Back in the summer of 2013, the Treasury Borrowing Advisory Group made a very explicit plea [9] to the Treasury (and the Fed): please help us, liquidity in bond markets is plunging; the plea was so loud the Fed - whose monetization of Treasurys remains the primary culprit for the collapse in fixed income liquidity - promptly announced it would first taper, and then completely halt QE (at least until such time as Congress provided a generous surge in monetizable debt).
It took the peanut gallery a few years to pronounce that "everyone is worried about bond market liquidity", even though this had been clear for years to anyone who actually traded bonds in these broken markets.
A far more concerning development over the same period has been not so much the evaporation of equity, or debt, liquidity but that of the one "market" which central banks have no qualms about intervening on a daily basis: currencies. A recent example of just how thin liquidity had become in FX took place on March 18 [10]during the infamous USD flash crash, when just after the market close, the USD index went, well, "crazy"
[11]
[12]
Furthermore, as we also discussed in the past year, one of the primary reason why FX liquidity has imploded is that after soaking up commodity, bond and equity markets in the past several years, HFT firms such as Virtu decided to make FX their latest trading arena [13].
[14]
And while the practical impact of collapsing liquidity in this last remaining asset class was also quite clear to market participants, especially those who relish the daily USDJPY stop hunts, there had been little theoretical work done on analyzing just how bad said liquidity vacuum had become.
Until today, when courtesy of DB's FX strategist Oliver Harvey, we have the answer: using a "framework for measuring liquidity in FX based on the ratio between average weekly volumes and realized vol" Harvey the obvious: FX liquidity is going, going, gone. To wit:
There has been further deterioration of this measure of liquidity across 8 out of 12 commonly traded G10 pairs. EUR/crosses in particular have suffered, with EUR/USD and EUR/JPY seeing the largest declines, further evidence of how painful euro funding unwinds have been for investors.
The bottom line: "liquidity in EUR/USD is closing in on the lowest levels save after the Lehman bankruptcy, with only late-2010 worse."
[15]
Other crosses reveals the same trend: "Liquidity has also fallen in USD/JPY but only towards its longrun average, while liquidity in GBP/USD has moderately improved but from low levels. While quoted bid-ask spreads are a different, and probably inferior, measure, interestingly they tell out the same story (see chart 5), with 3 year rolling percentiles very elevated in EUR/USD but low in USD/JPY. Turning to volumes, these continue a steady decline from late last year across G10 FX, a spike at this month’s Fed meeting notwithstanding (chart 6)."
So just in case anyone needed a reason to feel bad for big bank FX traders, especially after all the currency rigging revalations, here it is:
The above underscores the tricky current environment for risk allocation. The same message is provided by our Regime Machine indicator released today which shows our measures of trend (VHF), uniformity (cross correlation between currency pairs) and smoothness (inverted realized vol) to have collapsed to historical percentile lows, the worst possible conditions for momentum strategies.
And here, for the visual learners, is DB's complete chart pack showing that there really is no way out: the moments a massive inversion in sentiment arrives, the FX market will simply go bid-or-offerless as all the HFT-driven fake liquidity in the market disappears with the switch of a button.
[16]