I have got to imagine the bank owned volume is playing havoc with the matched pairs methodology. Imagine if they finally wise up and start including those.
just like consumer confidence, i would really like to know who they are not the people i talk to in the stores. but different in all areas,regions states,counties and cities. no wonder this unwinding is taking soooooooo long.
The only question being why we're watching two people in a closet staring at a big, unblinking glass eye, repeating whatever is fed to them in little earpieces.
And those earpieces are talking the book of their masters, the conglomerate GE.
How bizarre! Maybe I'll switch to Boomberg, it must be different.
/snip
June 30 (Bloomberg) -- American International Group Inc., the insurer bailed out by the U.S., said that valuation declines on credit-default swaps sold to European banks could have a “material adverse effect” on the company’s results.
The risk of losses on the derivatives may last “longer than anticipated,” the New York-based insurer said late yesterday in a regulatory filing updating the “risk factors” in its 2008 annual report. The firm had $192.6 billion in swaps allowing lenders to reduce the funds they had to hold in reserve as of March 31, AIG said.
/endsnip
This however is the creamy, fudge-colored middle of the crap sandwich:
/snip
AIG said it was unable to provide full details on the value of assets backed by the swaps because of confidentiality agreements with counterparties and lack of information about debtors on loans tied to the contracts.
/endsnip
I'd say the latter is the material point of contention, whereas the former is entirely relevant to the US Taxpayer Owned company. You and your clients have no confidentiality.
I have got to imagine the bank owned volume is playing havoc with the matched pairs methodology. Imagine if they finally wise up and start including those. ~RD
I dont like that they do. In fact it irks me quite a bit. Is there any rational for not including them?
Imagine if they finally wise up and start including those.
Yeah, but how would they quantify that? Wouldn't that require reworking the whole methodology? I suppose you could just treat the bank sale as a regular sale though sometimes the foreclosed regular sale is wiped out...and the deed goes from the prior owner (not the foreclosed) to the new owner (buyer from the bank). I suppose C-S just uses recording fees.
nades (homepage, profile) wrote (in reply to...) on Tue, 6/30/2009 - 6:30 am
I have got to imagine the bank owned volume is playing havoc with the matched pairs methodology. Imagine if they finally wise up and start including those.~RD
I dont like that they do. In fact it irks me quite a bit. Is there any rational for not including them?
When CSI was designed the rare instance of a bank REO usually meant a material change/event that would legitimately be excluded. But now that arms length transactions are the minority that rationale no longer applies. you just cannot throw out 60% of the transaction volume because they are assumed to be tainted.
So if we use the composite 20 as a reference and lets say it goes to 100 and then goes flat right from there... This means we still have another ~40% to fall? (140/100)? I dont think this was in the adverse scenario....
@Dawg When CSI was designed the rare instance of a bank REO usually meant a material change/event that would legitimately be excluded. But now that arms length transactions are the minority that rationale no longer applies. you just cannot throw out 60% of the transaction volume because they are assumed to be tainted.
It really doesn't make sense if the bank is buying back at foreclosure and then marketing the property through a realtor or the MLS.
CSI methodology:
In an arms-length transaction, both the buyer and seller act in their best
economic interest when agreeing upon a price. When they can be identified from a deed
record2, non-arms-length transactions are excluded from the pairing process. The most
typical types of non-arms-length transactions are property transfers between family
members and repossessions of properties by mortgage lenders at the beginning of
foreclosure proceedings. Subsequent sales by mortgage lenders of foreclosed properties
are included in repeat sale pairs, because they are arms-length transactions.
The pairing process is also designed to exclude sales of properties that may have been
subject to substantial physical changes immediately preceding or following the
transaction. Furthermore, since a property must have two recorded transactions before it
can be included as a repeat sale pair, newly constructed homes are excluded from the
index calculation process until they have been sold at least twice.
I have got to imagine the bank owned volume is playing havoc with the matched pairs methodology. Imagine if they finally wise up and start including those.
