by Calculated Risk on 1/20/2007 04:56:00 PM

Update: When I first started this blog, I promised to try to call the next recession - as a fun exercise. I didn't think we would have a recession in '05 or '06. Now I think there is a chance, so I'm going out on a limb.

For those interested in recession predictions, here is an excellent paper on why most forecasters miss recessions: The Arcane Art of Predicting Recessions

Since I've just turned bearish on the general economy (update: this is for '07, not next week or Q1 - sorry if that was confusing), I'd like to note what would change my mind.

1) Show me several hundred thousand residential construction jobs lost, and little or no impact on the general economy.

There are many stories like this one from the San Diego Union on the local economy: With real estate sector cooling, employment stays on simmer

Construction firms shed 1,800 workers during December and real estate firms cut 500 positions, according to data released yesterday by the California Employment Development Department.
...
During 2006, construction firms in the county lost a total of 5,000 jobs, more than 5 percent of their work force. Statewide, construction firms have axed 15,300 workers, a 1.7 percent loss.

Click on graph for larger image.

That sounds grim, but in reality very few residential construction jobs have been lost so far. Since residential construction employment peaked in February 2006 at 3,347.4 thousand, only 133.6 thousand jobs have been lost according to the BLS, or about 4%.

Based on the current level of starts, many more jobs will be lost in the coming months (note starts are shifted 6 months into the future on graph, since completions and employment follow starts by about 6 months).

2) Show me mortgage equity withdrawal (MEW) at less than 3% of disposable personal income (DPI), with little or no impact on consumer spending.

From the WSJ: Homeowners Borrow Less Against Equity in Their Homes, Data Show

"Homeowners extracted $113.5 billion ... via mortgage refinancing and other means in the third quarter, the lowest since the fourth quarter of 2003, according to new estimates by a Federal Reserve staffer and former chairman Alan Greenspan

That amount ... was down from $151.8 billion in the second quarter, and the high of $235.9 billion recorded in the third quarter of last year. The latest figure equals 4.7% of households' after-tax income, compared to 10.4% in the third quarter of 2005."

3) Show me near record foreclosure activity, with little or no impact on lenders or the general economy. I'll have more on foreclosure activity this week.

I don't ask for much.

by Calculated Risk on 1/19/2007 03:17:00 PM

From Reuters: First Horizon profits fall, hurt by mortgage unit

First Horizon National Corp. said on Wednesday quarterly earnings fell a steeper-than-expected 28 percent as mortgage banking revenue plunged.
...
First Horizon of Memphis, Tennessee said pretax income from mortgage banking slid 91 percent to $3.5 million, while revenue from mortgage banking fell 27 percent to $111.9 million.

UPDATE from Tanta (lifted from comments):

Lordy, the whole statement's worth a read. Looks like a perfect storm:

1. The carry trade died: "Net interest income decreased 41 percent to $20.4 million in fourth quarter 2006 from $34.7 million in fourth quarter 2005. An inverted yield curve resulted in compression of the spread on the warehouse, which was 1.24 percent in fourth quarter 2006 compared to 2.06 percent for the same period in 2005. Additionally, an 18 percent decrease in the warehouse related to lower origination activity negatively impacted net interest income."

2. Gain on sale dropped: "Net origination income decreased 22 percent to $69.2 million from $89.1 million in fourth quarter 2005 as loans delivered into the secondary market decreased 21 percent to $6.3 billion."

3. Keeping the loans rather than selling them didn't help credit quality: "Provision for loan losses increased to $23.0 million in fourth quarter 2006 from $16.2 million in fourth quarter 2005. This increase primarily reflects continued growth of the commercial and construction loan portfolios and an expectation of slowing economic growth. As a result of this increase, the allowance to loans ratio has increased from 92 basis points in fourth quarter 2005 to 98 basis points in fourth quarter 2006. Nonperforming assets were $139.0 million on December 31, 2006, compared to $79.7 million on December 31, 2005. The nonperforming assets ratio related to the loan portfolio increased to 58 basis points in fourth quarter 2006 from 33 basis points last year. The nonperforming asset ratio continues to migrate from historical low levels due to maturation of the loan portfolio, issues with several commercial credits in the retail commercial bank's traditional lending markets, and deterioration in the residential real estate portfolio reflecting the slow down in the housing market. The net charge-off ratio increased to 25 basis points in fourth quarter 2006 from 22 basis points in 2005 as net charge-offs grew to $13.5 million from $11.0 million during a period of strong loan growth."

