by Calculated Risk on 2/25/2016 03:34:00 PM

From housing economist Tom Lawler:

Yesterday the National Association of Realtors estimated that US existing home sales ran at a seasonally adjusted annual rate of 5.47 million in January, up 0.4% from December’s downwardly-revised (to 5.45 million from 5.46 million pace, and up 11.0% from January’s upwardly-revised (to 4.93 million from 4.82 million) seasonally adjusted base. The NAR’s unadjusted sales estimate for January was up 7.5% from a year ago. The January release incorporated annual seasonal adjustment revisions, which, as shown in the table below, worked to increase adjusted sales in the early part of the year (especially January and February), and decrease adjusted sales from late Spring through earl Fall.

Seasonally Adjusted Existing Home Sales in 2015, Old vs. New Seasonal Factors (Annual Rate, 000's)
	Old Seasonal Factors	New Seasonal Factors	% Change
Jan	4,820	4,930	2.3%
Feb	4,890	4,970	1.6%
Mar	5,210	5,250	0.8%
Apr	5,090	5,140	1.0%
May	5,320	5,290	-0.6%
Jun	5,480	5,410	-1.3%
Jul	5,580	5,480	-1.8%
Aug	5,300	5,290	-0.2%
Sept	5,550	5,440	-2.0%
Oct	5,320	5,290	-0.6%
Nov	4,760	4,860	2.1%
Dec*	5,440	5,450	0.2%
*Note: December’s preliminary sales pace using old seasonal factors was 5.46 million. However, the NAR revised unadjusted sales in December down to 436,000 from 438,000. The number in the table above uses the old seasonal factor but the revised unadjusted sales number.

The NAR’s seasonally-adjusted sales estimate for January was above consensus and slightly higher than my projection, but that was solely attributable to the larger-than-expected downward revision in January seasonal factors.

The NAR also estimated that the inventory of existing homes for sale at the end of January was 1.820 million, up 3.4% from December’s downwardly-revised (to 1.76 million from 1.79 million) level and down 2.2% from last January. The NAR’s estimate was somewhat higher than my projection based on local realtor/MLS reports available about a week ago, though reports released suggest that the NAR’s estimate is “reasonable.” Finally, the NAR estimated that the median existing SF home sales price last month was $215,000, up 8.2% from last January. This YOY increase was slightly higher than my projection.

CR Note: Tom Lawler also sent me the table below of short sales, foreclosures and all cash sales for a several selected cities in January.

On distressed: Total "distressed" share is down in all of these markets.

The All Cash Share (last two columns) is mostly declining year-over-year. As investors pull back, the share of all cash buyers will probably continue to decline.

	Short Sales Share		Foreclosure Sales Share		Total "Distressed" Share		All Cash Share
	Jan- 2015	Jan- 2014	Jan- 2015	Jan- 2014	Jan- 2015	Jan- 2014	Jan- 2015	Jan- 2014
Las Vegas	7.0%	9.7%	7.9%	9.4%	14.9%	19.1%	31.1%	36.0%
Reno**	4.0%	10.0%	6.0%	6.0%	10.0%	16.0%
Phoenix	2.3%	6.3%	3.9%	6.8%	6.2%	13.1%	27.4%	32.0%
Sacramento	4.8%	6.9%	4.2%	9.5%	9.0%	16.4%	19.8%	22.7%
Minneapolis	2.9%	4.3%	11.0%	16.0%	13.9%	20.3%
Mid-Atlantic	4.3%	5.8%	13.9%	15.2%	18.2%	21.0%	19.5%	21.4%
Riverside	3.1%	3.3%	4.2%	5.8%	7.3%	9.1%	18.2%	19.9%
San Bernardino	1.9%	4.3%	4.5%	6.1%	6.4%	10.4%	19.2%	21.9%
Bay Area CA*	2.5%	4.2%	2.5%	4.2%	5.0%	8.4%	20.5%	24.3%
So. California*	3.5%	5.7%	4.5%	5.7%	8.0%	11.4%	22.4%	26.7%
Florida SF	3.5%	5.3%	15.3%	25.7%	18.7%	31.0%	36.8%	43.4%
Florida C/TH	2.2%	3.2%	12.1%	20.6%	14.4%	23.9%	61.9%	69.5%
Miami MSA SF	5.7%	8.9%	16.8%	23.2%	22.5%	32.1%	36.3%	42.4%
Miami MSA C/TH	2.7%	3.7%	14.7%	23.5%	17.4%	27.2%	65.6%	70.9%
Chicago (city)					22.8%	24.1%
Rhode Island					11.8%	16.1%
Northeast Florida					24.8%	38.1%
Spokane					15.6%	24.2%
Spokane					15.6%	24.2%
Toledo							34.7%	37.6%
Tucson							29.7%	34.8%
Knoxville							26.3%	28.5%
Peoria							22.3%	24.6%
Georgia***							24.8%	31.3%
Omaha							19.5%	22.2%
Pensacola							32.9%	38.3%
Richmond VA			13.5%	18.3%			21.9%	22.8%
Memphis			15.6%	14.8%
Springfield IL**			16.8%	16.6%
*share of existing home sales, based on property records **Single Family Only ***GAMLS

