For the last several years, the behavior of Japan’s yen against the US dollar is near perfectly correlated with the eight-week average of US$ repo fails. From the outside, which means the mainstream, this sounds like one of those nonsensical relationships built up in fevered imagination by superstitious blasphemers. If I go outside three days in a row and it rains each time, this doesn’t mean I control the rain just by leaving my house.
But this JPY-repo connection isn’t some fly-by-night, temporary oddity. It has lasted well more than two years and the association is only getting stronger. With particular emphasis following the last weeks in February, first weeks in March, the yen is now at a multi-decade low while the average of total repo fails at a multi-year high equal only to March 2020 (going back half a decade).
This is Big Problem territory and then some, for both sides of the Pacific.
I wrote about what’s going on here several months ago, so no need to rehash the mechanics. Short version if you haven’t seen it: what’s wrongly called the “yen carry trade” is really Japanese banks redistributing eurodollars on a collateralized basis. Given the inherent maturity mismatch when doing this, Tokyo’s big firms end up short dollars and short collateral.
Thus, any dollar shortage which strongly implicates collateral problems (isn’t this all of them?) comes out in something like repo fails, tied now closely together with JPY as Japanese banks struggle mightily under the weight of being eurodollar-ed.
Having covered the “what”, the question today is why doesn’t anyone know this? JPY’s collapse has generated so much interest that on Wednesday it was actually trending on Twitter.
It is easy to dismiss this correlation even while recognizing just how strong it appears to be through nothing more than visual inspection (running the numbers only confirms the suspicion). Without knowing the background, it truly might appear little more than random voodoo; our modern science-y brains are conditioned to disbelieving unless the right people say otherwise.
In terms of repo and collateral, this reluctance to look deeper was exposed by the Treasury market on October 15, 2014. To this day, the government claims it was nothing more than computerized trades run amok. In its multi-agency report issued in July 2015, while admitting “no single cause is apparent in the data” it goes on to speculate about several contributing culprits anyway:
“In sum, record trade volumes, a decline in order book depth, changes in order flow and liquidity provision, and notable and unusual market activity together provide important insight into the factors that may have contributed to the heightened volatility, decreased liquidity, and round-trip in prices on October 15.”
Those aspects tell us next to nothing about why everything happened, merely what data shows when it did. As you might surmise, the word “collateral” does not appear once, not a single instance in its seventy-six fluffed up pages.
It is a curious omission, because it isn’t as if those who wrote the report aren’t aware of this monetary purpose. Speaking in advance of its release, in the middle of April 2015, Simon Potter, then the Executive VP of FRBNY’s Markets Group stood in front of a meeting of primary dealers to discuss October 15.
Maybe it would otherwise have seemed superfluous to state to this particular audience what it already knew, but what Mr. Potter said about the Treasury market wasn’t just matter-of-fact:
“The Treasury market serves as a liquid investment option for global investors, a ready source of collateral for other financial transactions, a benchmark for pricing other securities, an important component of the Federal Reserve’s framework for implementing monetary policy…” [emphasis added]
Yet, Potter wasn’t connecting the second of those with the third - referring instead to Treasury yields not the instruments’ use as collateral in the framework of Fed money-less policies. As so, this was the only occasion in his lengthy remarks Potter used the word collateral.
Though it was never mentioned anywhere in any official source, repo fails had indeed surged in the months leading up to October 2014. Throughout June and July, weekly and average fails more than doubled – right as the US$’s exchange value began its “unexpected”, seemingly superstitious surge at the time blamed on looming Fed rate hikes (which wouldn’t actually happen in any serious way for more than two years).
Fails would remain elevated for months, making an especially conspicuous rise at the end of Q3 2014 and then again two weeks later, which just so happened to be the week of October 15.
In an equally curious twist, we know what the level for repo fails had been because of data provided by Simon Potter; or, specifically, the primary dealers who supply the weekly figures to Mr. Potter’s group at the FRBNY via a call report called FR2004.
This next part is really going for superstition. The history of FR2004 traces back to July 1990. For reasons that I cannot pin down (and I have researched this for weeks), all of a sudden in that particularly hot summer, while the world was fixated on Saddam’s fanciful cash grab about to strike Kuwait, FRBNY decided it needed dealers to report more complete information on their repo market activities, especially fails.
Distracted by Iraq’s military buildup along the Kuwaiti border, over in the Treasury market something else equally absurd demanded notice: the Salomon Brothers scandal.
If you aren’t familiar with that episode, I wrote about it many years ago, as had Michael Lewis briefly in his book Liar’s Poker. It was also the subject of yet another exhaustive multi-agency government investigation which found, shockingly, Salomon Brothers’ antics were just the tip of the iceberg, hardly isolated. Dealers up and down the line were committing outright fraud in order to secure newly auctioned on-the-run (OTR) instruments in Treasuries as well as GSE debt.
From that report:
“On January 16, 1992, administrative proceedings were instituted jointly by the SEC, the OCC and the Federal Reserve against 98 registered broker-dealers, registered government securities brokers and/or dealers and banks (the "respondents"). In those proceedings, the three agencies found that, in connection with their participation in the primary distributions, each of the respondents made and kept certain records that did not accurately reflect the respondent's customers' orders for the GSEs' securities and/ or offers, purchases or sales by the respondent of the GSEs' securities.”
