Evidence-based policymaking sounds pretty good. Financial regulators take a look at what works, then respond accordingly. Simple and uncontroversial.
So for those poor unfortunates that waded through the 600 pages or so of CRD IV last Wednesday, the news that the European regulators would be “observing” bank liquidity for two years before bringing in Basel’s Liquidity Coverage Ratio in 2015 didn’t set the world alight.
Observation period looked suspiciously like code for a delayed decision, particularly when compared to the Basel III draft text. The BIS document lays out in some detail which instruments count for the LCR, and in what quantity (60% sovereign or SSA debt, 40% covered bonds and corporate, senior unsecured financials and ABS not allowed at all).
The European Commission’s text is much weaker, although it comes with weasel words about “respecting the Basel agreement”. Adding the observation period and removing the asset specification means that the EC is now free to observe anything it pleases and draw conclusions as it sees fit.
But a bigger issue is why it wants to observe at all. It is relevant which assets count towards the LCR in a crisis, but only then. The stated purpose of the LCR is to ensure a bank can stay out of the wholesale markets for 30 days of stressed conditions, selling “highly liquid” instruments to cover outflows.
It is quite explicitly a crisis measure. It is not useful when there is no crisis. If a bank can roll its CP or raise senior unsecured or pick up wholesale funding from any other source, it won’t need to sell down its LCR assets (although it is supposed to “test” the market regularly). So what is the point of observing liquidity when there is no crisis?
Actually, the EC does say in the FAQs that came with CRD IV that the European Banking Authority will “test different criteria for measuring how liquid securities are under stressed market conditions”. But simulating liquidity stress in opaque OTC markets is going to be heavily dependent on guesswork. Maybe imagining bid/offer spreads widening, or stressing inputs to VaR models will help. But nobody really knows where the bid is until you try to trade big volumes for real in a situation where liquidity is in short supply.
With a bit of luck and a fair wind, the EC will be able to conclude that nearly everything is liquid — risk will be back on, trading will be two-way, and bid/offer spreads will be tight. Of course, we won’t know anything about how durable that liquidity will be in a crisis — the only time when we need to care.
Worse than the sheer pointlessness of the observation period is that the EC has rigged the game in advance. If every bank in Europe is enthusiastically buying and selling LCR-eligible assets to “test” the market and avoid signalling distress (as regulators have told them to do), then there will be large volume two-way markets in LCR eligible assets. The EC will have proved in elaborate detail that 2+2=4.
Even worse is the unequal capital treatment. Under CRD III, securitisations held in the trading book are supposed to have equal treatment with securitisations in the banking book — uniquely among all asset classes. If banks need to hold more capital to trade securitisations, they will widen bid/offer spreads in order to maintain the same level of return on capital. If securitisations turn out to be less liquid as a result, this should surprise nobody.
The LCR is not a terrible idea — far from it. Liquidity ought to be a part of banking regulation. But the best thing the EC could do now is tell the market explicitly what instruments are going in the LCR. Observation is a waste of time, and the market craves certainty.