U.S. banking problems: Mortgage interest rate risk strikes again
Published in Housing Finance International.
Notable U.S. bank runs and failures in spring 2023 followed speeches from the Treasury Department and the Federal Reserve in 2022 and early 2023 about what a good shape the U.S. banking system was in. Of course, they wanted the non-bank financial companies (so-called “shadow banks”) to be regulated more, but they claimed the commercial banks were in great condition. This was thanks to the Dodd Frank Act and its greatly expanded regulation, it was suggested, and we could congratulate ourselves for this banking stability.
Surprise! Shortly thereafter came the second, third and fourth largest bank failures in U.S. history (First Republic, Silicon Valley, and Signature banks, respectively). These were accompanied by emergency government declarations that the failing banks were “systemically important”—in other words, that the banking system in general was at risk. Such declarations allowed bailouts of wealthy depositors who had accounts with the failing banks of far over the normal statutory maximum of $250,000 for government deposit insurance coverage. Those depositors should have taken a loss on the risky investments they had made and from their folly in maintaining hundreds of millions, in some cases, or even billions of dollars, in single small-enough-to-fail banks. That was quite unwise risk taking on their part, but big money venture capitalists, tech entrepreneurs and cryptocurrency barons, among others, got all their deposits back by being given other people’s money. The insurance fund of the Federal Deposit Insurance Corporation was reduced below its minimum statutory level by the cost of these failures.
Central to the failures were billions of dollars in losses resulting from investing in long-term, fixed rate mortgage securities and financing them with short-term, floating-rate liabilities. The interest rate risk of the American 30-year fixed-rate mortgage strikes again! Although these mortgages were often held in securitized form, making them theoretically liquid, once their market value was far below their cost, selling them would trigger the realization of the huge losses, which could and did set off a run on the bank. So just as in the case of the U.S. savings & loans in the 1980s, the shortfunded 30-year mortgages contributed to first massive interest rate risk and then fatal liquidity risk for the failed banks.
However, not only the failed banks, but hundreds of banks across the country have made this mistake. In doing so, they were following the Pied Piper of the Federal Reserve and its long lasting, but now ended, suppression of both short-term and long-term interest rates, its predictions of “lower interest rates for longer,” and quite remarkably, its own practice of investing trillions of dollars in long term fixed rate assets funded by borrowing short in its own balance sheet. Investing in long-term mortgages and also long-term Treasury securities, the banks and the Federal Reserve together created huge amounts of interest rate risk for the total banking system.
This is a classic banking risk, a classic banking mistake. Along with it comes a classic problem: although it is often said that bank runs reflect irrational fears, in fact runs are highly rational from the viewpoint of the depositor.
Keeping your money in a questionable bank instead of getting it out, when you can’t really know what the bank’s assets are worth or what is really going on inside it or who is lying, has no upside potential and large downside potential of serious losses on money you didn’t intend to be at risk at all. So, it is rational to run.
Every bank should be structured with the thought that runs are rational from the viewpoint of the depositors. This is an old and essential banking truth. As the great Walter Bagehot wrote in Lombard Street in 1873, “A banker, dealing with the money of others, and money payable on demand, must be always, as it were, looking behind him and seeing that he has reserve enough in store if payment should be asked for.” Bagehot adds, “Adventure is the life of commerce, but caution is the life of banking.”
One recent academic analysis has estimated the mark to market loss of all U.S. banks, including both their fixed-rate securities and fixed-rate loans, and concluded that the banking system has a mark to market loss of about $2 trillion on these assets. That’s $2 trillion.
In comparison, the book capital of the U.S. banking system is about $2.2 trillion, and of that $2.2 trillion, $400 billion are intangible assets. So, the tangible capital is approximately $1.8 trillion, while the mark to market loss may be something like $2 trillion. That would suggest that we have an entire banking system with something like zero real capital. While there is large variation among the 4,700 banks and savings associations, the overall situation is certainly sobering on a mark-to-market basis.
The way unrealized losses turn into real losses in financial institutions that have fixed rate, long term assets funded with short term floating rate liabilities, is that the cost of their liabilities goes dramatically up and the yield on their assets doesn’t.
The big question in the banking system has thus become: how resistant are the banks to the cost of their deposits rising to full market levels of 5% or so, in order to fund the deeply underwater fixed rate loans and securities on which there is a $2 trillion market value loss? How much is their interest cost going to rise? In other words, as the unrealized losses turn into greater interest expense, how much can you continue to cheat the savers in order to prop up the banks? What we will observe going forward is the contest between the inexorable rise in the funding cost, and the hope that banks can continue to cheat the savers and keep the deposit cost below market. This puts the banking regulators, who are in favor of bank safety, on the side of cheating the savers. All this is just like the U.S. savings & loans and their regulators in the 1970s and 1980s.