I think you're thinking of the FHFA (fka OFHEO) index, dawg. No reason C/S would exclude any pairs based on who the seller is.
It looks like we have a long, long way to go yet. Prices are still not historically low and yet we are in a recession/depression of historical magnitude.
It will keep grinding.
I think that if REO / FCs were included in the CSI it would inflate the price. The FC 'sales price' is above market value. If it was below there would still be some 'equity' left in it and it wouldnt be a FC. Still something tells me if you included FCs it would lower the index.....
non-arms-length transactions ...repossessions of properties by mortgage lenders at the beginning of
foreclosure proceedings. Subsequent sales by mortgage lenders of foreclosed properties
are included in repeat sale pairs, because they are arms-length transactions.
See the bias in the sales pair they exclude and the one they include? And just go to JtR and see the tens of thousands of dollars the banks are pouring into their REOs and you got an even greater upward bias in the CSI numbers.
Dawg - where did you find that? Can you post a link. That reads very strangely for a number of reasons, not the least of which is that whomever wrote it doesn't understand the concept of "arms-length transaction".
"repossessions of properties by mortgage lenders ? at the beginning of
foreclosure proceedings."
Last I checked, the lender took title to a property at the conclusion of a foreclosure proceeding.
Shnaps: My read (of RD's blurb) is that banks taking back the property (which occurs at some value determined by courthouse auction) is not included. It makes sense to exclude those, they are not "real" auctions in the current market.
nades - that's right. I just wonder if the Rob Dawg grasps this. And if he does, what's his beef with this approach?
The most typical types of non-arms-length transactions are property transfers between family
members and repossessions of properties by mortgage lenders at the beginning of
foreclosure proceedings. Subsequent sales by mortgage lenders of foreclosed properties
are included in repeat sale pairs, because they are arms-length transactions.
Excluding the repo transaction presumes that the bank is not acting in their own best fiduciary interests. The idea of taking the last sale, pretending there was no foreclosure and then matching the subsequent bank sale violates the other exclusion criteria of substantial improvements.
@nades Shnaps - they include bank to private party but not defaulted buyer back to bank. At least thats the way I read it....
Well, that seems right to me...you're not really missing any juicy drops in price by skipping over the trustee sale, if it goes back to the bank - that's going to be for at least the face of the note. If the bank takes a haircut on the subsequent sale...well, that will be reflected. Seems to work.
Dawg - I suggest you re-read Tanta's fine ubernerd on this topic. There are a number of reasons foreclosure transfers are not "normal sales" in the "open market" sense, and therefore it is entirely appropriate to exclude them when one is building a repeat-sales index.
Excluding the repo transaction presumes that the bank is not acting in their own best fiduciary interests.
I don't think so, in fact the opposite: It is the banks best fiduciary interest to bid $1 on the courthouse steps. You do not want those sales in the index.
The bank subsequently sells the house as REO. That sale is recorded.
As long as the bank didn't make any improvements physical or financial right?
Don't know what you mean by "financial" improvements.
But as for physical improvements: None of those are backed out of the index, right? So a non-foreclosure that has granite installed simply sells for more money (in theory (ha ha)), and that gets included in the index. Likewise for REO.
The Chicago Sun-Times has a web page through which you can access a database of actual recorded sales for Chicago and its surrounding suburbs.
Viewing the contents of that database and the MLS, one can see that hardly any homes are selling in the mid-high ($600-800k) sector of the market, and almost none above $800k, but also that asking prices haven't fallen much in the mid-high sector. I suspect that in the mid-high sector the only sales are being made to the few who still believe in real estate as a sure investment myth and to corporate transferees who are getting some subsidy, and that the owners of the homes that aren't selling are still holding out in vain hope the market will come back (many of the homes are now in their third year on the market). At the high end, most of the homes are vacant spec McMansions that remain unsold despite the builders having slashed prices by 20%.
Since the Case-Shiller stats are based on actual sale-to-sale changes in prices of the same homes, in the mid-high sector we are seeing only the prices of the small number of sales actually being made -- not the prices that would be needed to clear the market.
