dating back almost ten years to August 2007. Money markets are supposed to obey a rigid hierarchy, but they clearly haven't in that almost decade, leaving some commentators to speculate that this is just the way it is. The problem with that view should be obvious (but apparently isn't because these same commentators want desperately to believe this is all for the better), meaning that without any way to anchor expectations, let alone operations, it will be other relative changes that have nothing to do with the US central bank that money markets react to most - CNY and the rest of Chinese eurodollars, not 2a7.

The evidence for that is also readily apparent, starting with the primary dealer banks themselves. There is a clear relationship between dealer hoarding of collateral and repo fails; in other words, general illiquidity system-wide starts with whatever it might be that is spooking dealers.

Click to enlargeabook-nov-2016-repo-dealer-positions

Around January 2015, overall financed inventory of USTs began to increase again. That trend was amplified in the weeks just before the CNY break last August. It was all within the same timeframe where TIC data shows a decided increase in foreign central bank "selling USTs," including China. As I have shown repeatedly this year, central bank activities have only gained not abated, and thus dealer reports indicating greater net positive balances (hoarding) throughout this year, but particularly of late (especially when we include bills with the coupons) correspond directly.

abook-nov-2016-repo-dealer-positions-plus-bills Click to enlarge

The net "selling" was very large in both August and September, according to TIC, which would seem to correspond with the amplified "dollar" problems we find here in repo. In the dealer statistics for T-bills alone, we see the same outline as the T-bill rates; dealer net holdings rise as the equivalent yields for bills do. In the latest data through the week of November 9, where the bond/EM selloff accelerated, reported net dealer holdings of bills jumped to $30 billion as yields spiked to above the Fed's money market ceiling.

Click to enlarge

In terms of repo, that would seem to indicate a high degree of shortage of bills for collateral, which meant dealers are right in the middle as either the cause or sharing the same effect(s). The fact that GC rates have collapsed in relatively the same timeframe suggests fragmentation inside and out. In other words, it may be one reason why the PBOC has just let CNY drop in November as there is very limited capacity (in collateral especially) for the global dollar repo market to absorb further interventions. As it only progressed, it may have gone further than that where perhaps cash lenders have recoiled against them by taking "dollars" away from outlets beyond repo leaving it to suddenly pile up in domestic GC.

It was this same kind of onshore/offshore fragmentation that was most obvious in 2007 and 2008, only at that time it was largely between NYC and London, federal funds to LIBOR. And it isn't the first time this behavior has been observed over the past two years; the GC rate dropped sharply in October as CNY moved more quickly lower; in mid-August 2015, as repo fails were still rising after the CNY break, GC (and bill) rates were also moving in the opposite direction.

Again, I believe these are the consequences of money markets deprived of their natural hierarchy. They are being pushed in very obvious fashion by FX, not money market reform, and not Federal Reserve policy of RHINO. Dealers themselves confirm as much by what they are doing, not what they continue to say (which has been at times just plain absurd).

As I write constantly, there are direct costs to foreign central banks for "selling USTs" as a totally ad hoc mechanism by which they hope to cope with the ongoing "rising dollar." The fact that there is more than one euphemism in that previous sentence is a very good indication for the difficulties being faced by those participating in it ("selling USTs" in wholesale terms means other wholesale activities in "dollars" including repo borrowing of dollars via whatever collateral; the "rising dollar" is the most obvious manifestation of a global "dollar" shortage). Writing instead very plainly, global central banks are scraping together whatever "dollars" they can however they can, but when they do so, it has very direct effects for everyone else trying to do the same thing.

Picture a river or large pool of water on the African plain at the end of the rainy season. As the availability of water overall shrinks, it causes increasing conflicts among the many individual animals as well as between species. The growing intensity of conflict and struggle from those animals and those groups of animals has the effect of depriving still others their place at the water's edge where they might otherwise get at least a small drink to survive another day.