The 88% figure doesn't match the sources. The Fed did not buy 88% of all mortgages. It held a large share of the mortgage-bond market, and the 88% looks like a misreading of a different number.
Here is what the record shows:
- Holdings: The Fed's agency MBS holdings rose from about $1.4 trillion in March 2020 to a peak of roughly $2.7 trillion in 2022, according to the Dallas Fed and Axios reporting of the Fed's H.4.1 data.
- Share of the market: The Brookings paper by Aaron Klein and Alan Cui (October 8, 2025) puts the Fed's $2.7 trillion in 2022 at about 32% of an $8.5 trillion agency MBS market. I could only read this through a secondary summary, not the full paper.
- Share of growth: The same summary says the Fed's purchases were nearly 90% of MBS growth during the 2020–2022 QE period. That is probably the source of the 88% claim, but it is a share of new issuance, not of all mortgages.
- Mortgage debt: About 65% of home mortgage debt was securitized into MBS as of 2021, per a New York Fed staff report by Andreas Fuster. So the Fed's holdings are a share of that securitized pool, not of every mortgage.
The Brookings argument is that these purchases lowered mortgage rates and pushed up home prices. Other research, including a Kansas City Fed paper cited by Cato, finds a smaller spread effect, around 40 basis points from the Fed and similar from banks. Economists still debate how much of the housing price rise came from QE.
If you want to check the numbers yourself, start with the Fed's H.4.1 release and the Brookings paper. Add what you find to this thread.
Also discussed here:
- Two Sea Spiders Nobody Had Named in Nearly a Hundred Years