I also see they increased their portfolio of residential construction loans by 11%, and also took a $7MM loss, part of which was "the result of employee misrepresentation in our construction lending business."

by Calculated Risk on 1/19/2007 02:12:00 PM

I've recently turned bearish on the U.S. economy. This shouldn't be confused with my longer held negative views on housing.

For housing, I'm very confident that 2007 will be worse than 2006 by every material measure: prices, sales, residential construction employment, starts, MEW, percentage of homeowner equity, and the number of foreclosures.

But the impact of the continuing housing bust on the U.S. economy is far less certain. Although I think a recession is better than a "coin flip" in '07, the odds of a soft landing are still good.

Professor Leamer identified two key missing ingredients for a recession: enough job losses, and a credit crunch (see: Is a Recession Ahead? The Models Say Yes, but the Mind Says No). These are the two issues I wrestled with over the holidays, and I couldn't come to a definitive conclusion.

At the core, recessions are about jobs, and it is easy to imagine scenarios with job growth slowing to 100K per month, maybe even 50K per month. But that isn't a recession.

Click on graph for larger image.

This graph shows the average monthly job growth since 1997. In 2001, the economy lost an average of 147K jobs per month - and that was a fairly mild recession.

Over the last three years, the U.S. has added an average of 164K jobs per month. Right now, based on housing starts, it looks like residential construction employment will fall by 400K to 600K jobs by this Summer. Many of these workers will find new jobs, but it is easy to imagine job growth slowing to 100K per month. Adding in other housing related lost jobs, and some ripple effect, maybe growth will slow to 50K per month. Not enough for a recession.

So how can the U.S. economy slide into recession in '07?

Some possible sources: a credit crunch based on bad loans in the RE sector (and possibly in CRE and C&D too), less consumer spending based on falling MEW, and another downturn in the housing market. If all of these can be avoided, a recession is unlikely.

Right now I don't think these problems can be avoided.

by Calculated Risk on 1/18/2007 08:02:00 PM

Pulte Homes warns on Q4 results.

"Pulte Homes continues to navigate through a challenging operating environment, with demand for new homes during the fourth quarter still far below pre-2006 levels."
Richard J. Dugas, Jr., President and CEO of Pulte Homes, Jan 18, 2007

Pulte had already guided earnings expectations lower. Now they are reducing their Q4 earnings estimates again:

The Company expects fourth quarter results to be in the range of a loss of $.05 to earnings of $.05 per diluted share from continuing operations. Pulte Homes had previously issued earnings guidance in the range of $.30 to $.70 per diluted share, guiding to the lower end of the range in its earnings conference call for the third quarter.

Orders are getting worse:

The Company closed 12,566 homes during the fourth quarter of 2006, a decline of 20% from the fourth quarter of 2005. Net new orders for the quarter were 6,446 homes, a 34% decrease from last year's fourth quarter. For the full year 2006, home closings were 41,487, down 9% compared with 2005. Net new orders for the full year 2006 decreased 29% from the prior year to 33,925 homes.

But the cancellation rate improved slightly:

"... [Pulte] experienced a stabilization to a slight improvement in our fourth quarter cancellation rate compared with the third quarter, as this metric showed progressive improvement throughout the period."

Of course if orders fall to zero, the cancellation rate will decline to zero too. So this metric needs to be kept in context.

Impairments are much worse than expected:

On a preliminary basis, Pulte Homes anticipates that these impairments and land-related charges will be in the range of $330 million to $350 million for the fourth quarter, or $.83 to $.88 per diluted share. The Company previously issued guidance of $150 million for impairments and land-related charges.

Perhaps the only good news is Pulte is reducing their production:

"[Pulte] continue[s] to reduce the number of homes we are starting, as evidenced by our meaningful reduction in speculative units under construction during the quarter."

by Calculated Risk on 1/18/2007 02:39:00 PM

DataQuick reports: California December Home Sales

Last month's sales made for the slowest December since 1996 when 33,591 homes were sold.

Note that 1996 was the last year of the early '90s housing bust in California.

Click on graph for larger image.

A total of 41,100 new and resale houses and condos were sold statewide last month. That's up 4.8 percent from 39,200 for November and down 22.2 percent from a 52,800 for December 2005.

Last month's sales made for the slowest December since 1996 when 33,591 homes were sold.

The median price paid for a home last month was $474,000. That was up 1.1 percent from November's $469,000, and up 3.5 percent from $458,000 for December a year ago.

Although some areas in California are already seeing YoY nominal price declines - like San Mateo, Sonoma, San Diego and Ventura - the median YoY price in California increased slightly.

by Calculated Risk on 1/18/2007 11:06:00 AM

The Census Bureau reports on housing Permits, Starts and Completions. Seasonally adjusted permits rose slightly:

Privately-owned housing units authorized by building permits in December were at a seasonally adjusted annual rate of 1,596,000. This is 5.5 percent above the revised November rate of 1,513,000, but is 24.3 percent below the December 2005 estimate of 2,107,000.

