« Yes, There Is No [Social Security] Crisis, But... | Main | Chad Orzel Is Wrong! »

January 19, 2005

Nouriel Roubini Preprocesses the Wall Street Journal

Nouriel writes:

Nouriel Roubini's Global Economics Blog: January 2005 Archives: While Greg Ip wrote a very fine front page analysis of the risks of a dollar crash in the WSJ today (and thanks for the citations of my work with Brad Setser on financial crises), an even finer article by him on the risks of a bursting of housing bubbles made it only to page 2 of the paper (I guess the informal rule is that you cannot have more than one article per journalist in the front page). And his timely analysis was matched by another good WSJ piece by Agnes Crane in the Credit Market section on how the risk of a bursting of the housing bubble may affect US monetary policy (i.e. whether Greenspan & Co. may not tighten as much as is needed if there are risks to the housing markets from higher interest rates).

A lot - academic, policy and market-wise - has been written recently on housing bubbles in the US and abroad, the risk of a housing bubble bursting and the relation between asset bubbles and monetary policy (see my web site pages on housing and asset bubbles and the Fed). So, not too much original may be said. A missing link between the good Ip dollar piece and the good Ip housing piece is that of the effects of a dollar crash on US bond markets and thus on housing markets. The link is implicit but worth fleshing out: i.e. a dollar crash/hard landing that would be associated with a bond market rout (see my previous blog) would have severe consequences on all other risky and overvalued assets, including housing values.

In that scenario, the Fed would be in a ugly trade-off: to stave off a free fall of the dollar with its inflationary consequences it would have to sharply tighten monetary policy; but such tightening would be recessionary, it would also exacerbate the increase in short and long term interest rates and the fall in housing prices.

In that regard, one issue missed by the Crane piece is the following one: in the usual comparison between the US and UK/Australia, the standatd argument is that in the US a larger fraction of mortgages are at fixed rates (compared to UK/Australia) and that the ability of households to refinance and move from adjustable rates to fixed rates (at a hint of expected increased in long rates) may reduce the impact of short term interest rate tightening by the Fed on the housing market compared to UK/Australia where such tightening may be restrained by concerns about a collapse in housing prices. I.e. the Fed - compared to the Bank of England - should be less concerned about the effects of its overdue tightening on the housing market.

This argument is only partially - and comparatively - true. If households are not taking on the risk of rising long rates as they can refinance, someone else is gonna take such risk and the ability of banks, mortgage finance companies or, down the chain the large GSE's, to hedge such risk is limited. The hot potato of higher short-rates leading to higher long rates and losses deriving from households not taking the long rate increase risk will be borne by some other agents in the financial markets, banks/mortgage companies and/or GSE's. The general equilibrium effects of derivative hedging of the yield curve rising are not clear but, as it well know, the risk of a systemic crisis event are rising in such circumstances...

Posted by DeLong at January 19, 2005 08:23 AM


S & L - style bailouts perhaps?

Posted by: theCoach at January 19, 2005 08:43 AM

It seems unlikely to me that there will be a real estate crash. Negative price corrections due to rising interest rates, certainly, perhaps over several years. But corrections in home prices in the negative double digits? Only in certain overheated markets.

I'm not sure what the impact would be to issuers of fixed rate mortgages. Most homeowners will probably just sit tight. The default rate might go up a bit. Consumer spending will decrease from the negative wealth effect.

But I don't really see a meltdown.

Posted by: steve at January 19, 2005 09:35 AM

having just bought a condo (two weeks ago) in the bubble of bubbles (San Francisco) this has been a major, major concern of mine. I've seen a number of articles that show the proportion of adjustable rate mortages in SF has been over 50% for at least 18 months.

I have not been able to get data on the type of adjustable rate periods these have been for. 3 years, 5 years or 7 year fixed prior to going adjustable? If a majority have 5 or 7 year fixed, is this enough time to ride out a correction?

I think part of the run up in property prices has been a structural change in the availability of different financing. 5 or 10% down is the standard now and the availability of these adjustable rate financings has changed the mind set of the buyer.

When I was shopping for rates, I had to consider the length of time I'm planning to stay and balance that with any risk of a correction. Given me and my wife's age and family status, I'm putting my bet on the fact that we will want to move out of a condo in the city in the next 5 to 10 years. Why pay the extra interest for a 30 yr. fixed if we are not planning to stay?

On the other hand, if there is a sharp rise in interest rates, I want to be protected.

