By Jeffrey P. Snider
I'm not sure what popular perception believes of international holdings of "reserves", but I would wager there is a rather large disconnect between it and how the international system actually works. This is more than just theoretical notions of banking in eurodollar shadows, but the pipelines that connect and cross the globe. Central banks do not have massive stores of currency in a vault somewhere, either local or London (or New York). There isn't enough currency in existence to satisfy "official" claims of USD reserves.
That means that even foreign central banks have to have bank accounts with either London or New York banks (or with FRBNY directly). If Thai banks are short of dollars, for example, the Bank of Thailand can sell its "reserve" assets on the open market (or, more likely, bespoke and dark), but to do so requires a bank custody account. Raising dollar liquidity, then, means exchanging dollar-denominated financial assets for a deposit ledger at a dollar providing bank (again, mostly London but also NYC).
Any dollar liquidity crisis would be very evident in the custody data since there are no anonymous stores of generic currency to be deployed. A dollar crisis shows up in custody accounts exactly like I have described in the exchange above - less holdings of dollar assets. From there, we can infer intent, particularly since custody data is conveniently characterized as either private or official.
Private dollar exchanges would be more on the order of "hot money" or investment-driven flows. Official custody is central banking or central government. The fingerprints of "official" exchanges between assets and deposit account liquidity carry much less ambiguity, far less prone to multiple interpretations - central banks raise "dollars" because their local banks have little other recourse. That's not a trivial distinction, then, because these dollar mechanisms intricately connect global asset markets and trade itself.
There was a large alteration in the pattern of dollar "flow" in the middle of 2013, right exactly when the eurodollar market went haywire and U.S. credit sold off so dramatically. That was followed by an evident reprieve, particularly after the September FOMC decision against taper. It did not last, as we know, again, from more real-time data gathered elsewhere.
Where the custody accounts help in piecing together recent history is in those inferences of need/investment decision-making. Central bank activity is a vital clue in that regard since central banks are not directly responding to "investment" considerations.
On a scale that we have not seen, and a frequency not seen since the 1990's, central bank holdings of U.S. treasuries have been declining. That was a dramatic change over the pattern that had been established since the end of the 2008-09 panic. We can see it more clearly in a cumulative fashion:
The scale and duration is even greater than what was witnessed in the 2007-09 period. That does not necessarily lend itself to linear interpretations of the scale of current dollar problems vs. those of the panic period. The panic period was far more European in nature, and thus there were other avenues available for dollar financing (swap lines, for example) where central banks (ECB & SNB) did not have to directly appeal to their "reserves." But the eurodollar problem then did spillover into these other, more indirect nations and central bank systems as shown above. With a more "emerging" focus this time, it is actually not surprising to see a greater change in TIC custody in 2013 than 2007-09.
What is perhaps less expected is the degree to which dollar liquidity appears affected. Even private flows of dollars have "tapered" to the point at which dollar asset gathering in foreign custody is diminished. The product of both private diminishment and central bank mobilization is as the chart above shows. We can infer that dollar travails are more widespread than beyond the central bank axis.
To be clear here, central bank activity cited above includes not just emerging markets but also the Asian giants (though I suppose China may still count as "emerging", though far more dynamic than "developed" Japan). In fact, Japan and China have seen net selling in recent months, including the dramatic changes in June 2013.
While some focus has remained on Chinese internal financing markets, the fact is that China is also reliant on dollar availability itself. The activities in Chinese interbank markets fit this dollar pattern, with dramatic spikes in SHIBOR coinciding with custody changes (US dollar liquidity). And there is more than a fair bit of dollar activity predicated with yen markets.
The oil exporters also deserve greater attention here, but I think this bodes ill moreso for Brazil. Clearly there is a usage of dollar "reserves" here by the Banco do Brasil, contradicting the narrative it created through its swaps program. Mobilization of reserves is usually the last step before severe devaluation (which is why the central bank was so careful about structuring its programs to avoid that appearance).
Crisis periods themselves are far from linear - they are more like ripples. There is an ebb and flow, with serious, often hidden, dynamics that essentially tear down the foundation of leverage. You can see why that would be so damaging in the eurodollar case, where leverage was built on the premise that volatility in funding markets would remain low for(ever?) an extended period. That allowed a free flow of dollars into "markets" that probably should have experienced much more caution and prior trepidation.
Taper away that premise and the dollar edifice crumbles all over the world.