The Federal Reserve plunges into the mortgage market and takes a bath

Published in Housing Finance International Journal.

The most important central bank in the world, the Federal Reserve, owned on September 30, 2024, the staggering sum of $2.3 trillion in mortgage securities, making it by far the biggest investor in the “agency MBS” market which dominates U.S. housing finance. These are the mortgage securities guaranteed by Fannie Mae, Freddie Mac and Ginnie Mae, which total $9.3 trillion. Thus, the central bank owns 25% of the total relevant market, a massive concentration of interest rate and market risk. How has the Fed done on this investment?

The Fed’s mortgage securities portfolio had, as of September 30, a mark-to-market loss of $315 billion, or more than 7 times its stated total capital of $43 billion. The average yield on its mortgage securities portfolio is only 2.2%, while its deposits are costing it 4.65%, for a negative spread on these investments of 2.45%. This means the Fed has an annualized operating loss on its mortgages of $56 billion, equal to a cash loss per year of 130% of its stated total capital. (The Fed is in addition losing money on its portfolio of long-term Treasury bonds.) The underlying mortgages of the Fed’s mortgage securities are primarily 30-year fixed-rate loans, so they have a long time yet to run. In this respect, the Fed looks just like the biggest 1980s-style savings and loan in history. It is quite amusing to think about what Federal Reserve examiners would say to any regulated bank which ran such risk and produced such outcomes!

Moreover, the Fed must transfer most of any profits it makes to the U.S. Treasury, so when the Fed has losses, its losses are also losses to the Treasury and costs to the taxpayers, imposed on them without legislation or Congressional approval. The Fed’s losses increase the government deficit and the government’s debt. As the provocative financial observer Wolf Richter has written, we are “in the era of the money-losing Fed.” The U.S. has never been in such an era before.

In the process of making these losing investments, the Fed greatly distorted the U.S. housing market. Its purchases of mortgage securities with newly created money, or the monetization of mortgages, resulted in a rapid expansion of abnormally cheap mortgage credit, driving the interest rates on 30-year fixed rate mortgages to below 3%, and fueling a rapid acceleration in house prices which has made them unaffordable to large segments of the American population. This was the Second U.S. Housing Bubble of the still young 21st century.

Many people, including me, thought that when U.S. mortgage interest rates inevitably went back to normal levels of 5%-6% or more, average U.S. house prices would fall. They did fall a little, but then started back up. With the standard U.S. 30-year mortgage now costing 6.8%, U.S. house prices nonetheless went up about 4% over the last 12 months.

The most common theory to explain this surprising outcome is that the lucky borrowers of 3% mortgages with the rate fixed for 30 years don’t want to lose the large financial advantage of their abnormally cheap debt by moving, so they disproportionately stay in their houses, thereby reducing the normal supply of houses available for sale and holding up house prices. If this be true, it is another striking source of distortion in the housing market by the Fed.

Quoting Wolf Richter again: “If [house price] charts look absurd it’s because the housing market has become absurd. Housing market charts should never ever look like this. They’re documenting the crazy distortions triggered by the Fed’s monetary policies.” I believe Wolf is correct about this.

Does a huge position in mortgage securities belong on the balance sheet of the U.S. central bank? An authoritative Federal Reserve study of what assets the Fed might invest in, in addition to Treasury securities, is Alternative Instruments for Open Market and Discount Window Operations (2002), written by an expert Federal Reserve System Study Group. This careful study articulated as two of its essential principles that the Fed should:

“Structure its portfolio and undertake its activities so as to minimize their effect on relative asset values and credit allocation within the private sector.”

“Manage its portfolio to be adequately compensated for risks and to maintain sufficient liquidity in its portfolio to conduct potentially large actions on short notice.”

It is apparent that the Fed’s mortgage investments have violated both these principles.

As for the first, they had a major effect on relative asset values, pushing the price of houses up at double-digit annual rates. And they were a massive allocation of credit by the central bank to a particular sector within the private economy, namely housing.

Regarding the second principle, instead of being adequately compensated for its risks, the Fed is suffering, as discussed above, operating and market value losses on its mortgage investments far greater than its capital. As for liquidity, the Fed can’t get out of its mortgage position in any short term—the position is too big relative to the market. Selling it would entail realizing the portfolio’s market value deficit, thus turning paper losses into cash losses.

It is not a practical possibility for the Fed to liquidate its mortgage portfolio on short notice. If it did manage to sell, at least $315 billion in new cash losses would have to be reported in its profit and loss statement. Adding these to the $210 billion in operating losses it has already reported, that would bring its total cash losses to well over half a trillion dollars. Recall that this is also an increase in the federal deficit and the federal debt. Such losses would not only be an embarrassment to the Fed, but might trigger Congressional attention it doesn’t want.

Worse, the Fed would not be able to unload its mortgage securities for as much as its fair value estimate. Trying to quickly sell its massive position would drive the market prices down against it, causing a loss much greater than $315 billion--indeed it might crash the mortgage-backed securities market. Even the announcement that the Fed planned to sell its securities gradually over time would probably lower MBS prices and thereby increase mortgage interest rates—hardly a welcome political outcome.

So it looks like the Fed will have to just hold its losing mortgage securities until the underlying 30-year mortgages slowly mature over time, as “the era of the money-losing Fed” continues. It does not appear that anyone, including the new Trump Administration, can change this outcome.

Looking further forward, the troublesome results of the Federal Reserve’s mortgage investing experiments will certainly provide good material for future historians and theorists of central banking and housing finance.