How much longer can the owners of the mid-high sector homes now their third year on the market hold out? Many are up for rent, but of course find no tenants because the owners are fantasizing that they can demand rents that will pay the mortgage and taxes.
When the owners of these homes can hold out no longer and the prices drop low enough to clear the market, the Case-Shiller index decline will be breathtaking.
Obamma,s policies may not be working as well as many had hoped:
1st 100 days - There are 2.9 million more people unemployed in May than there were unemployed in January. The unemployment rate went from 7.6% to 9.4%. Since May 2008, we have lost 5.5 million jobs. The biggest losers were: Manufacturing 1.5 million lost Finance & Prof Serv 1.5 million lost Construction 1.1 million lost Retail & Leisure 1.3 million lost
A lot of folks commenting on the SPCS HPI report for April has completed misinterpreted the data, saying that the decline in the indes in April "shows" that home prices declined in April vs. March. That, of course, is not an accurate statement. As any competent analyst knows, the SPCS index for a given month reflects transactions over the rolling 3-months. As such, the April SPCS vs the March SPCS reflects prices during the February to April period relative to the January to March period. Looking at other sources or home price behaviior based just on a given month's transactions, the SPCS HPIs for April would suggest that home prices INCREASED in April vs. March! Here is an excerpt from Lawler Economic & Housing Consulting's daily commentary on the SPCS report:
Many press reports and headlines will probably interpret the release as indicating that while the pace of home price declines slowed in April, home prices were still declining in most metro areas in April. In fact that is not a correct interpretation. As I keep repeating, the monthly SPCS HPI reflects transactions over a 3-month period, and the % change in the April composite-20 HPI reflects transactions prices from February to April compared to transactions prices from January to March. The monthly % change in the SPCS composite-20 was -2.1871% in March and -0.564% in April, and that is a very large swing in a 3-month moving-average time series. In fact, given data from other repeat-transactions HPIs on the monthly pattern of home prices, such a swing would actually imply that a monthly SPCS composite-20 HPI based solely on transactions in that month actually INCREASED in April, which is broadly consistent with other information on home price behavior in April vs. March. Indeed, in some metro markets the SPCS HPI suggests that prices may have increased considerably in April, even though the monthly HPI declined!
For example, consider a monthly time series that behaves as follows:
Sep 100
Oct 98.1 -1.9%
Nov 96.24 -1.9% 98.11333
Dec 93.45 -2.9% 95.93 -2.2% -2.6%
Jan 91.11 -2.5% 93.6 -2.4% -2.8%
Feb 89.01 -2.3% 91.19 -2.6% -2.2%
Mar 88.12 -1.0% 89.41333 -1.9% -2.2%
Apr 89.44 1.5% 88.85667 -0.6% -0.6%
Note that for this time series – which is actually based on the LP-derived composite-20 HPI through March -- even though April shows an increase of 1.5%, the 3-month moving average shows a decline of 0.6%.
For some metro areas, the SPCS HPI data would suggest a startlingly large jump in April home prices. E.g., the Minneapolis HPI in March, reflecting transactions from January through March, was an astonishingly large 5.9% below the February HPI, which reflected transactions from December through February. The April HPI – which (duh!) reflected transactions from February through April – was just 0.7% below the March HPI, suggesting that an HPI based just on April transactions would have jumped considerably from an HPI based just on March transactions.
One possible reason for the improvement in home prices in many areas in April vs. March was the decline in the share of transactions that were foreclosure sales. While there are no hard data nationally on that number, numerous local realtors/MLS reported that the foreclosure share of resales peaked in the first few months of this year, and declined in April and fell even further in May. For various reasons (home condition, desire of lenders/servicers to get rid of REO, etc.) foreclosure sales tend to sell at substantial discounts to “comparable” (at least at one time) homes sold in a traditional fashion. Given the further drop in the foreclosure share of home sales in May, there is a pretty good chance that the SPCS composite-20 HPI will be little changed in May, and that it will show an increase for Q2/09 as a whole.
c'mon!
pigged again. rats
CR,
The spin is that they are not falling as fast, and therefore recovery, in one of its many forms, is around the corner...
