Nouriel Roubini writes a thoughtful piece on the Citibank super-SIV.

“Such banks played a reckless game of regulatory arbitrage by creating risky off-balance sheet SIVs, loaded with risky assets and funded with the most short term asset backed CP in order to avoid the Basel capital charges for similar on balance sheet assets. The whole point of bank capital regulation is that banks that get the lender of last resort support of the central banks need to have enough capital to avoid the gamble for redemption games of playing at a casino with the money of depositors. But banks first avoided those capital charges by creating the off-balance sheet SIVs with lower capital charges and then, when the roll-off of the liabilities of such SIVs occurred amply relied on the Fed’s lender of last resort lending – and on explicit Fed bending of strict rules on how much the banks could re-lend to their affiliated and SIVs – to avoid the losses that they would have incurred by their reckless creation of illiquid SIVs (as well as accepting for repo operations assets whose true quality is dubious as the Fed has not clearly explained what assets it now accepts in its repo and discount window operations). This is moral hazard of the first order: avoid capital regulations via off balance sheet dangerous schemes characterized by serious maturity mismatches, high liquidity risk and gambing for redemption by investing in toxic waste securities; and then get free lender of last resort support when the liquidity roll-off occurs.”

The Fed stepped in to support the share markets. Even though they are basically only flat for the year, the Wall Street boyz felt that they needed a bailout and so they got one. As I mentioned before, the timing of the Fed move and its surprise were chosen for effect, which admits of only one conclusion as to the intent of the move.

The Fed apparently went on to hold a conference call to urge the major banks to use the discount window, freshly rolled down, to borrow money and then to lend same to the mortgage industry. If there was any doubt about the Fed’s desire to reflate the bubble, set it aside. The Fed clearly understands the problem, that they can print all the money they want, but the credit industry must put it to work.

Unfortunately, burst bubbles don’t reflate easily. Nouriel Roubini puts it well: “The economic slowdown is already underway and given the glut of housing, autos and durable goods in the economy the demand for these goods will be relatively insensitive to interest rates.”

People liken this to the 1998 LTCM crisis, but it is very different. In 1998, a major player, LTCM, “got it wrong” in the financial markets and a derivatives crisis was narrowly avoided by the Fed’s intervention. It is different in 2007. First of all, the root of the problems is economic, not financial. Too much money has been badly invested, most visibly in residential housing, so that there is a glut. More than people can use or occupy, let alone afford. As a result, prices will decline and leverage will implode. The Fed may be able to slow this process, but it is very doubtful that it can restart the upward trajectory of debt that is needed to sustain the bubble economy.

Anyhow, enough whining. The Fed is now sitting at the table, helping the bulls with extra cards from time to time. It will be a rough ride for the bears, because the Fed can and will make life difficult for us. As Marty Zweig famously said “Don’t fight the Fed”. Well Marty, sound advice, but when the Fed is playing stupid, someone has to do it. The good news is that the Fed will keep the bear brigades thinned out, which will make it more profitable to tough it out. This is the end of a Ponzi finance bubble. It isn’t “subprime”. It isn’t just housing. It is a debt mountain four times the size of GDP, augmented by a derivatives tower so high no-one knows how tall it it.

Nouriel Roubini asks: “Are We at The Peak of a Minsky Credit Cycle?” Well, if you had checked here first, Nouriel, you wouldn’t need to ask rhetorical questions, would you?

And by the way, another hedge fund went casters up to the tune of $1.5 billion in losses. That’s deleveraging in action. The WSJ provides a copy of their “Oops, sorry ’bout that” letter.

Q1 2007 GDP was revised lower this morning, down to 0.6%. Economic bulls and the Fed continue to forecast an abrupt change in this trend.

On the other hand, Nouriel Roubini writes: “This writer has been a serious bear on housing for a long time: but after listening to these most sophisticated analysts of housing, mortgage lending and the MBS markets from a top global financial firm my concerns seemed almost not bearish enough. The main message from these analysts and the data is that the housing recession, the subprime carnage and the broader mortgage mess are getting worse, not better; and things will get worse well into 2008. There is no end in sight to the housing recession and we are only in the first innings of the mortgage credit crunch.

As for private consumption (that is about 70% of GDP) its prospects are now weak and cast a serious shadow for US economic growth in the quarters ahead.”

Nouriel, that’s right. You are not nearly bearish enough. That’s what being an economist will do for you. Economics isn’t the dismal science anymore, now it’s the science of cheerleading market and economic manipulation. Pump ‘em up, pump ‘em up, pump ‘em up, waaaay up! So far the game is Keynesians 1, Austrians 0. But it is far from over.

Widely followed economic prognosticator ECRI (”Economic Cycle Research Institute”) has a good long-term record. They provide a variety of leading indicators, and are currently forecasting a resurgence in economic growth, as well as calling a bottom in housing prices. It is well known that the Greenspan Fed relied heavily on ECRI’s forecasts, and one suspects that the Bernanke Fed does too. These forecasts may, in fact, be the basis of Bernanke’s rosy public statements that growth will continue, unaffected by the “softness” in the housing market.

Other forecasters, such as Paul Kasriel and Nouriel Roubini, are far less sanguine and much more aligned with my views (that the barbarians are at the gate). Is ECRI right? I will be amazed if they are.

Given their long-term record, how dare I doubt them?

First of all, their forecasting is, as I understand it, model-based. Like a trading system, they plug a bunch of data into their model and it spits out the forecast numbers based on historical relationships. All trading systems break down with time because the statistical relationships are non-stationary. Of course, ECRI knows this and doubtless has provision to re-optimize or re-train their model over time. But does this process work fast enough?

Secondly, a lot of data out there is visibly distorted. A prime example is the weekly Mortgage Bankers Association applications report (which I believe is input to the ECRI model). This data, which has tracked home sales accurately in the past, has stopped doing so. A paper by David Berson, Fannie Mae’s chief economist, asks:

“Why has the relationship between purchase applications and home sales weakened recently? One likely explanation is that the stricter guidance of depository institution regulators over the past year with respect to mortgage loans has made it more difficult for some households to qualify for a loan. As a result, those households have had to make multiple applications in order to get a mortgage loan — thereby pushing up purchase applications without increasing home sales.”

Another example is the NAR’s reporting. Looking at things through a bubble is a good metaphor. Does the ECRI model take this distortion into account? I doubt it.

Thirdly, I suspect we are “out of range” on many input variables, such as debt and savings-related data (We haven’t previously had such negative savings), or even the change in house prices (which have not in the past shown y-o-y declines). In other words, when the model was built, back-tested and run out-of-sample, each input variable had a range over which it varied. The presumption is that relationships established under those circumstances will hold even if the variables move out of their historical ranges. Maybe, but it is the same as a pilot who ignores the published limitations of his aircraft - he has just become a test pilot. Extrapolation can be fatal.

We’ll see.