On top of its interest rate risk, we find out the U.S. banking system has, once again, a large risk exposure to troubled commercial real estate. Consider this observation: “The unfavorable conditions in banking were greatly aggravated by the collapse of unwise speculation in real estate.” That is from the report of the U.S. Comptroller of the Currency in 1891. Little about this connection has changed since then.
The famous Fed Chairman Paul Volcker observed that “There are no new problems in banking, only new people.”
But why do crises keep recurring if we have comprehensive regulation, like the Dodd Frank Act and its thousands of pages of implementing regulations? The banking scholar, Bernard Shull, insightfully wrote, “Comprehensive banking reform, traditionally including augmented and improved supervision, has typically evoked a transcendent, and in retrospect, unwarranted optimism.”Writing in the 1990s, Shull continued, “Confronting the S&L disaster with yet another comprehensive reform, the Secretary of the Treasury proclaimed ‘Two watchwords guided us as we undertook to solve this problem. Never again.’”
That was the savings & loan reform of 1989, and then came the banking reforms of 1991, and the housing finance reforms of 1992. But in the following decade, instead of “never again,” another massive crisis happened again anyway. Then after that 2007-09 crisis, comprehensive reform was once again debated with the political result of the Dodd Frank Act. But another decade or so later, once again we have bank runs and failures.
We are up against the reality of the politics of banking and finance. All finance is political, especially housing finance. No matter how we politically organize any bank regulation, it will over time have to be reorganized, because the perfect answer does not exist. Whenever we try to engineer and control human behavior, the attempts at control themselves induce unexpected adaptations and reactions in the behavior of financial markets, and also in the behavior of the regulators and politicians themselves. Hence, as we observe historically, every reform requires another reform to address the unexpected results and failures of the prior reform, and so on, ad infinitum. The current troubles are part of the ongoing crisis of 2020. Starting in 2020, we experienced the Covid crisis, the financial panic, the sharp economic contraction, the vast expansion of money printing and government expenses in response, the following bubble markets, the everything bubble, the runaway inflation, the correction with rising interest rates, the deflation of the everything bubble in 2022, and now the banking problems of 2023. I think of that as all one big crisis-- we’re still in the aftermath of 2020.
In that aftermath, with assets inflated to $8.4 trillion, the Federal Reserve is by far the biggest bank in the country. It is also the biggest savings & loan, because inside the Fed is a mortgage portfolio of $2.6 trillion of long-term fixed rate mortgages, funded short with floating rate liabilities. This makes it far and away not only the biggest savings & loan in the country, but in all of history. The Fed is now earning more or less 2% on its mortgages, and more or less 2% on its long-term Treasuries. The duration of its mortgages has become longer than expected with their negative convexity and increasing interest rates. Its Treasuries are also very long, over $1.4 trillion with remaining maturities of more than 10 years. While earning 2%, the Fed is now paying over 5% to carry those investments. If your income is 2% and your cost is 5%, it’s hard to make money that way! From September 2022 to early June 2023, the Fed has had a net loss of $68 billion, and is on its way to an annual loss of about $100 billion, more than twice its capital.
The Fed’s own balance sheet was and is jammed with gigantic interest rate risk. One might not unreasonably ask, how could the Fed criticize banks like the failed Silicon Valley Bank for doing exactly the same thing that the Fed itself was and is doing? A fair question!
Of course, if the failed banks or the technically insolvent Fed had been right in the forecast that short-term interest rates would stay abnormally low, that kind of balance sheet would have been a profitable bet. They could have had 2% income and say, 0.25% interest expense. So let us ask: what was the Fed’s own interest rate forecast as it was accumulating its massive portfolio and interest rate risk?
Consider the Fed’s interest rate forecast done in June 2021 for the end of the year 2022. The median forecast for fed funds target was 0.25%. Of Federal Open Market Committee members submitting estimates, the highest rate was 0.75%. Compared to the 4.25%-4.5% reality, a big miss indeed!
What about the projection for 2023 at that same meeting? What did they think the fed funds target interest rate would be by the end of 2023? The median forecast was 0.75%. The highest submitted estimate was 1.75%. Another big miss, to say the least, with the rate now at 5%-5.25% five months into 2023.
The Fed’s poor forecasts bring to mind that famous American baseball poem, Casey at the Bat. In trying to forecast interest rates, and inflation, and banking problems, just like “Mighty Casey,” the mighty Fed has struck out.
In the meantime, the spring of 2023 provided some extreme drama in U.S. banking, tied significantly to the risk of the American 30-year fixed rate mortgage-- and this cycle is not over. Whatever may the coming in the way of further increases in interest rates, or interest rates continuing at current levels, or an often-predicted recession with lower interest rates but increased loan defaults, all will cause risks to the banking system. In any case, financial markets should remember not to put their trust in the forecasts of the Federal Reserve.