by Calculated Risk on 2/25/2016 12:38:00 PM

Update on August 16, 2018: A recent article by Ryan Cooper referenced this post. Unfortunately he misunderstood what I wrote. I argued that most of the decline in the overall participation rate since the great recession was due to demographics and long term trends - and was not cyclical, and I've suggested using the prime working age ("25 to 54" years old) participation rate to track the economy (not perfect, but better than the overall rate). Strangely Mr. Cooper used the "25 to 54" years old rate to argue I was wrong.  Time has proven me correct, and Mr. Cooper should correct his mistake.

Please see this post (with a spreadsheet) for an explanation.

Original post:

Yesterday I was a guest on Bloomberg's What'd You Miss? talking about housing. One of the earlier guests was Professor J.W. Mason discussing the labor force participation rate (LFPR).

Professor Mason argued that most of the decline in the LFPR is cyclical.

I've written extensively about the LFPR, and I've argued that most of the decline is due to ongoing trends and demographics.

Dr. Mason said that if you looked at each age group by sex, you could construct an "adjusted" LFPR. He presented the following graph:

Click on graph for larger image.

NOTE: I believe this approach is incorrect.

This graph from Bloomberg shows the BLS reported LFPR and an "adjusted" LFPR using the 1999 LFPR for each age group and gender - and the current population for each group.

As an example, if we look at men in the 40 to 44 age group (these are my calculations using Dr. Mason's method):

• 1999 Labor Force Participation Rate, Men, 40-44: 92.2%
• Jan 2016 LFPR: 90.4%
• Current Population 40-44 Men: 9,572,000
• Current Participation: 8,652,000
• Adjusted Participation (using 1999 LFPR): 8,829,000
• "Missing" Male workers in the 40-44 age group: 177,000

This is the approach that Dr. Mason used to construct the above graph - and this is incomplete and I believe his conclusion is incorrect.

Here is a look at the trend for 40 to 44 year old men (BLS data, only available Not Seasonally Adjusted since 1976).

Click on graph for larger image.

This graph shows the 40 to 44 year old men participation rate since 1976 (note the scale doesn't start at zero to better show the change).

There is a clear downward trend, and a researcher looking at this trend in the year 2000 might have predicted the 40 to 44 year old men participation rate would be about the level as today (see trend line).

Clearly there are other factors than "economic weakness" causing this downward trend.   I listed some reasons a couple of years ago, and newer research from Pew Research suggests stay-at-home dads is one of the reasons: Growing Number of Dads Home with the Kids

Just looking at this graph, I don't think there are many 40 to 44 year old men that will be returning to the labor force - and a careful analysis of each age group by sex shows few "missing workers".

If Dr. Mason had used this trend line for 40 to 44 year old men - instead of the 1999 participation rate - the number of missing workers would be close to zero.

The bottom line is most of the decline in the Labor Force Participation rate is due to ongoing trends (as above) and demographics - and is not cyclical.

by Calculated Risk on 2/25/2016 11:00:00 AM

From the Kansas City Fed: Tenth District Manufacturing Activity Declined Further

The Federal Reserve Bank of Kansas City released the February Manufacturing Survey today. According to Chad Wilkerson, vice president and economist at the Federal Reserve Bank of Kansas City, the survey revealed that Tenth District manufacturing activity declined further.

“Factories reported a slightly larger decline in February than in previous months,” said Wilkerson. “Energy-related firms generally had a negative outlook, but firms overall remained slightly optimistic about future factory activity.”
...
Tenth District manufacturing activity declined further in February, while producers’ expectations for future activity remained slightly positive. Price indexes were mixed, but most remained in negative territory.

The month-over-month composite index was -12 in February, its lowest level since 2009, down from -9 in January and December. ... The employment index dropped from -15 to -26, its lowest level in nearly six years. The new orders for exports index fell from -4 to -13, while the capital expenditures index remained stable but weak.
emphasis added

The Kansas City region continues to be hit hard by lower oil prices and the stronger dollar.

by Calculated Risk on 2/25/2016 08:34:00 AM

The DOL reported:

In the week ending February 20, the advance figure for seasonally adjusted initial claims was 272,000, an increase of 10,000 from the previous week's unrevised level of 262,000. The 4-week moving average was 272,000, a decrease of 1,250 from the previous week's unrevised average of 273,250.

There were no special factors impacting this week's initial claims.

The previous week was unrevised.