Ninety-eight dealers had “kept certain records that did not accurately reflect” what they were doing, how desperate (greedy, in a sense) they’d become for these OTR securities.
Why?
For all the dozens maybe hundreds of contributing staffers, all the government could figure was a kind of repo-engineered short squeeze. “Squeezes in the repo market also can be created or exacerbated by market participants that hold a relatively large portion of a security. For example, a participant that holds a large amount of a scarce security can increase its scarcity value by financing a portion of the holdings away from the ‘special’ repo market.”
Ninety-eight dealers were risking their entire operations over some backdoor arbitrage of GC repo spreads? If only computer trading had been a bigger “thing” back then, the government could’ve gotten a serious head start on July 2015.
Though I haven’t uncovered proof, it is one of those coincidences that begs legit consideration; the “modern” FR2004 asking dealers more robust information about repo and collateral came about just as the Salomon scandal was reaching its peak absurdity (my story on it was published in these pages [The Crony Pretense Behind Warren Buffett's Banking Buys] just three days before the government’s report on October 15 came out; how’s that for another coincidence!)
You really have to wonder if someone deep down on the staff level didn’t get the inkling there was much more going on here than short squeezes and accounting fiction, enough so to get some of the empty suits upstairs to sign on to purposefully gather more information while the episode was still playing out.
In any event, the dealer form was updated again in 2004, and once more in January 2022. Unlike 2004’s reform, this year’s had a specific purpose: collateral, damn it!
The new FR2004C – where repo fails come from – commands primary dealers (therefore doesn’t cover the whole repo or collateralized marketplace, which includes derivatives) to spill much more granular information about their shadow repo activities, specifically:
“While other data collections cover the tri-party, GCF, and cleared bilateral market, the revised Form FR2004 provides us, for the first time, with a comprehensive view into the uncleared bilateral repo market.”
Initial results were released only last month, and in one sense there wasn’t much surprising in them. To me, it was a total letdown. Volumes are huge, we knew that, and the primary collateral backing both shadow repo and reverse repo is…Treasuries.
Yes, but that’s just one side of the equation. These updated collections collect nothing about collateral transformation, the very thing which nearly ended AIG but did end the pre-crisis era for money, markets, and economy.
What is collateral transformation, or a collateral swap? Here’s former Federal Reserve Governor Jeremy Stein describing it in February 2013, a year and a half before those darned computers got loose:
“Imagine an insurance company that wants to engage in a derivatives transaction. To do so, it is required to post collateral with a clearinghouse, and, because the clearinghouse has high standards, the collateral must be ‘pristine' - that is, it has to be in the form of Treasury securities. However, the insurance company doesn't have any unencumbered Treasury securities available-all it has in unencumbered form are some junk bonds. Here is where the collateral swap comes in. The insurance company might approach a broker-dealer and engage in what is effectively a two-way repo transaction, whereby it gives the dealer its junk bonds as collateral, borrows the Treasury securities, and agrees to unwind the transaction at some point in the future. Now the insurance company can go ahead and pledge the borrowed Treasury securities as collateral for its derivatives trade.”
Insofar as anyone knows, or what’s reported on FR2004C, it’s almost exclusively Treasury collateral when in fact, and in money reality, there’s a whole bunch of junk of questionable liquidity characteristics lurking in these shadows. The very sort of junk which was already hitting a serious rough patch in the weeks leading up to the Day Of Computers, as I recalled at the time:
“In the first two weeks of October [2014], however, dealer holdings sharply increased as the junk problem got worse, meaning that even dealers had reached their limit of collateral supply and were now looking to replenish for their own books at the same time the rest of the repo market was left then to fend for itself. The 12 minutes of illiquid buying on October 15, again a ‘buying panic’, clearly had nothing to do with computer trading.”
In short, it was a margin or really collateral call which provoked a reverse of too many collateral-for-collateral swaps, leaving those suddenly exposed to panic-bid whatever good and useful collateral they could get at whatever price it took to get it.
We still get these calls, or scrambles for collateral, as anyone in Tokyo nowadays might honestly attest to.
So, why does all this remain in the dark? Officials can’t make sense of it in any detailed fashion. They come upon small bits and pieces, conditioning themselves to think in anomalies or one-offs rather than fundamental systemic operation, even the occasional top-level policymaker like Stein (who didn’t stick around long at the FOMC), but no one can ever put it all together in a cohesive framework then give it the “necessary” imprimatur for public acceptance.
Authorities can tell there must be something here, yet because it is entirely opposite from bank-reserves-are-everything it doesn’t get beyond raw data collection. And even when advancing into information gathering, as the new FR2004C, it never goes nearly far enough.
No matter how robust or clearly visible correlations and evidence shows otherwise, the Earth (Fed) must remain firmly planted at the center of the solar (dollar) system. Though it might be troubling to the Church of Bank Reserves, quite a few operating in it cannot help but recognize how everything instead points to a more straightforward arrangement of celestial (collateral) mechanics in this fuzzy outside realm of monetary space.
Until sanctioned by the highest authorities, those seated upon their gilded thrones set within the wondered majesty of FOMC Conference Room, the faithful must remain forever faithful to the Ancient Way. Even then, FR2004C’s next update just might go behind the veil of bilateral reverse repo to the heart of collateral swapping because truth has a way of coming out no matter the institutional resistance to it; but that’s a drawn-out process, as it obviously has been.
It’s not just those in Japan who can’t afford to wait for however many more years this will surely take.