--bh
I have got to imagine the bank owned volume is playing havoc with the matched pairs methodology. Imagine if they finally wise up and start including those.
money quote from Mark Haynes re Case-Shiller report: ..."there's something there."
can they save my poor SPY puts before end of July though.......puts on those darned green shoots
So, how many trillion dollars did we pour into producing this little 2-3% change in the rate of decline?
Good news per CNBS.
just like consumer confidence, i would really like to know who they are not the people i talk to in the stores. but different in all areas,regions states,counties and cities. no wonder this unwinding is taking soooooooo long.
Arizona looks to pulling away from the pack in so many awful economic categories, in bizarro-phoenix fashion.
The only question being why we're watching two people in a closet staring at a big, unblinking glass eye, repeating whatever is fed to them in little earpieces.
And those earpieces are talking the book of their masters, the conglomerate GE.
How bizarre! Maybe I'll switch to Boomberg, it must be different.
OT on AIG & Euro def swaps from bloomberg:
/snip
June 30 (Bloomberg) -- American International Group Inc., the insurer bailed out by the U.S., said that valuation declines on credit-default swaps sold to European banks could have a “material adverse effect” on the company’s results.
The risk of losses on the derivatives may last “longer than anticipated,” the New York-based insurer said late yesterday in a regulatory filing updating the “risk factors” in its 2008 annual report. The firm had $192.6 billion in swaps allowing lenders to reduce the funds they had to hold in reserve as of March 31, AIG said.
/endsnip
This however is the creamy, fudge-colored middle of the crap sandwich:
/snip
AIG said it was unable to provide full details on the value of assets backed by the swaps because of confidentiality agreements with counterparties and lack of information about debtors on loans tied to the contracts.
/endsnip
I'd say the latter is the material point of contention, whereas the former is entirely relevant to the US Taxpayer Owned company. You and your clients have no confidentiality.
--bh
Oh good, Phoenix is only declining 20%/yr as opposed to 40%/yr declines seen last month.
I have got to imagine the bank owned volume is playing havoc with the matched pairs methodology. Imagine if they finally wise up and start including those.
~RD
I dont like that they do. In fact it irks me quite a bit. Is there any rational for not including them?
Imagine if they finally wise up and start including those.
Yeah, but how would they quantify that? Wouldn't that require reworking the whole methodology? I suppose you could just treat the bank sale as a regular sale though sometimes the foreclosed regular sale is wiped out...and the deed goes from the prior owner (not the foreclosed) to the new owner (buyer from the bank). I suppose C-S just uses recording fees.
NM - I'm sure it's not that complicated.
I suppose you could just treat the bank sale as a regular sale
Edzachary, heck that number could still be too high....
nades (homepage, profile) wrote (in reply to...) on Tue, 6/30/2009 - 6:30 am
I have got to imagine the bank owned volume is playing havoc with the matched pairs methodology. Imagine if they finally wise up and start including those.~RD
I dont like that they do. In fact it irks me quite a bit. Is there any rational for not including them?
When CSI was designed the rare instance of a bank REO usually meant a material change/event that would legitimately be excluded. But now that arms length transactions are the minority that rationale no longer applies. you just cannot throw out 60% of the transaction volume because they are assumed to be tainted.
CS excludes bank sales? Really?
Dawg, I never did track down your pointer for excluding first time + second time sales.
So in short, I have no idea what CS is based on really.
Mint's $15.3 M golden dilemma: Was there a heist?
Mint's $15.3 M golden dilemma: Was there a heist?
That is 1200 pounds of gold. Not something you just put in your pocket and walk out with.
So if we use the composite 20 as a reference and lets say it goes to 100 and then goes flat right from there... This means we still have another ~40% to fall? (140/100)? I dont think this was in the adverse scenario....
Truss no one...
