QE is intended to put downwards pressure on longer dated yields, but its most obvious impact is on the front end. Repo and short dated bills are pinned around the leaky ON RRP floor, with the zombification spreading up along the bill curve. This outcome stems from of our two-tiered monetary system interacting with the constraints of Basel III. In general, every dollar of QE creates two dollars of money – one dollar of reserves (money for banks) and one dollar of bank deposits (money for non-banks). Banks and non-banks can be thought of as two distinct classes of investors, each with their own constraints and opportunity costs. Banks and non-banks take the new money and rebalance their portfolios. Their eligible investment universe most strongly overlaps at the front end, leading to a flood of investments into money markets. Massive QE eventually pushes all front end rates to the ON RRP floor (or below). In this post we review QE money creation, the investment constraints of the bank/non-bank investor classes, and how the QE experience abroad is a preview of what is to come in dollar money markets.
Reserves and Bank Deposits are Both Money
In our two-tiered monetary system, reserves are created when the Fed buys assets or makes a loan, while bank deposits are created when a bank buys an asset or creates a loan. Banks transact with each other using reserves and non-banks transact with each other using bank deposits. The two tiers are connected through banks, who own reserves and owe bank deposits. Non-banks spend their bank deposit money, and banks settle the payment amongst themselves using reserves.
In a superabundant reserves world, both tiers of money become independent. That is easy to see from a non-bank’s perspective – no one buys an item worrying about how their bank will settle the payment with the seller’s bank. It is different from a bank’s perspective. Historically, a bank would carefully manage its reserve levels with its deposit outflows in mind. But payments management has become less important as reserve levels have risen significantly (the pre-Crisis banking system worked with around $20b in reserves, today we have around $4 trillion). Reserves management today is more about meeting regulatory liquidity requirements and relative value returns. Although banks have much more reserves, they are constrained by Basel III in how they deploy them.
Bank Are Now Government MMFs
Basel III’s LCR requires banks to hold liquidity portfolios comprised of High Quality Liquid Assets (“HQLA”). The size of these portfolios are substantial, with JPM alone required to hold over $600b. HQLA is defined to include reserves, Treasuries, Treasury repo, and Agencies (with a slight regulatory penalty). In practice, banks fulfill their HQLA requirements through a mix of reserves, Treasuries and lending in Treasury repo. The mix fluctuates according to relative value, where banks opportunistically shift out of reserves and into Treasury repo/Treasuries when their returns exceed interest on reserves (“IOR”). When repo rates traded consistently above IOR in 2018, banks rebalanced their HQLA portfolios towards more Treasury repo (see here). The LCR essentially requires a bank to manage their reserves holdings like an internal government money market fund, with IOR as their hurdle rate. As QE adds reserves into the banking system, the collective AUM of those ” internal MMFs” has risen to the trillions. These HQLA portfolios act as a soft ceiling for Treasury repo and bills at a rate slightly above IOR.
QE Bank Deposits Overflow Into MMFs
Non-bank investors are free to rebalance into any assets, which in part explains narrow credit spreads and high equity prices. But on the margins the new QE created deposits are flowing into MMFs (and other liquidity vehicles). This is because Basel III sets limits on the size of a bank’s balance sheet (leverage ratio), and imposes a regulatory cost for each type of liability (LCR/NSFR/GSIB surcharge etc., see here). In prior QE periods, Basel III had not yet fully taken effect and the pace of QE was a fraction of today’s $120b/month. Banks have too many deposits and are optimizing their balance sheet by reducing high cost deposits. Pushed out deposits have no where to go but MMFs (despite their offering 0% or negative net yields).
Like bank HQLA portfolios, MMFs are heavily constrained in their investments and largely hold bills, agencies, and Treasury repo. They are essentially competing with bank HQLA portfolios for the same investments, but with the lower ON RRP offering rate as their hurdle rate. They are deploying their trillions on any investible assets that yield even a fraction above their hurdle rate. MMFs thus create a soft ceiling just above the ON RRP rate, while cash investor without ON RRP access fight over investments yielding below the ON RRP rate.
Money Market Zombification is Global
Massive QE and Basel III have led to predictable outcomes across the world. Switzerland, Japan, and the Eurozone all have money market curves pinned slightly below their lowest administered rates. While their banking systems are not identical, they share the same theme: on-going QE creates money, regulatory constraints impose costs on banks for holding that money, and uneven access to administered rates. They were first to zombify their money markets, but the Fed is catching up via its $120b/month purchases.
The dollar system is more market based (as opposed to bank based), so it has two administered rate floors – the standard bank centered IOR and also a market centered ON RRP. At the same time, it is a global currency with vast amounts held beyond the reach of any administered rate. The IOR floor was easily breached, and the market is banging on the ON RRP floor with short dated bills and overnight repo touching negative. There are trillions in reserves and bank deposits weighing upon the floor. Negative overnight rates would force investors to move a few months out to stay positive, and then a few more months out..
Note: Just as the effects of QE are most strongly felt at the front end, so its reversal via QT was also most strongly felt at the front end. Of all the things that could break, it was the super deep Treasury repo market that broke in September 2019. A steady increase in demand for repo financing (as shown in growing SOFR volumes), was met with steady declines in “bank internal MMF” assets until something snapped.