Ut-oh

Agency spreads have widened. See John Jansen for all the gory details.

The New York Fed’s custodial holdings of Agencies haven’t grown at anything like their typical pace over the past few weeks. In the first two weeks of August, custodial holdings of Agencies fell by $6.7b while custodial holdings of Treasuries rose by $24.7b.

And -- via Yves Smith -- comes word that foreign central banks have been a bit more reluctant than usual to buy the debt the Agencies issue to refinance their retained mortgage portfolio. Lynn Adler of Reuters writes:

Overseas investors took an atypical back seat in Fannie Mae’s three-year note sale this week.

Central banks bought just 37 percent of the $3.5 billion issue, down from 56 percent in May’s $4 billion offering of the same maturity. Asia accounts took just 22 percent of the notes, down from 42 percent in May. "Most fixed income investors to whom we have spoken believe that a capital infusion by the government into Freddie and Fannie is a prerequisite for turning sentiment around in mortgage-backed securities and, by extension, in the broader fixed income markets," Barclays Capital analysts Rajiv Setia and Philip Ling wrote in a report The longer the debate drags on, the more tentative foreign interest in the sector is likely to become. Even though the GSEs are adequately capitalized, investor confidence has been shaken," the analysts wrote. "A slowdown in international investor interest remains the major risk factor for agency spreads, in our view."

No wonder there is a lot of interest in Asian (read "central bank") demand for Tueday’s Freddie Mac auction.

The Treasury’s bailout plan for Agencies sought to retain the Agencies current "hybrid" structure, one where the Agencies continued to be privately owned even as they borrow in the market at (relatively) low spreads based more on the expectation that they are too big, too important and too Chinese to fail than on the strength of their balance sheets. Larry Summers noted: "almost every outside observer agrees that pre-crisis, the GSEs could only borrow because of their implicit government guarantees. Since the crisis their position has sharply deteriorated, and will deteriorate further." The Treasury’s plan hinged in the first instance on strengthening the perception that the US government stood behind the Agencies rather than strengthening the Agencies actual balance sheets. And with ongoing deterioration in the housing market -- and rising losses on Alt-A mortgages, it seems like the world’s central banks aren’t buying it.

It isn’t all that hard to see why. If one of the Agencies goes bust, no reserve manager wants to be responsible for having lost a ton of money. Not now. Not after the size of some countries’ Agency holdings started to attract a bit of attention. That may force the Treasury’s hand. The spreads the Agencies can afford to pay are too small to compensate reserve managers for the reputational cost -- and political fallout -- they would face should the Agencies ever default.

And I would bet the fact that the US is at odds, politically, with one of the largest holders of Agencies hasn’t helped; the Russia’s central bank was scaling back its Agency portfolio even before the conflict over Ossetia broke out.

Most analysis of the housing boom tends -- in my view -- to leave out the role of Agencies played in channeling central bank funds into the housing market.

That strikes me as a mistake. True, the Agencies weren’t growing their balance sheets that rapidly at the peak of the housing boom. And foreign central banks generally didn’t buy the kind of risky repackaged mortgages that caused the most trouble over. The central banks tended to stick to more traditional kinds of "product."

But in more subtle ways, the rise in central bank reserve growth combined with the United States’ almost unique capacity -- through the Agencies -- to transform a pool of mortgages into an acceptable central bank reserve asset to fuel the boom. Central bank demand for Treasuries and Agencies kept overall interest rates down. Particularly after 2004 -- when the fiscal deficit started to fall and the Japanese stopped intervening -- central banks started to buy a lot of Agency bonds.

Low long-term rates pushed up housing prices, helping to create expectations that home prices would rise and it was safe to buy a home with little down -- and that it was safe to lend money to someone buying homes with little down.

The inverted yield curve made it hard for a lot of vehicles that borrowed short (by issuing short-term paper) and lent long (by buying longer-term securities) to make money. Or to make money without taking on a lot of credit risk. That contributed to the rise in demand for riskier securities. Pension funds that needed a bit of yield to honor their commitments also started taking on more risk.

Central banks bought the safe stuff, holdings its price down. And private actors took the money freed up by the sale of Agency bonds to central banks and bought riskier debt. It all sort of worked, at least for a while.

The irony of course is even though the Agencies ability to tap central bank demand to support the housing market contributed to the current crisis, the Agencies ability to continue to tap central bank demand to support the housing market is also central to the resolution of the current crisis. Or at least central to hopes that the current crisis can be resolved without much larger falls in home prices, much bigger financial losses and the need for the government to step in and backstop the banking system more directly. Right now the Agencies account an ungodly share of new mortgage lending.

I think former Treasury Secretary Larry Summers got this right in the Financial Times a few weeks back. He wrote:

we need the GSEs to be highly active in support of the housing market and financial system in the months ahead. If the authorities can see a path to their being able to play such a role in a framework where it can honestly be said that their borrowing is based on confidence in their financial position rather than primarily on federal guarantees, then this is obviously the preferred alternative. But after what we have seen, such a judgment cannot be based on the GSEs’ own claims, the understandable desire of government officials to maintain confidence and attract private capital, or the fact that they are able to borrow – which only reflects the strength of federally provided credit assurances.

If this preferred alternative is, as I fear, not realistic given the state of GSE finances, the government should use its new receivership power to protect taxpayers and the financial system. In the process, payments to stock holders, holders of preferred stock and probably subordinated debt holders would be wiped out, conserving cash for the benefit of taxpayers. The GSEs’ borrowing costs would fall considerably, helping prospective homeowners.

In this scenario, the government would operate the GSEs as public corporations for several years. They would then be in a position to extend credit where appropriate to support resolution of the current housing crisis. Once the crisis has passed, the federal government would divide their functions into government and private components, the latter of which would be sold off in multiple pieces. The proceeds could be used to fund the low-income housing support activity that was previously mandated to the GSEs. (emphasis added)

This option is obviously not good for the GSE’s current share holders. But it is hard to ask private shareholders to risk large loses in the pursuit of public policy goals. And right now the US government wants the GSEs to lend, not to conserve their capital.

And it certainly seems like the GSE’s creditors aren’t in the mood to extend credit just on the expectation that the GSE’s will be backed by the government if there is a need. The world’s central banks want to the Treasury to show them some money. That means changing the GSE’s current structure.