@Dawg When CSI was designed the rare instance of a bank REO usually meant a material change/event that would legitimately be excluded. But now that arms length transactions are the minority that rationale no longer applies. you just cannot throw out 60% of the transaction volume because they are assumed to be tainted.
It really doesn't make sense if the bank is buying back at foreclosure and then marketing the property through a realtor or the MLS.
JP: don't doubt the dawg
You have to count that as a pretty good heist, eh? No one knows when, how, or, really, if, a $15m theft occurred.
CSI methodology:
In an arms-length transaction, both the buyer and seller act in their best
economic interest when agreeing upon a price. When they can be identified from a deed
record2, non-arms-length transactions are excluded from the pairing process. The most
typical types of non-arms-length transactions are property transfers between family
members and repossessions of properties by mortgage lenders at the beginning of
foreclosure proceedings. Subsequent sales by mortgage lenders of foreclosed properties
are included in repeat sale pairs, because they are arms-length transactions.
The pairing process is also designed to exclude sales of properties that may have been
subject to substantial physical changes immediately preceding or following the
transaction. Furthermore, since a property must have two recorded transactions before it
can be included as a repeat sale pair, newly constructed homes are excluded from the
index calculation process until they have been sold at least twice.
A deeply flawed tool for the current market.
I have got to imagine the bank owned volume is playing havoc with the matched pairs methodology. Imagine if they finally wise up and start including those.
I think you're thinking of the FHFA (fka OFHEO) index, dawg. No reason C/S would exclude any pairs based on who the seller is.
It looks like we have a long, long way to go yet. Prices are still not historically low and yet we are in a recession/depression of historical magnitude.
It will keep grinding.
Homes were just the whore d'oeuvres in the ongoing financial fiasco, an appetizer course...
Interesting chart here :
http://www.gimmiethescoop.com/img/AIG-insurance-bailout2.png
I think that if REO / FCs were included in the CSI it would inflate the price. The FC 'sales price' is above market value. If it was below there would still be some 'equity' left in it and it wouldnt be a FC. Still something tells me if you included FCs it would lower the index.....
non-arms-length transactions ...repossessions of properties by mortgage lenders at the beginning of
foreclosure proceedings. Subsequent sales by mortgage lenders of foreclosed properties
are included in repeat sale pairs, because they are arms-length transactions.
See the bias in the sales pair they exclude and the one they include? And just go to JtR and see the tens of thousands of dollars the banks are pouring into their REOs and you got an even greater upward bias in the CSI numbers.
Thanks Dawg. (And volker: I trust no one. Nothing personal. But respect is another issue.)
Subsequent sales by mortgage lenders of foreclosed properties are included in repeat sale pairs, because they are arms-length transactions.
Foreclosure sales are in the index.
newly constructed homes are excluded from the index calculation process until they have been sold at least twice.
So the initial (=new home) sale does not count, but the next one does.
Is that correct or do I need more coffee.
EDIT: Regardless, I'm off to get more coffee, then I'll parse dawg's next comment.
Oh, on occasion you must doubt the dawg - just know you are going to be in a dawg fight!
Dawg - where did you find that? Can you post a link. That reads very strangely for a number of reasons, not the least of which is that whomever wrote it doesn't understand the concept of "arms-length transaction".
"repossessions of properties by mortgage lenders ? at the beginning of
foreclosure proceedings."
Last I checked, the lender took title to a property at the conclusion of a foreclosure proceeding.
Here is the current S&P/CS methodology document. I see nothing about exlcuding REO sales (page 5).
Shnaps: My read (of RD's blurb) is that banks taking back the property (which occurs at some value determined by courthouse auction) is not included. It makes sense to exclude those, they are not "real" auctions in the current market.
The subsequent sale by the bank is included.
fwiw.
Shnaps: Scribd (Page 6) Bottom
or here
HOME PRICE INDICES S&P/CASE-SHILLER
JP - I think you're right. They could have worded that better.
In any event, It appears they do in fact include the 'bank-to-private party' sales that Dawg seems to think they are omitting.
Also, I see we've been pigged. We'll have to 'Doubt the Dawg' elsewhere today.
