[Skip to the end]
Hopefully, when Goldman talks, people listen:
Clarification from author Franesco Cafagna: Views expressed in this piece are his own and are not necessarily reflect the view of Goldman Sachs

1. Do excess reserves really matter and does the FED really need to drain them?

The short answer is: I don’t think so. The total amount of reserves currently in the banking system is the sum of all Required Reserves (including a certain amount that banks hold for precautionary reasons) and Excess Reserves. The FED HAS to provide the banking system with the amount of Required Reserves it needs otherwise rates spike higher (potentially to infinity if the discount window or other forms of “marginal lending facilities” did not exist): the amount required is the result of banks’ individual credit decisions (how many loans they make) and the FED’s job is to estimate that amount and provide it to the system. But the FED does not control this number. When it comes to Excess Reserves, lots of people worry about the potential long-term inflationary impact they may have. The truth is that they don’t matter because they bear no weight in banks’ credit decisions (how many new loans they make). They simply appear on banks’ balance sheets as an Asset that gets “invested” every night in the form of a deposit that they leave at the FED and on which they currently get a 25bps remuneration. If the FED decided to drain excess reserves via Reverse Repo the impact on the system as a whole would be zero because the system as a whole is “self contained”. To understand this let’s think of the most extreme case: the FED drains all excess reserves via one giant Overnight Reverse Repo executed with all the
banks in the banking system. At a macro level all that’s happened is that each bank has changed its Excess Reserve asset (which is effectively an O/N asset) into and O/N Reverse Repo and the two are virtually identical. Another way to think of this is that Excess Reserves are ALREADY being drained every night because banks leave them on their account at the FED every night. The only thing that will change is the liquidity profile of banks IF the FED decided to execute Reverse Repos longer than 1 day: in that case a 1-day assets (excess reserve) would be transformed into a longer asset (Reverse Repo longer than 1 day). Whilst this may affect individual institutions, the system as a whole is unaffected because this amount “extra cash” in the system (excess reserves) is NOT being used for anything. It just sits at the FED every night. So effectively it’s being “drained” already every night. So all this talk about excess reserves and their potential inflationary impact seems misplaced: they are just irrelevant and the FED simply does not need to drain them because they are “self-drained” every night anyway.

2. Does the FED really need to execute Reverse Repos with Non-Primary dealers?

This item has gained press coverage following the Fed’s release of the last Fomc minutes in which it was clear that it debated the possibility of executing large scale reverse repo operations with non-primary dealers: the motivation behind this discussion is the perceived balance-sheet capacity constraint that the 16 Primary dealers might face (a Reverse Repo increases the assets of the broker-dealer entity facing the Fed). This statement by the Fed has created all kind of debate across the street with various dealers coming up with all kinds of estimates of the overall size that the Primary dealers can handle (with some estimates being as low as 100-150bn out of a total of over 800bn that the Fed might want to execute). Leaving aside the actual need to execute Reverse Repo in the first place (point 1 above) and assuming that the Fed will, in fact, choose to execute these operations because it has stated that they are part of the exit strategy policy, I think the alleged Primary Dealers’ balance sheet capacity constraint has been VASTLY exaggerated. It’s true that a Reverse Repo increases the assets of a broker-dealer entity, but this is an issue only for stand-alone broker-dealers (Jeffreys and alike). For Primary Dealers with big commercial banks operations (JPM, Citi, BOA) I don’t believe that this is an issue at all: since they are already sitting on big amounts of Excess Reserves and because 23A (which regulates the activity between a bank entity and its affiliates) does not impose any restriction on the amount of UST, Agencies and Agencies MBS repos that a bank can execute with an affiliate broker-dealer entity, this means that the JPMs of the world could potentially execute reverse repo operations with the Fed up to the amount of excess reserves they are already sitting on without increasing their balance sheet by 1 single cent: it would simply be a transformation of an asset (excess reserves of the bank entity) into another (reverse repo of the broker-dealer entity). So, in my view, the conclusion has to be that the Primary Dealers can in fact absorb a much bigger amount of Reverse Repo than originally thought even by the Fed itself and that realistically the only other counterparties that the Fed might engage directly for these kind of operations are the GSEs: but in this case the reason would not be balance sheet driven but would be driven by the distortion that the GSEs’ participation in the fed funds mkt creates (call me if you would like to discuss this further).

By Franesco Cafagna

[top]