personal views of a former fed trader

ON RRP Take-Up Will Go Much Higher

The humble ON RRP is in the spotlight as take-up marches steadily upwards. It will go much, much higher. Increasing participation is largely a function of two structural forces in the financial system: on-going Fed QE ($120b/month) and Basel III constraints. On the margins, changes in the level of the TGA have some impact as well. Water pouring into a glass remains in the glass, until the glass is full and then every incremental drop overflows. The Fed has been adding tremendous amounts of liquidity into the financial system over the past year, and there was initially very little take-up in the ON RRP. But now it appears the banking system is full – Basel III constraints are becoming binding. The incremental QE deposits are flowing out of banks and into MMFs, and then down the ON RRP drain. The system is working as intended. In this post we review the cause of increasing ON RRP take-up, note that high take-up will be a permanent feature going forward, and suggest that money market rates will fall below the ON RRP offering rate.

Why is ON RRP Take-Up Increasing?

When the Fed conducts QE, it increases the level of reserve assets and bank deposit liabilities in the banking system (see this post). This mechanically increases the size of the banking system’s aggregate balance sheet. Some of the money can also be absorbed by the Treasury General Account (“TGA”) which is the Federal Government’s checking account at the Fed (it is outside the banking system). The $120b per month of liquidity pouring into the financial system has largely ended up in the banking system and the TGA (see graph).

QE mechanically increases the size of the banking sector

Basel III regulations limit the size and composition of an individual bank’s balance sheet. Like an office building with fixed square footage, banks have a limit on how many deposits they can hold and must decide whose deposits to accept. Under Basel III, GSIBs are incentivized to push out the deposits of institutional investors because they incur higher regulatory costs (see this post). These institutional investors take their deposits and put them into money market funds (“MMFs”), whose assets have grown to around $4.6 trillion. Note that some of these investors may also prefer MMFs for their lower credit risk. MMFs take that money and invest them largely in Treasury bills, repo (overnight loans secured by Treasuries) or the ON RRP.

Pushed out deposits end up in MMFs, who place them in ON RRP

The ON RRP is like a checking account offered by the Fed to MMFs. MMFs that cannot find any investments yielding above the ON RRP offering rate can park their money in the ON RRP. This mechanism is how the Fed controls overnight rates – if the ON RRP is offering 0% then private sector borrowers must offer MMFs an equal or higher rate to attract MMF investments. Today MMF assets are growing, but they cannot find enough investments yielding above the ON RRP rate so they just park the money in the ON RRP. As a result, ON RRP take-up increases.

In theory, the banking system as a whole has more than enough capacity to hold the extra QE deposits, but in practice that capacity is not evenly distributed. The U.S. banking system is highly consolidated, with the four largest GSIBs holding 40% of the banking system’s assets and the remaining 4000 banks holding the rest. GSIBs fall under the most stringent version of Basel III so are most incentivized to push out deposits. GSIBs are not substitutable with smaller banks as they have global platforms and a range of services that smaller banks cannot offer. The banking sector is thus bifurcated where GSIB constraints have a disproportionate impact on the banking system’s capacity to absorb deposits.

It’s also worth noting that the U.S. banking system (relative to GDP) is smaller than many foreign jurisdictions who have an even larger (relative to GDP) QE programs. The U.S. is a more capital markets centric, rather than bank centric financial system. It thus has less space than other banking systems to hold QE deposits. At the same time, the U.S. decided to “gold plate” its version of Basel III such that U.S. GSIBs are subject to higher standards than their global peers.

U.S. is much less bank centric than other countries

On the margins, changes in the TGA account have also influenced ON RRP participation. When an investor buys a Treasury security, his money moves from his account in the banking system into the TGA, which is outside the banking system. This frees up space in the banking system to hold more deposits. The TGA declined by $200 billion over the past month and is projected to decline further heading into a potential debt ceiling episode in July. This suggests that more deposits will be flowing into the banking sector and ballooning balance sheets.

Higher ON RRP Participation is Permanent

Higher ON RRP participation stems from a quantity problem, and not a price problem. It would not change if the Fed tweaked its ON RRP rate, and only modestly change if there was an increase in the issuance of money market investments. An ON RRP tweak directly influences money market rates, but would not create more balance sheet space. An increase in the availability of money market investments would temporarily temper ON RRP participation, but it also does not solve the fundamental problem.

Suppose Treasury issued $1t in bills to fund new spending, then MMFs would move from the ON RRP and into newly issued bills. ON RRP take-up would decline. That money would flow into the TGA account, then be spent into the banking system. Some of it could end up in banks with balance sheet space, but some of it will also end up in constrained backs where it will again be pushed out to a MMF, who will then turn around and put it back into the ON RRP.

Bill issuance does not solve balance sheet constraints

Of course, $120b/month of liquidity continues to flow into the banking system while all this is happening.