Hmmm that S&P doc seems to indicate that only after three sales is a new home included:
The S&P/Case-Shiller indices do not sample sale prices associated with new construction,
So after the new home sale, two more sales must occur.
Shnaps - they include bank to private party but not defaulted buyer back to bank. At least thats the way I read it....
nades - that's right. I just wonder if the Rob Dawg grasps this. And if he does, what's his beef with this approach?
The most typical types of non-arms-length transactions are property transfers between family
members and repossessions of properties by mortgage lenders at the beginning of
foreclosure proceedings. Subsequent sales by mortgage lenders of foreclosed properties
are included in repeat sale pairs, because they are arms-length transactions.
I think his concern was improvements made to the house? We'll have to wait to see if he checks in...
Excluding the repo transaction presumes that the bank is not acting in their own best fiduciary interests. The idea of taking the last sale, pretending there was no foreclosure and then matching the subsequent bank sale violates the other exclusion criteria of substantial improvements.
@nades Shnaps - they include bank to private party but not defaulted buyer back to bank. At least thats the way I read it....
Well, that seems right to me...you're not really missing any juicy drops in price by skipping over the trustee sale, if it goes back to the bank - that's going to be for at least the face of the note. If the bank takes a haircut on the subsequent sale...well, that will be reflected. Seems to work.
I concur....
Dawg - I suggest you re-read Tanta's fine ubernerd on this topic. There are a number of reasons foreclosure transfers are not "normal sales" in the "open market" sense, and therefore it is entirely appropriate to exclude them when one is building a repeat-sales index.
Excluding the repo transaction presumes that the bank is not acting in their own best fiduciary interests.
I don't think so, in fact the opposite: It is the banks best fiduciary interest to bid $1 on the courthouse steps. You do not want those sales in the index.
The bank subsequently sells the house as REO. That sale is recorded.
That is how you would also design the index, no?
Thanks for the ubernerd post: I missed that in the first go around.
"Weirdos in Georgia" LOL! I raise my coffee cup in her general direction.
As long as the bank didn't make any improvements physical or financial right?
Dawg - only if you consider a fresh coat of paint and steam-cleaning vomit stains an 'improvement'.
As long as the bank didn't make any improvements physical or financial right?
Don't know what you mean by "financial" improvements.
But as for physical improvements: None of those are backed out of the index, right? So a non-foreclosure that has granite installed simply sells for more money (in theory (ha ha)), and that gets included in the index. Likewise for REO.
The Chicago Sun-Times has a web page through which you can access a database of actual recorded sales for Chicago and its surrounding suburbs.
Viewing the contents of that database and the MLS, one can see that hardly any homes are selling in the mid-high ($600-800k) sector of the market, and almost none above $800k, but also that asking prices haven't fallen much in the mid-high sector. I suspect that in the mid-high sector the only sales are being made to the few who still believe in real estate as a sure investment myth and to corporate transferees who are getting some subsidy, and that the owners of the homes that aren't selling are still holding out in vain hope the market will come back (many of the homes are now in their third year on the market). At the high end, most of the homes are vacant spec McMansions that remain unsold despite the builders having slashed prices by 20%.
Since the Case-Shiller stats are based on actual sale-to-sale changes in prices of the same homes, in the mid-high sector we are seeing only the prices of the small number of sales actually being made -- not the prices that would be needed to clear the market.
How much longer can the owners of the mid-high sector homes now their third year on the market hold out? Many are up for rent, but of course find no tenants because the owners are fantasizing that they can demand rents that will pay the mortgage and taxes.
When the owners of these homes can hold out no longer and the prices drop low enough to clear the market, the Case-Shiller index decline will be breathtaking.
We need to stop the corruption by Goldman, bank of Amerika, and Jp Morgan. Stop the Geithner and bernanke. .
Financial Opinions Updated Daily iamned.com
read the full Rolling Stone [Taibbi] The Great American Bubble Machine article.