Starts were up for the second straight month:

Privately-owned housing starts in December were at a seasonally adjusted annual rate of 1,642,000. This is 4.5 percent above the revised November estimate of 1,572,000, but is 18.0 percent below the December 2005 rate of 2,002,000.

And Completions were flat, at just below the recent record levels:

Privately-owned housing completions in December were at a seasonally adjusted annual rate of 1,900,000. This is 0.4 percent above the revised November estimate of 1,893,000, but is 2.7 percent below the December 2005 rate of 1,953,000.

Click on graph for larger image.
The first graph shows Starts vs. Completions.

I'm a little surprised at the slight rebound in starts, especially since completions and inventories are still near record levels. Still Starts have fallen "off a cliff", and completions have just started to fall.


This graph shows starts, completions and residential construction employment. (starts are shifted 6 months into the future). Completions and residential construction employment are highly correlated, and Completions lag Starts by about 6 months.

Based on historical correlations, it is reasonable to expect Completions and residential construction employment to follow Starts "off the cliff". This would indicate the loss of 400K to 600K residential construction employment jobs by this Summer.

by Calculated Risk on 1/17/2007 05:51:00 PM

If you missed Tanta's most recent post, it is definitely worth reading (and rereading): Information is Power, Which is Why You Don’t Get Any.

Make sure you read the comments too.

I just hope she wasn't referring to me when she wrote:

"... next time I'll write something calm and polite and professional and bland and get my case jumped by the readers who are tired of calm and polite and professional and bland because that's all you ever get in the newspaper and we come to blogs for some juice."

by Calculated Risk on 1/17/2007 05:14:00 PM

From DataQuick: Bay Area home prices flat, slow sales

Bay Area home prices were flat last month while the sales pace was the slowest pace in a decade ...

A total of 7,488 new and resale houses and condos sold in the Bay Area last month. That was up 3.9 percent from 7,204 in November, and down 19.9 percent from 9,347 for December last year, according to DataQuick Information Systems.

Sales have declined on a year-over-year basis the last 21 months. Last month's sales count was the lowest for any December since 1996 when 7,180 homes were sold. The average for all Decembers since 1988 is 8,339.
...
The median price paid for a Bay Area home was $612,000 in December. That was down 0.6 percent from $616,000 in November and up 0.5 percent from $609,000 for December a year ago. The median peaked last June at $644,000.
...
Indicators of market distress are still in the normal range. Financing with adjustable-rate mortgages is flat. Foreclosure activity is rising but is still in the normal range. Down payment sizes are stable. Flipping rates and non-owner occupied buying activity are down, DataQuick reported.

Just six months ago, DataQuick reported "indicators of market distress are still largely absent" and "foreclosure rates are coming up from last year's low point, but are still below normal levels". Now foreclosures are in the "normal range", and from the information I'm receiving, foreclosures will be significantly above normal shortly.

by Calculated Risk on 1/17/2007 01:12:00 PM

From Fed Governor Frederic Mishkin: The Role of House Prices in Formulating Monetary Policy

Once again the Fed is arguing against taking action when a possible bubble is forming. This is the same argument Greenspan made during the stock market bubble, but this time the issue is housing. However, the Fed stands ready to take action if the bursting of the bubble impacts the general economy.

Mishkin's conclusion:

Large run-ups in prices of assets such as houses present serious challenges to central bankers. I have argued that central banks should not give a special role to house prices in the conduct of monetary policy but should respond to them only to the extent that they have foreseeable effects on inflation and employment. Nevertheless, central banks can take measures to prepare for possible sharp reversals in the prices of homes or other assets to ensure that they will not do serious harm to the economy.

by Calculated Risk on 1/17/2007 12:59:00 PM

From NAHB: Builder Confidence Improves in January

Click on graph for larger image.

Excerpts:

The HMI increased from an upwardly revised 33 in December to 35 in January, its highest level since July of 2006.
...
Two out of three component indexes registered improvement in January. The index gauging current single-family home sales and the index gauging traffic of prospective buyers each gained three points, to 36 and 26 respectively, while the index gauging sales expectations for the next six months remained unchanged at 49.

Meanwhile, three out of four regions surveyed in the HMI posted gains in January. Two-point gains were registered in the Northeast, Midwest and South, to 39, 24 and 41, respectively. The HMI for the West remained unchanged from the previous month at 32.