The following graph shows the 4-week moving average of weekly claims since 1971.

Click on graph for larger image.


The dashed line on the graph is the current 4-week average. The four-week average of weekly unemployment claims decreased to 272,000.

This was close to the consensus forecast of 270,000. The low level of the 4-week average suggests few layoffs.

by Calculated Risk on 2/24/2016 09:06:00 PM

Thursday:
• At 8:30 AM ET, The initial weekly unemployment claims report will be released.  The consensus is for 270 thousand initial claims, up from 262 thousand the previous week.

• Also at 8:30 AM, Durable Goods Orders for January from the Census Bureau. The consensus is for a 2.0% increase in durable goods orders.

• At 9:00 AM, FHFA House Price Index for December 2015. This was originally a GSE only repeat sales, however there is also an expanded index.  The consensus is for a 0.5% month-to-month increase for this index.

• At 11:00 AM, the Kansas City Fed manufacturing survey for February.

by Calculated Risk on 2/24/2016 03:55:00 PM

The automakers will report February vehicle sales on Tuesday, March 1st.

Note:  There were 24 selling days in February, unchanged from 24 in February 2015.

From WardsAuto: Forecast: February Sales Set to Reach 15-Year High

A WardsAuto forecast calls for U.S. automakers to deliver 1.34 million light vehicles in February, a 15-year high for the month. The forecasted daily sales rate of 55,868 over 24 days represents a 7.1% improvement from like-2015 (also 24 days).

The report puts the seasonally adjusted annual rate of sales for the month at 17.50 million units, compared with year-ago’s 16.32 million and January’s 17.45 million mark.
emphasis added

And from J.D. Power: New-Vehicle Sales in February Expected to Increase 8.1%

New-vehicle sales in February 2016 are expected to increase 8.1% from a year ago, according to a monthly sales forecast developed jointly by J.D. Power and LMC Automotive.

New-vehicle retail sales in February are projected to reach 1,046,700 units from 968,316 in February 2015, with the same number of selling days (24 days).
...
The SAAR for total sales is projected to reach 17.7 million units in February 2016, up 1.4 million units from 16.4 million a year ago and the highest rate since 2000 (18.9 million).

Looks like another strong month for car sales.

by Calculated Risk on 2/24/2016 01:48:00 PM

The new home sales report for January was below expectations at 494,000 on a seasonally adjusted annual rate basis (SAAR), however combined sales for October, November and December were revised up.

Sales were down 5.2% year-over-year (YoY) compared to January 2015.   However, we have to remember January 2015 was a pretty strong month at 521,000  SAAR.  Sales for all of 2015 were 501,000 (up 14.5% from 2014) - and since January (and February) were especially strong months last year, the YoY comparison is difficult.

Earlier: New Home Sales decreased to 494,000 Annual Rate in January.


Click on graph for larger image.

This graph shows new home sales for 2015 and 2016 by month (Seasonally Adjusted Annual Rate).

The comparisons in early 2016 are difficult.  And I also expect lower growth this year.

Houston (and other oil producing areas) will have a problem this year. Inventory of existing homes is increasing quickly and prices will probably decline in those areas. And that means new home construction will slow in those areas too.

And here is another update to the "distressing gap" graph that I first started posting a number of years ago to show the emerging gap caused by distressed sales.  Now I'm looking for the gap to close over the next few years.

The "distressing gap" graph shows existing home sales (left axis) and new home sales (right axis) through January 2016. This graph starts in 1994, but the relationship had been fairly steady back to the '60s.

Following the housing bubble and bust, the "distressing gap" appeared mostly because of distressed sales.

I expect existing home sales to move more sideways, and I expect this gap to slowly close, mostly from an increase in new home sales.

However, this assumes that the builders will offer some smaller, less expensive homes.

Note: Existing home sales are counted when transactions are closed, and new home sales are counted when contracts are signed. So the timing of sales is different.

by Calculated Risk on 2/24/2016 10:13:00 AM

The Census Bureau reports New Home Sales in January were at a seasonally adjusted annual rate (SAAR) of 494 thousand.

The previous three months were revised up by a total of 28 thousand (SAAR).

"Sales of new single-family houses in January 2016 were at a seasonally adjusted annual rate of 494,000, according to estimates released jointly today by the U.S. Census Bureau and the Department of Housing and Urban Development. This is 9.2 percent below the revised December rate of 544,000 and is 5.2 percent below the January 2015 estimate of 521,000."
emphasis added

Click on graph for larger image.

The first graph shows New Home Sales vs. recessions since 1963. The dashed line is the current sales rate.

Even with the increase in sales since the bottom, new home sales are still fairly low historically.

The second graph shows New Home Months of Supply.

The months of supply increased in January to 5.8 months.

The all time record was 12.1 months of supply in January 2009.

This is now in the normal range (less than 6 months supply is normal).

