Ebb Tide
Rebuilding the Treasury General Account will drain liquidity and bring Fed pivot forward
Morgan Stanley suggests (by way of the Financial Times) the depleted the Treasury General Account (TGA), will be rebuilt by US$ 725 billion in the next 3-months and maybe US$1.25 trillion by end-2023. The TGA at latest observation (24th May) held just US$49.5 billion, down from US$410 billion at the end of 2022, and nearly US$ 1 trillion in May 2022. What is good for Peter (the TGA rebuild) is not necessarily good for Paul (the liquidity of the banking sector). A simple estimation suggests a T-bill issuance programme that is LESS aggressive than assumed in the FT article could reduce reserve balances to US$ 2.5 trillion by October, well below the ‘safe limit’ of liquidity for the banking system. This adds to the chances of a change in Fed policy around early fall.
As I’ve outlined several times, the moving parts of Fed liabilities are concentrated in Reserve Balances (held by commercial banks on which they earn 5.15%), Overnight Reverse Repurchase Agreements (held by money-market funds, earning them 5.05%), and Treasury General Account (earning 0%). Together with Currency in Circulation these components account for 93% of all Fed liabilities and are by far the most important.
Fed staff found system reserve demand can be estimated as a log function of total deposits. Ironically, this approach mis-fired during the recent US bank liquidity problems - or rather the signs of stress were suggested by the Fed studies, but they did not appear in actual repo stress, as predicted. The introduction by the Fed of the Standing Overnight Repurchase Facility, which limits the upward movement in repo rates may have sufficiently dampened market signals as to render them useless as ‘early warnings’ of system-wide liquidity limits. If so, this represents another example of policy implementation innovation which has obscured market signals.
That said, the liquidity constraint suggested by the Fed’s modelling accurately warned of liquidity problems approaching in the US banking system at the end of 2022, well in advance of the problems with SVB and other banks. We shouldn’t dismiss its continued relevance.
In a steep tightening cycle, ONRRP was highly attractive for its short duration (overnight) offering full exposure to highest risk-free rates available due to the steep inversion of the yield curve. One way to improve the attractions of the T-bill issuance for money markets is for the attractions of T-bill rates to increase. This would increase costs in funding markets while banks remain fragile. An alternative would be to reduce the attractions of the ONRRP. This could be achieved either by introducing some uncertainty about the future elevated level of the ONRRP, or by limiting access to the facility. In other words, the funds would have to be convinced the Fed was close to cutting rates (carrot) or forced out of the facility (stick). I believe the Fed will switch in late Q3 towards easing, which means by late summer the T-bill market may become incrementally more attractive for funds. It is possible the Fed limit access to the ONRRP, but frankly that seems less likely than a change in policy direction.
Which means, absent a rate cut, the Treasury General Account will be re-built aggressively with the most likely supplier of liquidity being reserve balances of commercial banks. At current levels of deposits an estimated ‘safe’ level of reserves is between US$3 trillion and US$ 2.8 trillion. If the TGA rises to US$ 800 billion by end-2023, I can see a decline in reserve balances to US$ 2.5 trillion, or US$ 2.8 trillion by end-September. Both levels are WAY below the estimated safe level of ‘ample reserves’ for the system for current level of deposits. That’s a recipe for market disruption right around end-August 2023, which if true, may assist the Fed’s decision to pivot on policy.