Friday, February 1, 2019

The Fed's Trouble with Leverage

QT is coming to end after a very short run. Since 2017, the Fed was attempting to "wind down" its massively bloated balance sheet of worthless mortgage securities and monetized debt (Treasuries) from the last financial crisis. While the truth was long overdue, in a recent WSJ article [1], Lorie Logan, an executive at the NY Fed, finally admitted last May that she saw, "virtually no change of going back to the pre-crisis balance sheet size." Unfortunately, what this means is that with an end to QT, there will be an inevitable transition to QE.

Even more troublesome is looking at what's already on the Fed's balance sheet. On first glance at the balance sheet as of Jan. 16th, 2019 [2], the Federal Reserve has a leverage ratio of approximately 103:1 (total assets: total capital). However, the real problems are swept under the footnote rug. An article by Alex Pollock [3] delves into the Fed's accounting rules, and sheds light on some financial shenanigans. Supplemental Information #2 in the Sep. 30th, 2018 financials [4] shows information on "System Open Market Account Holdings," In other words, these are the securities that the Fed purchased after 2008 as part of the QE programs.

It writes: "Treasury securities, government-sponsored enterprise (GSE) debt securities, and federal agency and GSE mortgage-backed securities (MBS) are reported at amortized cost in the Combined statements of condition." (emphasis mine)

First of all, is it allowed? According to rules from 1931 and 1992, it is. In 1931, it was allowed that banks could value certain "investment-grade bonds" at "intrinsic value" instead of at market. This allowed for other methods which were deemed more appropriate at coming to the real value of the securities, where the market would mark them too low. A government accounting rule enacted in 1992 pertaining specifically to real-estate related assets, extended the old rule, allowing their net worth to be valued by their "intrinsic value" as well. Unfortunately, it is also permitted that a government and the Fed can make up their own accounting rules.

It is important to point out that amortized cost is not the same as mark-to-market. Mark-to-market method is much more conservative, and if there are losses in these securities, it underestimates the number. As it happens, the Fed does have unrealized losses on those QE-purchased securities. Contrary to what economists and mainstream media practitioners say, those securities are indeed worthless.

Moving onto the table below the footnote, it shows just how big these losses are, at amortized cost. While the cumulative unrealized losses on Treasury Securities notes are approximately $27.76 billions dollars (as of 9 months ended, 2018), but the total Treasury Securities unrealized loss is only $770 million. Meanwhile you also have to factor in the other row: the total cumulative unrealized losses on Federal agency and GSE MBS are $66.453 billion.

And what's paid in capital again? As of Jan. 16th, 2019, the total capital paid-in is $32 billion. 
According to Mr. Pollock from an interview in the Grant's Pub Podcast [5], the Fed can raise another $32 billion from shareholders. That is, from commercial banks who are members of the Federal Reserve.

But, it's still not enough. How will the Fed cover the rest of the losses? With interest rates rising, the unrealized losses will grow bigger, and the Fed's net worth will be more negative. As noted earlier, there is a strong probability that the Fed will attempt QE yet again to reverse the losses.

Since I do not claim to be a forecaster, I will invoke Mark Twain on this one: History seldom repeats, but it does rhyme. This is to say that this instance of financial shenanigans is quite reminiscent of the 1980s savings and loans crisis. Readers of history are all too aware of the ending in that story…