A long time ago, in a regulatory regime not so far away, bank deposits were simple things.
Under new Basel III rules, things are a touch more complicated.
Banks now have two types of deposits to contend with: operating and non-operating. The former are considered relatively "sticky" forms of funding in that they are unlikely to be suddenly pulled from the bank. The latter are classified as a form of short-term wholesale funding that has the potential to rapidly evaporate. Under Basel III, banks need to hold a buffer of ostensibly high-quality liquid assets (HQLA), such as government bonds, to cover non-operating deposits and protect themselves from a funding run.
There's a trade-off here, and it is one that tends to make banks reluctant to hold bigger HQLA reserves.
Maintaining a big war chest of safe, and therefore low-yielding, assets to cover potential deposit outflows means lower returns for the banks. As JPMorgan Chase put it recently in a client bulletin for corporate treasurers: "HQLA provide security, liquidity and reliability in a stress event, but those benefits come at a price, notably limited and potentially negative returns. When compared to the return potential for operating deposits used to fund loans, the return on non-operating deposits deployed against HQLA is significantly less, limiting return potential for both the bank and the company."
Here's JPMorgan, for instance, warning that some clients might need to hit the road through use of a not-so-subtle infographic . (The bank isn't alone in cautioning customers that they may have to move their money).
Why do we bring this up?
With an interest rate increase looming, speculation about its impact on bank deposits is intensifying.
The thinking here is that a rate hike would put an end to six years of extra cash pouring into banks as the Federal Reserve flooded the financial system with liquidity. Instead, as the central bank raises rates, the difference between rates paid by bank accounts and money market funds is expected to rise in tandem. That could encourage a bunch of depositors to move money from banks to newly higher-interest accounts.
Potentially exacerbating the dynamic is the new overnight reverse repo (RRP) facility created by the Fed in an effort to help it better control short-term rates as it normalizes its monetary policy and that now allows non-banks such as money market funds to have reserve accounts at the U.S. central bank.
In a recent note, Credit Suisse analysts Zoltan Pozsar and James Sweeney highlight the potential impact on bank deposits and the potential for turbulence as the Fed approaches its "historic liftoff from the zero lower bound." At issue are the non-operating accounts held by big institutional investors, which the analysts estimate equate to about 60 percent of the $1.1 trillion in deposits held by U.S. financial institutions.
For banks with lots of HQLA reserves and high-quality assets, the prospect of a bunch of hot money departing for the greener pastures of money market funds may not prove very worrisome. In fact, as deposits depart, such banks may find their balance sheets benefiting from a surplus of capital with which they can do all sorts of amusing things, such as buy back their stock and debt or (gasp!) raise deposit rates.
But for banks that haven't loaded up on stickier funding or higher-quality assets to offset their non-operating deposits, well, things have the potential to become more interesting. As Pozsar and Sweeney put it: "A rising tide – rising interest rates – may not lift all boats as is typically the case during hiking cycles."
Here's what the analysts say:
Fast flows — the deposits to government-only money fund flows — may lead to unforeseen dislocations ... This is because the large U.S. banks that hold most of the hot money (in the form of non-operating institutional deposits) that could leave fast once interest rates move higher are positioned very differently for potential deposit outflows. As it turns out, the largest U.S. banks hold similar amounts of non-operating deposits, but the HQLA portfolios backing them vary significantly from bank to bank (see Exhibit 15).
Someone won’t get things right.
Estimating the exact amount and interest rate sensitivity on non-operating institutional deposits is not an exact science. The concept of non-operating deposits itself is a new concept (a creation of Basel III) and banks are presently building their systems and models to estimate their exact volume and behavior. The potential for model risks abounds.
Model risk. Fun times!
Banks whose non-operating deposits eclipse their reserves do have a few levers they can pull to try to stem outflows, including paying depositors higher rates or lending out securities from their HQLA portfolios. But the point here should be clear. Two grand experiments, one conducted in the technical backwaters of monetary policy and the other operating in the realm of new banking rules, are about to collide.
It bears watching.