How the Treasury Repo Market Clears
This chapter focuses on how the overnight Treasury repo market turns securities into funding demand and how money market funds turn portfolio cash into funding supply. Understanding that operating model gives readers a concrete way to predict the direction of repo-rate pressure when Treasury issuance, dealer inventories, or money market fund assets change, rather than treating “liquidity” as one undifferentiated quantity.
A repo, short for repurchase agreement, is a transaction in which one party obtains cash against securities and agrees to reverse the exchange. For a dealer, it is a way to finance a Treasury position. For a cash investor, it is a way to lend against Treasury collateral. The rate is not a decorative price attached to the transaction: it is the cost of carrying the securities and the return available to the cash provider. When the number of securities that need financing grows faster than the cash available to finance them, the rate can rise.
Clearing does not require every borrower to find a single lender or every lender to accept the same trade. It is the result of many institutions adjusting at the margin. A dealer can seek more cash, sell or reduce inventory, or accept a higher financing cost. An MMF can lend more, buy a Treasury bill, or keep cash in another eligible instrument. The market rate is the price that makes those choices compatible for that day. This is why a repo market can have enormous gross volume and still become sensitive: high turnover does not guarantee that the next dollar of funding is easy to obtain.
Who brings collateral and who brings cash?
The scale of this market is easy to underestimate because it operates mainly in overnight transactions. The Federal Reserve’s note reports more than $8 trillion in daily trading volume across the repo market and examines the tri-party segment in particular, which has around $1.5 trillion in daily volume. The detailed mechanism in the note is therefore about an important segment, not a claim that every repo transaction has the same participants or structure.
In tri-party repo, primary dealers are the largest borrowers. They buy Treasury securities in the primary market and then finance those holdings by borrowing cash, with money market mutual funds—MMFs—supplying much of the lending. The dealer’s problem is not simply that a Treasury security exists. It is that the security occupies capital and must be carried until it is sold or otherwise financed. Repo gives the dealer a short-term funding channel against the position.
Hedge funds and other leveraged investors also borrow in repo to finance longer-dated, coupon-bearing Treasury securities. This matters because a Treasury coupon is not just a claim on future interest and principal. While it is held by a private investor, it is also an asset that may need financing. If more coupon securities move into private hands, or if dealers temporarily hold more of them, the demand for repo borrowing can increase even when the securities themselves remain safe and liquid in a broader sense.
The financing need can also vary with who is holding the security and why. A dealer that expects to distribute an auction position may need overnight funding only while the position is being placed. A leveraged investor may be financing a more persistent position. The repo market does not need to know the investor’s entire strategy to price the loan; it responds to the amount of collateral seeking cash and the willingness of lenders to meet that request. The same quantity of Treasury securities can therefore create different pressure at different times, depending on how much is being financed and how readily borrowers can adjust their positions.
The cash provider faces a different decision. An MMF with more assets under management has a larger potential pool from which to lend, but the whole pool is not automatically offered in repo. The fund allocates its portfolio among instruments that meet its objectives and constraints. The Federal Reserve note identifies Treasury bills as a close substitute for repo lending. An MMF can hold a Treasury bill or lend cash against Treasury collateral; the two choices serve different legal and portfolio functions, but both can be ways to hold a short-term, relatively secure asset.
A fund’s total AUM therefore differs from its marginal repo supply. Some assets may already be committed to bills or other holdings, and a fund may prefer one maturity or instrument over another as rates and expectations change. An increase in AUM can make more lending possible without guaranteeing that the additional cash will appear in repo immediately. Conversely, a fund sector that is not shrinking in total can still offer less repo funding if bills become more attractive or if cash is needed elsewhere in its portfolio.
This creates an interface between the borrower’s balance sheet and the lender’s portfolio. The dealer asks, in effect, “How much cash must I raise to carry these Treasuries?” The MMF asks, “Where should I place the cash in my portfolio?” The repo rate is one of the prices that helps reconcile those choices. A change on either side can matter, and the two sides do not respond to the same signal.
The market also has a vocabulary that distinguishes the funding rate from the policy rate. The tri-party general collateral rate, or TGCR, is a measure of the rate on tri-party repo transactions. The Federal Reserve note compares TGCR with the interest rate on reserve balances, IORB, and studies the spread between them. A TGCR–IORB spread can therefore be read as a way of tracking how secured funding is behaving relative to a rate associated with bank reserve holdings, while keeping the two markets conceptually distinct. Repo rates also determine SOFR, the secured overnight financing rate, which the note describes as the primary reference rate for trillions of dollars of financial contracts. These measures are closely related to the funding market, but they are not interchangeable labels for every short-term rate.
Federal funds are a separate market. Commercial banks and Federal Home Loan Banks, or FHLBs, are its main participants, and FHLBs account for most federal-funds lending. Repo is most liquid early in the morning, while federal-funds activity is concentrated later in the day. This segmentation is important: it means a change in repo rates is not automatically an equal change in EFFR. It also leaves room for a lender active in both markets to compare the two opportunities and move cash between them. The existence of that choice is the bridge through which repo conditions can matter for monetary policy.
The basic flow can be represented without pretending that every transaction follows an identical path:
The diagram shows two different channels rather than one generic “Treasury supply” effect. Coupon securities commonly create a need for financing on the dealer or leveraged-investor side. Bills can instead compete for the MMF’s lending capacity. Both can put upward pressure on repo rates, but they do so by changing different sides of the market.