In the end we went for a 7 yr fixed period with adjustable rates after that. The loan also has 10% cap. If things start to look really bad, I'm sure I enough time to refi into a 30 yr fixed well before the market gets to 10%.

And why, do you ask, am I buying?

Well, tax breaks for one. We don't own now and can use the breaks. Also, sometimes you just get to that point in life where you want a place of your own.

Posted by: I can't belive at January 19, 2005 09:40 AM

We don't have a housing bubble, we have a permit bubble. Land costs are reasonable if the land does not have a permit to build attached to it. It's even more reasonable if the land does not have a federal, state, or local government agricultural subsidy attached to it.

The interest rate/inflation risk has been transferred from the banks to the public in the form of US government guarantees of pension funds (you know, like social security) that have invested in bonds, in bond and money market retirement accounts held by the general public, and in overseas holdings of treasury debt.

Posted by: walter willis at January 19, 2005 10:57 AM

walter willis wrote, "We don't have a housing bubble, we have a permit bubble. Land costs are reasonable if the land does not have a permit to build attached to it."

So what? Suppose the land comes with a provision that *nothing* may be done with it. Then (aside from natural conservation) the land is worthless anyway. That it would also command a low price means nothing.

In fact, you actually have it precisely backwards. Excessively strong zoning restrictions act to make land scarcer, and drive up the resulting demand.

Posted by: liberal at January 19, 2005 01:17 PM

On the fixed rate/variable rate question, it has always seemed obvious to me that this is just shifting, not reducing, the loss from sharp interest rate rises. Putting the risk into the financier's court, rather than the consumers', just means the contractionary mechanism is different - not necessarily less at all.

What has happened to interest rates on new fixed rate mortgages in the US? They ought to be rising sharply for the same reasons long bond rates should be. Do we have a similar puzzle here to that for long bonds?

Posted by: derrida derider at January 19, 2005 05:02 PM

I don't think the US has ever experienced a situation where a significant proportion of houseowners are "in negative equity", i.e. the value of their house (or condo) is less than their mortgage balance. Owners in such a situation cannot sell, since to do so they would have to come up with the difference between what the house would fetch and what they owe. They can't refinance either, since they can't borrow enough on the new mortgage to pay off the old. They therefore have to pay whatever their current mortgage holder demands--and on an adjustable mortgage, such demands may become exorbitant--or go into foreclosure.

People who took 90%, 95%, 100% loan-to-value mortgages are most at risk to end up in negative equity on a fall in house prices. Those whose mortgages were also adjustable rate will be the ones to get worst squeezed. Does anyone know how many such (high ltv, adjustable rate) mortgages are out there?

Posted by: jam at January 19, 2005 05:09 PM

"I don't think the US has ever experienced a situation where a significant proportion of houseowners are "in negative equity", i.e. the value of their house (or condo) is less than their mortgage balance."

Can this be true? Even in the 1930s?

Posted by: otto at January 19, 2005 06:42 PM

I don't think it is a bubble in the strict sense of the term, which requires people buy with a plan to sell to the greater fool.

Rather, my guess is that home prices have responded forcefully to the mortgage market rally because that rally has not been associated with any serious deterioration of trend income growth in the household sector. Rather, this time around, the mortgage rally has been driven mostly by developments in the corporate and foreign sectors. Specifically, the corporate sector has moved sharply toward financial surplus and the foreign sector has recycled its surplus savings into the US debt market. The resulting decline of interest rates should have delivered a uniquely potent stimulus to the real estate market -- and it has. We do not have to introduce the notion of a bubble to get this result.

Of course, when these special factors reverse, home prices will probably go down -- and by more than normal. Another way to put this is that the duration of the housing stock is longer than normal in this cycle. But this does not mean that housing is a bubble. The dynamics here do not require that anyone -- except perhaps Asian central banks -- behave irrationally.

Posted by: Gerard MacDonell at January 19, 2005 06:48 PM


I don't know. Brad's the economic historian here. But I don't think so. The modern mortgage, long-term, used to finance home purchase the cost of which is amortized over its life, dates, I understand, from 1934 and Title II of the National Housing Act. Prior to that, mortgages were short term (3-5 years) and for much less than the property was worth. Think of them as the real estate equivalent of pawning chattels.

So in the '30s, there weren't very many mortgages used to buy houses with and the worst of the house price falls had already occurred by the time they were issued.

Posted by: jam at January 19, 2005 07:42 PM

[comment spam]

Posted by: at February 9, 2005 11:33 PM