"The seasonally adjusted estimate of new houses for sale at the end of January was 238,000. This represents a supply of 5.8 months at the current sales rate."

On inventory, according to the Census Bureau:

"A house is considered for sale when a permit to build has been issued in permit-issuing places or work has begun on the footings or foundation in nonpermit areas and a sales contract has not been signed nor a deposit accepted."

Starting in 1973 the Census Bureau broke this down into three categories: Not Started, Under Construction, and Completed.

The third graph shows the three categories of inventory starting in 1973.

The inventory of completed homes for sale is still low, and the combined total of completed and under construction is also low.

The last graph shows sales NSA (monthly sales, not seasonally adjusted annual rate).

In January 2016 (red column), 37 thousand new homes were sold (NSA). Last year 39 thousand homes were sold in January.

The all time high for January was 92 thousand in 2005, and the all time low for January was 21 thousand in 2011.

This was below expectations of 520,000 sales SAAR in January, however prior months were revised up slightly - although sales were down year-over-year.  Still a decent report.  I'll have more later today.

by Calculated Risk on 2/24/2016 07:00:00 AM

From the MBA: Mortgage Applications Decrease in Latest MBA Weekly Survey

Mortgage applications decreased 4.3 percent from one week earlier, according to data from the Mortgage Bankers Association’s (MBA) Weekly Mortgage Applications Survey for the week ending February 19, 2016. This week’s results include an adjustment to account for the President’s Day holiday.
...
The Refinance Index decreased 8 percent from the previous week. The seasonally adjusted Purchase Index increased 2 percent from one week earlier. The unadjusted Purchase Index decreased 4 percent compared with the previous week and was 27 percent higher than the same week one year ago.
...
The average contract interest rate for 30-year fixed-rate mortgages with conforming loan balances ($417,000 or less) increased to 3.85 percent from 3.83 percent, with points increasing to 0.42 from 0.36 (including the origination fee) for 80 percent loan-to-value ratio (LTV) loans.
emphasis added

Click on graph for larger image.


The first graph shows the refinance index since 1990.

Refinance activity was higher in 2015 than in 2014, but it was still the third lowest year since 2000.

Refinance activity has picked up recently as rates have declined.


The second graph shows the MBA mortgage purchase index.  

According to the MBA, the unadjusted purchase index is 27% higher than a year ago.

by Calculated Risk on 2/23/2016 06:39:00 PM

Here is the earlier post on Case-Shiller: Case-Shiller: National House Price Index increased 5.4% year-over-year in December

The year-over-year increase in prices is mostly moving sideways now around 5%. In December 2015, the index was up 5.4% YoY.

In the earlier post, I graphed nominal house prices, but it is also important to look at prices in real terms (inflation adjusted).  Case-Shiller, CoreLogic and others report nominal house prices.  As an example, if a house price was $200,000 in January 2000, the price would be close to $274,000 today adjusted for inflation (37%).  That is why the second graph below is important - this shows "real" prices (adjusted for inflation).

It has been almost ten years since the bubble peak.  In the Case-Shiller release this morning, the National Index was reported as being 3.7% below the bubble peak.   However, in real terms, the National index is still about 18% below the bubble peak.

Nominal House Prices


The first graph shows the monthly Case-Shiller National Index SA, the monthly Case-Shiller Composite 20 SA, and the CoreLogic House Price Indexes (through December) in nominal terms as reported.

In nominal terms, the Case-Shiller National index (SA) is back to October 2005 levels, and the Case-Shiller Composite 20 Index (SA) is back to April 2005 levels, and the CoreLogic index (NSA) is back to June 2005.

Real House Prices

The second graph shows the same three indexes in real terms (adjusted for inflation using CPI less Shelter). Note: some people use other inflation measures to adjust for real prices.

In real terms, the National index is back to December 2003 levels, the Composite 20 index is back to September 2003, and the CoreLogic index back to December 2004.

In real terms, house prices are back to 2003 levels.

Note: CPI less Shelter is down 0.6% year-over-year, so this has been pushing up real prices recently.

Price-to-Rent

In October 2004, Fed economist John Krainer and researcher Chishen Wei wrote a Fed letter on price to rent ratios: House Prices and Fundamental Value. Kainer and Wei presented a price-to-rent ratio using the OFHEO house price index and the Owners' Equivalent Rent (OER) from the BLS.

Here is a similar graph using the Case-Shiller National, Composite 20 and CoreLogic House Price Indexes.

This graph shows the price to rent ratio (January 1998 = 1.0).

On a price-to-rent basis, the Case-Shiller National index is back to August 2003 levels, the Composite 20 index is back to April 2003 levels, and the CoreLogic index is back to July 2003.

In real terms, and as a price-to-rent ratio, prices are back to 2003 levels - and the price-to-rent ratio maybe moving a little sideways now.