Why bills and coupons pull the market in different directions
Start with a coupon-issuance day. When new longer-dated Treasury coupon securities settle, someone in the private sector must hold them. A primary dealer may have purchased them in the auction or may receive them through market-making activity. If the dealer does not want to fund the position from its own cash, it turns to repo. More inventory means more borrowing demand. If lenders are willing to supply the additional cash, the rate may not move much. If they are not, the dealer must offer a higher rate or reduce the position.
This is why coupon issuance can affect repo even though the Treasury securities are high-quality collateral. “Safe collateral” describes the credit characteristics of the asset; it does not guarantee unlimited balance-sheet space to finance it. The relevant question for repo is whether a borrower can find cash at a rate that makes carrying the position worthwhile. The financing need can be temporary, but an overnight market must clear it each day while the position remains on the balance sheet.
Consider two shocks of similar dollar size. If the first adds coupon securities to dealer inventories, it primarily asks the borrowing side to raise more cash. If the second adds Treasury bills that MMFs prefer to hold, it primarily removes cash from the lending side. If both occur together, the market faces more funding demand and less immediately available repo supply. The rate response need not equal the sum of two fixed effects, because borrowers and lenders can change their behavior, but the direction of the pressure is easier to understand once the two channels are separated.
The bill channel begins from the MMF’s side. Treasury bills are close substitutes for repo lending, so a larger supply of bills held by the private sector can give MMFs an attractive alternative use for their cash. If an MMF buys more bills, that cash is not simultaneously being offered in repo. The result can be upward pressure on repo rates even when no dealer has just added a matching quantity of coupon inventory. The security has changed the lender’s portfolio choice rather than directly creating a borrower’s financing request.
In practical terms, the dealer’s need and the MMF’s alternative can meet at the same rate. A dealer seeking cash may be willing to pay more to keep its Treasury inventory. An MMF that prefers bills may lend only if repo compensation becomes sufficiently attractive relative to the bill yield and the fund’s other choices. The market clears when those adjustments balance. A rise in the rate can therefore be a signal of a tighter marginal match between collateral and cash, not necessarily a sign that the entire financial system has lost liquidity.
That difference helps explain why the Federal Reserve note estimates separate effects for bill issuance and coupon issuance. It reports that, in its daily specification, both are associated with higher repo rates, but the mechanisms are not the same. Coupon issuance is tied to the quantity of longer-dated securities that dealers and hedge funds finance. Bill issuance works chiefly through the supply of cash that MMFs are prepared to lend. Treating both as a single Treasury-supply variable would hide which balance sheet is doing the adjusting.
MMF assets under management add another layer. An increase in government MMF AUM can expand the potential supply of repo lending, which tends to put downward pressure on repo spreads. But the note finds that the relationship between changes in government MMF AUM and daily repo spreads is only marginally statistically significant. That qualification matters. AUM may move gradually, while repo rates can react within a day to issuance, settlements, or cash movements. A larger fund sector can therefore make the market more capable in general without determining the exact amount of cash it will offer on a particular morning.
The phrase “repo capacity” should be read in that conditional sense. It is not the same as the gross volume of repo transactions, the total assets of MMFs, or the sum of reserves and ON RRP balances. Capacity depends on whether the relevant institutions want to use their balance sheets, whether their available assets are suitable for the transaction, and whether the return compensates them for doing so. An institution can possess cash and still prefer a Treasury bill, another market, or a different maturity. A dealer can possess Treasury inventory and still reduce its position rather than bid aggressively for financing.
This allocation problem is also why aggregate Federal Reserve liquidity does not automatically determine day-to-day repo rates. Reserves are held by banks and ON RRP balances by nonbanks, so their sum is a useful measure of liquidity supplied through the Fed. But the sum does not reveal how much cash is actually deployed into repo. Dealer and MMF behavior determines the market-facing allocation. A high aggregate total can coexist with little additional lending capacity if the cash is held by institutions that are not the marginal repo lender, or if attractive alternatives absorb it.
There are limits to this actor map. The Federal Reserve note concentrates its empirical analysis on the tri-party segment even though the repo market contains other segments, including centrally cleared activity discussed later in the note. The tri-party figures should not be treated as the volume of the entire market, and the mechanisms described here should not be extended to every participant without checking the relevant segment. The distinction between TGCR and SOFR also matters: the former is the rate used in the note’s tri-party analysis, while the latter has a broader reference-rate role.
The model nevertheless gives a reliable first prediction. More coupon inventory tends to increase the demand for repo funding. More bills can reduce MMF cash available for repo. More MMF AUM can expand potential lending supply, although the day-to-day relationship is not exact. If a shock raises funding demand and another shock reduces the supply of lendable cash, the pressure is compounded. If available cash expands or borrowers reduce inventory, the same market can clear with less rate pressure.
The Federal Reserve’s balance sheet enters this market by affecting both sides at once. Reducing the Fed’s Treasury holdings can leave more securities with private dealers, raising the amount they need to finance. It can also change where cash sits, including whether MMFs can draw on ON RRP balances or must attract deposits from banks. The next step is therefore not to add another definition of reserves. It is to trace how a balance-sheet decision changes private Treasury holdings, repo demand, and the source of the cash that meets it.