Cml Latest Figures Released This Morning "october Saw Expected Decline In Mortgage Lending "

[koala_bear]

http://www.cml.org.uk/cml/media/press/2802

Headline "October saw expected decline in mortgage lending"

An expected decline in mortgage lending towards the end of the year has now become apparent, data released by the Council of Mortgage lenders shows. Lending for both house purchase and remortgaging were affected by a lull in activity in October. There were 46,000 loans for house purchase (worth £6.7 billion), down 4% in number and 6% by value from September. The total was 16% lower (12% by value) than in October 2009, but lending numbers in the final quarter of 2009 were boosted as buyers brought forward transactions to take advantage of the stamp duty holiday.

Important notes at the bottom of the page

"4. The November RMS data will be released on Friday 14 January 2011.5. New reporting arrangements for our Regulated Mortgage Survey will take effect in January 2011. As a result of this, next month’s figures, including all historic data back to April 2005, will be based on a new version of the dataset, and may show revisions over this period."

Expect more Halifax style post revisionism in the next set!

[maffo in oxford]

Headline "October saw expected decline in mortgage lending"

Not an 'unexpected' decline then?!

[Bloo Loo]

Not an 'unexpected' decline then?!

course not, they saw the October figures at the beginning of november and have gotten over it.

[koala_bear]

More entertaining quotes, have they really seen the light?

"The take-up of repayment mortgages in October was the highest in more than five years for both home-buyers and those remortgaging. 93% of first-time buyers took out a repayment mortgage in October, the highest proportion since records began in 1974, showing a clear shift away from a pre-2007 norm of around 30% of first-time buyers opting for interest-only mortgages. This shift shows lenders have been adjusting their loan criteria in anticipation of possible regulatory changes, and a recognition that repayment mortgages may be in the best interests of less experienced borrowers such as first-time buyers."

FTB lending and affordability

# FTB loans: 17,000

Total lent: 2.000bn

LTV: 80%

Earnings ratio: 3.19

proportion of income spent on interest payments:13.3%

Doing some quick calculations gives:

average loan 117.6k

average price paid 147.1k

average earnings 46.1k --> which implies that 80-90% of loans are to couples or they are just loaning to singles in London...

Given that avg. price paid by home movers averaged 231.7k it would seem the Haliwide standardised basket of homes and customers are heavily skewed to FTBs so that the while they are showing the bottom dropping out of the market the middle and top are still doing ok as academagix and LR (both laggers I know) as still doing wiley e coyote style gravity defying...

[Bloo Loo]

More entertaining quotes, have they really seen the light?

"The take-up of repayment mortgages in October was the highest in more than five years for both home-buyers and those remortgaging. 93% of first-time buyers took out a repayment mortgage in October, the highest proportion since records began in 1974, showing a clear shift away from a pre-2007 norm of around 30% of first-time buyers opting for interest-only mortgages. This shift shows lenders have been adjusting their loan criteria in anticipation of possible regulatory changes, and a recognition that repayment mortgages may be in the best interests of less experienced borrowers such as first-time buyers."

FTB lending and affordability

# FTB loans: 17,000

Total lent: 2.000bn

LTV: 80%

Earnings ratio: 3.19

proportion of income spent on interest payments:13.3%

Doing some quick calculations gives:

average loan 117.6k

average price paid 147.1k

average earnings 46.1k --> which implies that 80-90% of loans are to couples or they are just loaning to singles in London...

Given that avg. price paid by home movers averaged 231.7k it would seem the Haliwide standardised basket of homes and customers are heavily skewed to FTBs so that the while they are showing the bottom dropping out of the market the middle and top are still doing ok as academagix and LR (both laggers I know) as still doing wiley e coyote style gravity defying...

everyone earns 46.1K dontchaknow.

the paperwork confirms

[FreeTrader]

Once again the CML seems to be having some problems getting its numbers right, and its apparent bias towards producing figures in a way that softens the true picture of what’s happening in the mortgage market is probably why no one on this site takes much notice of its statistical releases any more.

