By Jeffrey P. Snider
I have to say that I am more than a little amused by the "news" that the US Treasury Dept. is asking for comment from the primary dealers about the recent repo fails. As Bloomberg put it succinctly, "The Treasury also asked the dealers to explain the causes for an increase recently in "fails-to-deliver" in the market for U.S. government debt." I have little doubt that they already know what is taking place, so it seems rather like a fishing expedition for the banks to supply a ready and somewhat plausible (given the esoteric nature of repo, it doesn't even need to be logically consistent, just sound like it) excuse that doesn't involve QE.
The current probing is running toward treasury shorts as a primary culprit. Shorting certainly plays a significant part of demand for individual securities, creating a lot of specials particularly around auctions, but I fail to see how that would create systemic, regular shortages. Just a year ago, we saw exactly the same thing as we see now, including the impulse to pin blame on short positions.
Earlier in June, as repo rates, particularly in the 10-year, were heading down toward the 3% fail penalty, Bloomberg noted that shorts were present "at depth."
"It is typical for debt like this to get very special before an auction," said Kenneth Silliman, head of U.S. short-term rates trading at Toronto-Dominion Bank's TD Securities unit in New York. "There certainly seems to be a deep short base in the issue."
That might be true to some extent, but it again fails to take into account the regularity of this liquidity bottleneck. Are we to see a ban on shorting UST? In the very same article, this idea was contradicted by positions in futures markets, "Hedge funds and other large speculators have net short positions in 10-year Treasury note futures of 43,295 as of June 3, according to Commodity Futures Trading Commission data. That's down from a peak net short this year of 162,278 in April." In other words, all of this is exactly the conjecture I was referring to above that isn't backed up by much hard data.
There are all sorts of technical confluences at work, including the reopening of UST auctions that distort on-the-run availability. None of that explains why this problem has gained such depth in the same spaces at the same time. It's almost like the weather excuse for the first quarter - if the economy had actually been running robustly, winter storms would not have made much difference (assuming they did at all). If the repo market was as resilient as Janet Yellen perpetually asserts, technical factors would be a minor issue without major interruption. In short, it's not shorts but something far more systemically deficient, to which QE has played an inordinately enormous part.
The need to invent some mystery is, as I said above, a matter of politics more than anything. The full manner of what is happening in repo starts and ends with total collateral availability, tracing back to collateral formation. "Total" collateral involves not just the US treasury market, but MBS, agency debt and even small amounts of corporate securities (at high haircuts that make this class too inefficient to make up a large part of the mix).
MBS issuance, both in gross and net, has been decimated by last year's selloff - and judging by bank reports so far this quarter it is not coming back anytime soon. Repo dealers and participants that once relied on MBS for funding have essentially been "herded" to UST as the only available alternative (agency debt has all but ceased as a significant contributor).
But here too is another problem, as the US Treasury Department has been issuing far fewer bills as a cash management tool. According to SIFMA, gross issuance of UST surged from a paltry $95 billion in May to $271 billion in June - the very same time as the fails problem. However, SIFMA's data is "gross", not taking into account redemptions that reduce supply.
The Daily Treasury Statement for June 30 offers an answer here - the government issued $337 billion in bills plus another $251 billion in notes and $20 billion in bonds. Against that, it redeemed $422 billion in bills and $176 billion in notes - a net increase in securities of just $10 billion. In other words, repo collateral seekers were further "herded" into notes and away from bills, assuming they could get any at all (which fails indicate was more than a little difficult).
The more demand crowds into a single space, the more participants are going to amplify each other's actions, including those shorting UST - whatever specials are made so by short positions are thus made even more special, which has a snowball effect all down the line of rehypothecation. With banks at quarter end looking to "dress up" their risk and leverage, rehypothecation itself strains even further.
But all of these are actually just details in the bigger picture; the real issue. Collateral creation is as much economic in nature as anything else. Yet credit markets have been so distorted by years (decades) of alterations in the name of monetary economics that collateral creation itself has been stunted, weakened and, in some places, disabled. The repo market used to run on mostly MBS, and behaved "orderly" in doing so until that realization dawned; financialization was creating what amounted to "false" collateral. Thus the pool of available securities necessarily shrunk, including violent panic in the mix, and has never regrown or replaced what was "lost."
The only way to do so would be a robust economic expansion, as that would be a collateral machine. The FOMC fully expected that would have happened by now, allowing any downside in QE distortions via repo to be cleaned up under far more favorable conditions. Sound lending to intermediated obligors that can actually meet financial requirements would allow collateral to be processed and efficiently utilized (at reasonable haircuts that maintain a quasi-limit on leverage). What do we have now to pick up that slack, other than subprime auto loans and junk corporates? It is financialism itself that is holding back repo markets, not the least of which is the lack of true economic expansion in the fullest sense (not just the mathematical imputations of the Establishment Survey that don't even make much sense in wider context).
If the primary dealers were to be truthful to the Treasury Dept. on this repo account, they would note the failure of the economy, due fully to their own part in distorting the very functions of finance. That won't happen, though, so I eagerly await what I hope, in consolation, are creative if ultimately nonsensical excuses.