Skip to Content
Enter
Skip to Menu
Enter
Skip to Footer
Enter
Blog
Fundamental Analysis

Net Liquidity as a Formula: Fed Balance Sheet Minus TGA Minus RRP

Category:
Fundamental Analysis

min read

Share this post
Net Liquidity as a Formula: Fed Balance Sheet Minus TGA Minus RRP

In the autumn of 2021, the Federal Reserve had already begun signaling that its bond-buying program would slow. Yet risk assets kept climbing. The S&P 500 made fresh highs into January 2022. Many analysts, watching only the headline balance sheet, could not explain why the tightening rhetoric produced no visible tightening in markets.

The answer was hiding in two accounts that most investors never think about. The Treasury was drawing down its checking account at the Fed, and the Fed's reverse repo facility had not yet begun absorbing the excess. Both actions pushed cash into the banking system faster than the Fed's tapering pulled it out. The relevant number was never the balance sheet alone. It was the residual.

The residual that actually reaches markets

Net liquidity is the pool of cash genuinely available to the financial system after the Treasury and the Fed's money-market plumbing take their share. The formula is simple to write and easy to underestimate:

Net Liquidity = Fed Balance Sheet − Treasury General Account (TGA) − Reverse Repo (RRP)

The Fed's balance sheet is the total stock of assets the central bank holds, mostly Treasuries and mortgage bonds. When the Fed buys a bond, it creates reserves, which are the electronic cash banks hold at the Fed. That is the number the headlines track.

But not all of that cash is free to circulate. The Treasury General Account is the U.S. government's checking account at the Fed. When the Treasury raises money by issuing bonds and lets the cash sit in the TGA, that cash is drained from the banking system. It exists on the Fed's books but does not fund anything in markets.

The reverse repo facility, or RRP, is where money market funds park cash overnight in exchange for Treasuries, earning a fixed rate set by the Fed. Cash sitting in the RRP is likewise removed from circulation. It has left the private system and gone back to the central bank for safekeeping.

Subtract those two drains from the balance sheet and you get the cash that actually reaches banks, dealers, and ultimately risk assets. That residual, not the gross balance sheet, is what moves with the market.

The practical implication is direct: when you read that the Fed is tapering, do not assume liquidity is falling. Check whether the TGA and RRP are draining faster than the balance sheet is shrinking.

Why the two drains move markets as much as the Fed does

The mechanism rests on a single accounting truth. The Fed's liabilities must sum to a fixed total at any moment, so a dollar that moves into the TGA or RRP is a dollar that leaves bank reserves. These are not independent dials. They are competing claims on the same pool.

Consider the TGA. When the Treasury spends down its account, say by sending stimulus payments or paying contractors, that cash flows out of the government's checking account and into private bank deposits. Bank reserves rise. The system has more usable cash without the Fed buying a single bond. That is precisely what happened through 2021: the TGA fell from roughly $1.6 trillion in early 2021 toward a few hundred billion by that autumn, releasing over a trillion dollars into the system.

Now consider the reverse repo facility. Between 2021 and 2023, money market funds had more cash than they could safely deploy, so they parked it at the RRP, which peaked above $2.5 trillion in late 2022. Every dollar there was a dollar not chasing Treasury bills, corporate paper, or riskier assets. When the RRP later drained through 2023 and 2024, that cash re-entered the system and cushioned markets even as the Fed was actively shrinking its balance sheet.

The causal chain runs cleanly. A drain in the TGA or RRP raises bank reserves, which lowers the marginal cost of funding, which supports the prices of assets that trade on leverage and liquidity. A build in either account does the reverse. The Fed's headline actions are only one of three forces, and often not the largest one in a given quarter.

For you, the takeaway is that Treasury financing decisions are a monetary tool the Treasury does not admit to wielding. When the Treasury announces it will rebuild the TGA after a debt-ceiling standoff, it is scheduling a liquidity drain, regardless of what the Fed is doing.

Reading net liquidity in the present

Today the balance sheet is smaller than its 2022 peak, but the story is again about the residual, not the headline. The 10-year Treasury yield sits at 4.70% and the 2-year at 4.22%, leaving a positive 2s10s spread of about 0.48%. A steepening curve with elevated front-end yields keeps the RRP largely drained, because money funds earn more buying bills directly than parking cash at the Fed. That drained RRP has already delivered most of its cash to the system.

SHY over the thesis window.
SHY over the thesis window.

The next swing factor is the TGA. Every time the Treasury raises its cash target, whether to term out debt or rebuild a buffer after a financing crunch, it withdraws reserves on a schedule set by auction sizes rather than by the Federal Open Market Committee. A reader tracking only the policy rate would miss this entirely.

The mechanism also explains a persistent source of confusion. Markets sometimes rally on a rate-hike day and sell off on a pause. The rate decision is the visible event; the liquidity residual is the mechanism. If a hawkish meeting coincides with a large TGA drawdown, reserves can rise even as the policy rate climbs, and risk assets follow the reserves.

The practical implication is to build a simple weekly habit. The three components are all published: the balance sheet in the Fed's H.4.1 release, the TGA in the Treasury's Daily Statement, and the RRP in the New York Fed's operating data. Track the net of the three, not any one in isolation.

Where the formula misleads

Net liquidity is a powerful lens, but it is not a mechanical trading rule, and treating it as one is the most common error.

The relationship between the residual and asset prices is a correlation of flows, not a law. In 2018 the balance sheet was shrinking and the RRP was near zero, yet the market's fourth-quarter selloff owed more to earnings fears and a hawkish Fed communication error than to any precise liquidity number. The residual framed the backdrop; it did not dictate the day-to-day path.

The formula also omits channels that matter at the margin. Foreign central bank activity, the Fed's standing repo facility, and shifts in bank demand for reserves all affect how much of the residual is truly spare versus needed to satisfy regulatory buffers. When banks decide they need more reserves for their own safety, a given level of net liquidity feels tighter than the same level a year earlier.

And the causation can reverse. A market crash can drive cash into money funds and back into the RRP, so the RRP build is the consequence of stress, not its cause. Reading the build as a fresh drain in that moment would get the direction exactly wrong.

The discipline is to treat net liquidity as the tide, not the wave. It tells you whether the structural backdrop is supportive or draining over weeks and months. It will not time a single session, and anyone who backtests it as a daily signal will find the fit falls apart precisely when volatility spikes.

What the residual tells you

Net liquidity reduces a diffuse question, whether money is flooding or draining the system, to a single subtraction: the Fed's balance sheet minus the Treasury's checking account minus the reverse repo pile. The residual matters because it, not the headline balance sheet, is the cash that actually reaches banks and risk assets, and the two drains can overwhelm the Fed's own actions in either direction. Watch the net of all three, treat it as a slow tide rather than a daily signal, and you will understand why markets sometimes move opposite to what the policy rate alone would predict.