There is no real limit to how high the ON RRP can go. The ON RRP will always have enough collateral as the Fed is increasing its collateral holdings at the same time it is pumping deposits into the banking system. Nor are ON RRP counterparty limits practically binding, as large fund complexes can simply shuffle investments internally such that child funds approaching the $80b ON RRP counterparty limit receive more Bills/private repo while other child funds increase ON RRP participation. The Fed could also just increase counterparty limits at any time. The current pace of QE suggests that take-up could reach $1t by year-end.

The System Is Working As (Poorly) Intended

The purpose of the ON RRP is to control the Fed funds rate, which is the Fed’s stated policy rate. The funds rate can never print below the ON RRP, as all lenders in that market have access to the ON RRP. But the funds rate has long ceased to have any relevance to the real or financial economy. Money market rates that actually matter have and will continue to move below the ON RRP offering rate. Absent a technical adjustment, that means slightly negative rates.

Eventually, MMFs will also begin pushing out investors because their portfolio of 0% investments do no cover their operating fees. Those poor investors will end up with no where to go but in assets yielding below the ON RRP rate. This is without factoring in the oceans of dollar liquidity off-shore who do not have access to a on-shore MMF. The ON RRP is not and will never be a floor for dollar rates (see this post).

Investors seeking to escape negative yields can find refuge in the FX basis (see this post) or on the balance sheet of foreign GSIBs. Foreign GSIBs calculate their Basel III leverage ratio in a more lenient way so they have more room to hold deposits. In the 2014 QE regime foreign banks actually absorbed more QE deposits than domestic banks. Investors placed their money in prime MMFs, who in turn deposited them at foreign banks. This mechanism was broken by Money Market Fund Reform in 2016, which led $1t in assets to shift from prime funds to government funds. The run on prime funds last March also did not help investor confidence. However, negative yields can be motivating and the market usually finds a way to rewire itself. There will be new ways to manage credit risk and park cash on foreign banks without using prime MMFs.

Foreign banks absorbed the bulk of QE deposits in 2014

Slightly negative rates is not systemically threatening or even slightly ominous. Note that the Eurozone, Japan and Switzerland have had negative rates for many years without obvious problems. It may even be slightly risk positive, as investors rebalance along the risk curve to avoid the pain of negative rates. If the Fed perceives market rates as too low it could always just implement a technical adjustment, which it hinted at in the April FOMC Minutes.


  1. wabuffo

    Quick question:
    I imagine that the Fed is trading short-term T-bills (90 days or less duration) for reserves in order to minimize price risk due to a sudden change in Treasury security prices (even overnight)

    With overnight reverse repo running at $480 billion, I checked the Fed’s latest H.4.1 report to see their current inventory of Treasury securities. The Fed has $62.3 billion of T-bills due in less than 15 days and $316.5 billion of 16-90 day bills. That’s a total of $379 billion of short-term bills.

    What other securities is the Fed lending in reverse repo? Do they lend out from their inventory of 91-days +? What happens if reverse repo climbs further?

    Thanks in advance.

    • Fed Guy

      The securities put up as collateral in the RRP don’t have to be bills but can also be coupons. Note that the money funds do not rehypothecate their collateral, so all the Treasuries offered up as collateral in the RRP are essentially still out of the market. Also note that RRP is GC so participants do not choose collateral. Fed has a separate facility where it does lend specific securities, which you can read about here.

  2. Misua

    Good day, I have a question:

    Could one consider the pressure on MMFs and the ON RRP facility as a mechanism to eventually achieve one or more of the following goals:
    1) improving global $ liquidity (&thus recovery) by means of institutional reallocation into FX basis/foreign GSIB balance sheets.
    2) spurring corporations/(wealthy)households to either spend in the real economy or into duration, thereby supporting private and/or public spending.

    It seems to me that 1) & 2) would complement eachother well; Due to globalized nature of the economy, a domestic recovery would also require a global one. Further, foreign/smaller domestic banks taking up reserves would free up BS space (for said increase in private/public spending). This would explain why the TGA ceiling/SLR reinstatement aren’t a problem, and why an ON RRP hike/expansion or operation twist won’t be necessary. Lastly, it would fit within the context of an inequality fighting administration, since move into duration entails negative real yields, and because there’s quite some empirical evidence that robust local banks support a strong middle class.

    Having said that, the above would require that a practical ON RRP limit would eventually be reached, and that MMF outflow would encounter quite negative deposit yields at US GSIBs. If this is the case, perhaps the recent counterparty limit expansion (to $80B/counterparty) was carefully picked because it implies that, by the time a practical limit would be reached, the bulk of the TGA drawdown will be over, so that the Fed would have a stronger grip over the quantity of reserves in the system. This could enable a ‘testing period’; monitoring wether an at first small MMF outflow would achieve the intended effects, and to expand/raise rate of RRP facility in case it does not.

    The MMF outflow into duration might even function as a cushion to slow and eventually end QE?

    Apologies for the somewhat lengthy question/remarks.
    Kind regards, Misua

Leave a Reply

Your email address will not be published.

© 2021 Fed Guy

Theme by Anders NorenUp ↑