naked capitalism

The Treasury Department, as reported by Bloomberg, and commented on by Rolfe Winkler and Huffington Post (among others) noted, considerably increased its Freddie and Fannie safety net, by removing all limits on the amounts on offer (an increase from a ceiling of $400 billion) and simultaneously allowing the two GSEs to increase their balance sheets near term. Previously, they had been required to shrink their portfolios by 10% per annum; now it is their ceiling which will be lowered by 10% a year, and that ceiling is much higher than their current exposures ($900 billion versus roughly $760 billion for Freddie and $770 billion for Fannie as of the end of November).

A stunner reported at the same time was that Treasury bought $220 billion of MBS last year, in addition to the massive Fed purchases.

The timing of the announcement (Christmas Eve) was designed for it to receive as little notice as possible, already a bad sign. Skeptical observers focused on two possible explanations: first, that losses at the GSEs will be high, troubling investors, and the offer of unlimited support will calm them. Freddie and Fannie debt is now de facto a full faith and credit obligation of Uncle Sam, as if there was any doubt of that once the conservatorship was announced. But the idea that nearly $300 billion of firepower (the agencies had used roughly $110 billion of an authorization of $400 billion) is troubling.

The relaxation of the requirement that the GSEs shrink their balance sheets is also not pretty. I’m sure the argument would go something like “after all this extraordinary support to the housing markets, we don’t want Fannie and Freddie reducing their commitments at the same time when the Fed is ending its support.” That could have been achieved merely by pushing out the shrinkage timetable a year, rather than allowing for Freddie and Fannie to conceivably increase their exposures by over $200 billion between them. So one has to conclude that the agencies might well (ahem, are likely to) throw their firepower behind the “prop up the mortgage market” program, particularly with Obama’s ratings plunging and mid-term elections coming this year.

But if this comes to pass, what might the collateral damage be? Well first, even if you are not of the Austrian (malinvestment) persuasion, subsidizing the housing market to this extent is simply not a good idea. Housing is not a productive investment, and some research suggests that high levels of consumer indebtedness are correlated with lower levels of GDP growth (business debt, by contrast, is positive provided it is funding useful projects, versus, say, land and stock market speculation, as was the case in bubble era Japan).

But second, having the GSEs increase their balance sheets (assuming that does occur) creates other complications. John Dizard explained this back in 2008:

In the case of the US government sponsored enterprises, the biggest of which are Fannie Mae and Freddie Mac, for example, we are now about to get into the same mess we only crawled out of about three years ago..

At the beginning of this decade, derivative risk management geeks, interest rate swaps traders and central bank econometricians filled up entire server farms with what-ifs on the balance-sheet hedging activities of the GSEs. The essential problem was that the GSEs were balancing ever-larger portfolios of fixed-rate mortgages on tiny equity bases. Fortunately, as we all knew, the credit risks of those portfolios were limited because homeowners rarely default on their mortgages {Yves here, Dizard is being ironic]. But that still left very large interest rate risks.

The core problem for the housing GSEs is, and has been, the prepayment option embedded in US fixed-rate mortgages. That has meant that the term of the GSE assets extends or contracts depending on whether homeowners can refinance at an advantageous rate. However, most of the long-term debt on the liability side of the GSE balance sheets has a fixed term. So the GSEs must more or less continually offset this imbalance between the average maturity of their assets and liabilities through the derivatives market, specifically the interest rate swap market. Otherwise the mark-to-market losses would overwhelm their small equity bases.

This process of risk control on the part of the GSEs creates systemic risk for the fixed-income markets. GSE hedging tends to be pro-cyclical. As interest rates rise, the average term of the GSEs’ assets extends, since homeowners are not refinancing. As rates fall, the average term contracts, as homeowners prepay the mortgages on the GSE books. So the hedging activities tend to accentuate market moves. As rates rise and bond prices fall the GSEs are, in effect, selling fixed-income derivatives into a falling market. As long as the derivatives books are small relative to the size of the market, that is not a big problem. When the GSE derivatives books got big, that was a problem.

By 2001 Fannie and Freddie together had more than 10 per cent of the total market in dollar-based interest rate derivatives [Yves here, that also happened to be the position LTCM got itself into right before its meltdown...]. That concentration of risk was worrisome for the central banks. As we wrote at the time, they were concerned that the banks and brokers who were the counterparties for the GSEs would need back-up for these commitments from the Federal Reserve Board…
Yves here. Since everyone who counts is now backstopped, we of course have no reason to worry now. Back to Dizard:
Then Mr Greenspan, the GSE regulators and their geeky allies got lucky. A management compensation scandal broke at the GSEs that quickly turned into a more general accounting scandal. The reformers had the political wind at their back, and as the accountants and lawyers sifted through the books, the portfolio growth reversed. Even better from a systemic stability point of view, the GSEs’ share of the interest rate derivatives markets dropped by more than two-thirds by 2005. As homeowners took on more adjustable rate mortgages, they assumed some of the rate risk the GSEs shed.

Unfortunately, the squeezed balloon of mortgage credit just bulged out elsewhere. The GSEs, and the rest of the financial markets, assumed more credit risk, and they are now incurring those very real losses.

This recent history seems to have been forgotten by the government and the financial institutions. The caps on GSE portfolio growth have been lifted, and Congress and the markets are now asking them to take on the mortgage assets that everyone else wants to sell…

If this balance sheet growth does happen, the GSEs will be back to assuming the same rate risks that were so alarming four or five years ago, only bigger. And they will be attempting to hedge their rate risks using counterparties that are far more capital constrained than before.

Yves again. A lot of commentators have worried about credit risk at the GSEs, and that is a real concern. But they have missed the increased interest rate assumption taking place, and that is occurring independent of portfolio growth as more mortgages are refinanced from floaters to fixed. And that could conceivable be exacerbated if the GSEs expand their portfolios in 2010. But that very growth increases systemic risk. A lovely equation.

One has to wonder if this calculus contributed to the Treasury’s decision to give Freddie and Fannie open-ended support.