The mechanism is a balance-sheet swap, and getting this right dissolves most of what is confusing about it. When the Federal Reserve buys a Treasury bond or an agency mortgage-backed security, it credits the seller's bank with reserves held at the Federal Reserve. The central bank's assets rise by the securities and its liabilities rise by the reserves. No currency is printed, no appropriation is spent, and nothing is deposited in anybody's checking account. What has happened is that a long-dated bond has moved from private hands onto the central bank's books, and an overnight claim on the central bank has moved the other way.
The intended effect follows from that swap. With fewer long-dated bonds available to private buyers who want them, their prices rise and their yields fall, and other long-term borrowing costs priced off those yields fall with them. The Board describes the purpose of the 2008 through 2014 programs in those terms: it "greatly expanded its holding of longer-term securities through open market purchases with the goal of putting downward pressure on longer-term interest rates and thus supporting economic activity and job creation by making financial conditions more accommodative."
The household-facing channel is mortgages, and one of the purchases carries that purpose in the Board's own sentence. The first program the Board enumerates is $175 billion in direct obligations of Fannie Mae, Freddie Mac and the Federal Home Loan Banks, bought between December 2008 and August 2010, and the reason sits inside the sentence that reports it: the Federal Reserve made those purchases "to help reduce the cost and increase the availability of credit for the purchase of houses." The Board then records, without restating a purpose, a further $1.25 trillion in mortgage-backed securities guaranteed by Fannie Mae, Freddie Mac and Ginnie Mae between January 2009 and August 2010, which is the largest single program on the list. Both bought housing-related paper, so the tool was not aimed only at abstract financial conditions; one of its two main targets was the mortgage market a household actually borrows in.
Quantitative tightening is the reverse, and it usually works by not reinvesting rather than by selling. In October 2017 the Committee began a balance sheet normalization program that, in the Board's description, "gradually reduced the Federal Reserve's securities holdings by decreasing its reinvestment of principal payments received from securities held in the SOMA." When a bond the central bank owns matures, it is simply not replaced, and the balance sheet shrinks passively. That distinction matters because a program of outright sales into the market would be a much sharper intervention than letting holdings run off.
Two honest complications, both from Federal Reserve sources. The first is the relationship between the two directions. Governor Christopher Waller argued in a March 2024 speech that "it is very important that QE be credibly followed by QT," reasoning that if the purchases were "viewed as nothing more than a permanent injection of money into the economy, it would likely create inflation," and that the widely predicted inflation after 2009 did not arrive, in his view, because "the Fed credibly committed to withdrawing the injected reserves at a later date." That is one governor's assessment rather than a Committee position, and it is the clearest available statement of why the two operations are treated as a pair.
The second is that shrinking the balance sheet has its own limit. Waller described the money-market stresses of autumn 2019, when reserves had been reduced through the normalization program and heavy Treasury issuance arrived in September, and concluded that "the level of reserves likely went a bit too low." Reducing the balance sheet is not simply undoing the purchases; it runs into the banking system's own demand for reserves.