Obamma,s policies may not be working as well as many had hoped:
1st 100 days - There are 2.9 million more people unemployed in May than there were unemployed in January. The unemployment rate went from 7.6% to 9.4%. Since May 2008, we have lost 5.5 million jobs. The biggest losers were: Manufacturing 1.5 million lost Finance & Prof Serv 1.5 million lost Construction 1.1 million lost Retail & Leisure 1.3 million lost
good articles Financial Opinions Updated Daily iamned.com
recommended reading
Why isnt the stock market falling when the economy is sinking? Why cant we get change in Washington that we really need? Why vote?
Real Estate Values?
The Question is .......... How low can it go?
Some write there are 8 million more homeowners under water in the near future.
Is this a great banking system or what.
MLB
A lot of folks commenting on the SPCS HPI report for April has completed misinterpreted the data, saying that the decline in the indes in April "shows" that home prices declined in April vs. March. That, of course, is not an accurate statement. As any competent analyst knows, the SPCS index for a given month reflects transactions over the rolling 3-months. As such, the April SPCS vs the March SPCS reflects prices during the February to April period relative to the January to March period. Looking at other sources or home price behaviior based just on a given month's transactions, the SPCS HPIs for April would suggest that home prices INCREASED in April vs. March! Here is an excerpt from Lawler Economic & Housing Consulting's daily commentary on the SPCS report:
Many press reports and headlines will probably interpret the release as indicating that while the pace of home price declines slowed in April, home prices were still declining in most metro areas in April. In fact that is not a correct interpretation. As I keep repeating, the monthly SPCS HPI reflects transactions over a 3-month period, and the % change in the April composite-20 HPI reflects transactions prices from February to April compared to transactions prices from January to March. The monthly % change in the SPCS composite-20 was -2.1871% in March and -0.564% in April, and that is a very large swing in a 3-month moving-average time series. In fact, given data from other repeat-transactions HPIs on the monthly pattern of home prices, such a swing would actually imply that a monthly SPCS composite-20 HPI based solely on transactions in that month actually INCREASED in April, which is broadly consistent with other information on home price behavior in April vs. March. Indeed, in some metro markets the SPCS HPI suggests that prices may have increased considerably in April, even though the monthly HPI declined!
For example, consider a monthly time series that behaves as follows:
Monthly Monthly % ch3-mo moving average % chg, 3-mo mov. avg. SPCS-20 % chg
Sep 100
Oct 98.1 -1.9%
Nov 96.24 -1.9% 98.11333
Dec 93.45 -2.9% 95.93 -2.2% -2.6%
Jan 91.11 -2.5% 93.6 -2.4% -2.8%
Feb 89.01 -2.3% 91.19 -2.6% -2.2%
Mar 88.12 -1.0% 89.41333 -1.9% -2.2%
Apr 89.44 1.5% 88.85667 -0.6% -0.6%
Note that for this time series – which is actually based on the LP-derived composite-20 HPI through March -- even though April shows an increase of 1.5%, the 3-month moving average shows a decline of 0.6%.
For some metro areas, the SPCS HPI data would suggest a startlingly large jump in April home prices. E.g., the Minneapolis HPI in March, reflecting transactions from January through March, was an astonishingly large 5.9% below the February HPI, which reflected transactions from December through February. The April HPI – which (duh!) reflected transactions from February through April – was just 0.7% below the March HPI, suggesting that an HPI based just on April transactions would have jumped considerably from an HPI based just on March transactions.
One possible reason for the improvement in home prices in many areas in April vs. March was the decline in the share of transactions that were foreclosure sales. While there are no hard data nationally on that number, numerous local realtors/MLS reported that the foreclosure share of resales peaked in the first few months of this year, and declined in April and fell even further in May. For various reasons (home condition, desire of lenders/servicers to get rid of REO, etc.) foreclosure sales tend to sell at substantial discounts to “comparable” (at least at one time) homes sold in a traditional fashion. Given the further drop in the foreclosure share of home sales in May, there is a pretty good chance that the SPCS composite-20 HPI will be little changed in May, and that it will show an increase for Q2/09 as a whole.