Last month’s press release is here, and today’s is here. Look at the loan numbers, and the claimed one-month percentage falls. Here’s a snapshot highlighting the issue:

Not very subtle, but then again who needs to be with our unquestioning mainstream press?

[Democorruptcy]

Once again the CML seems to be having some problems getting its numbers right, and its apparent bias towards producing figures in a way that softens the true picture of what’s happening in the mortgage market is probably why no one on this site takes much notice of its statistical releases any more.

Not very subtle, but then again who needs to be with our unquestioning mainstream press?

Just had a look at the CML site.

They approve of what their chums at the FSA have done:

The Council of Mortgage Lenders welcomes the statement by the Financial Services Authority that it is deferring implementation of extending the approved persons regime to all those who sell mortgages.

http://www.cml.org.uk/cml/media/press/2798

So anyone who sells a mortgage is "approved". Were the FSA going to bring in some rules but as usual it's all left until the tomorrow that never comes?

FSA statement:

In June 2010 as part of the Mortgage Market Review we announced that we would be extending the approved persons' regime to include anyone who advises on or sells mortgages. We remain committed to making these changes to the approved persons' regime, but as part of our ongoing reprioritisation of work; particularly around the regulatory reform agenda, we are deferring introduction of the changes to 2012/2013.

Once the rules are finalised, we will give firms sufficient time to put changes in place to comply with the approved persons' regime, as with any new rules.

We will be publishing a full economic analysis of all our Mortgage Market Review proposals next year which will inform the final rules

http://www.fsa.gov.uk/smallfirms/updates/persons_regime.shtml

So does this mean that any Tom, Dick or Harry can sell mortgages?

[Pent Up]

Once again the CML seems to be having some problems getting its numbers right, and its apparent bias towards producing figures in a way that softens the true picture of what’s happening in the mortgage market is probably why no one on this site takes much notice of its statistical releases any more.

Last month’s press release is here, and today’s is here. Look at the loan numbers, and the claimed one-month percentage falls. Here’s a snapshot highlighting the issue:

Not very subtle, but then again who needs to be with our unquestioning mainstream press?

I emailed them some time regarding their dodgy percentages and to their credit got a reply almost straight away saying that it's due to the approval number being rounded and the percentage being calculated on the actual numbers. But the descrepency then was about 2%.

This seems a bit more than can be attributed to rounding.

[Pent Up]

I emailed them some time regarding their dodgy percentages and to their credit got a reply almost straight away saying that it's due to the approval number being rounded and the percentage being calculated on the actual numbers. But the descrepency then was about 2%.

This seems a bit more than can be attributed to rounding.

50k to 46k is 8% drop. They are giving a 4% drop.

They best they could do with rounding is 46500 and 49500 giving a 6% drop. So it's not rounding.

Must be a revised figure they are using for last month.

[Democorruptcy]

Once again the CML seems to be having some problems getting its numbers right, and its apparent bias towards producing figures in a way that softens the true picture of what’s happening in the mortgage market is probably why no one on this site takes much notice of its statistical releases any more.

Last month’s press release is here, and today’s is here. Look at the loan numbers, and the claimed one-month percentage falls. Here’s a snapshot highlighting the issue:

Not very subtle, but then again who needs to be with our unquestioning mainstream press?

CML say

"In September, the data we released (grossed up to reflect all of the market) showed that there were 50,000 house purchase loans in the month but the figures are always an estimate as we do not always get every lender’s data in on time. A month later, the actual figures show that house purchase lending for September was 48,000 not 50,000, hence a 4% drop is 46,000."

[Pent Up]

CML say

"In September, the data we released (grossed up to reflect all of the market) showed that there were 50,000 house purchase loans in the month but the figures are always an estimate as we do not always get every lender’s data in on time. A month later, the actual figures show that house purchase lending for September was 48,000 not 50,000, hence a 4% drop is 46,000."

I thought that would be it. So I wonder what Octobers figure actually was? 44k?

[FreeTrader]

CML say

"In September, the data we released (grossed up to reflect all of the market) showed that there were 50,000 house purchase loans in the month but the figures are always an estimate as we do not always get every lender’s data in on time. A month later, the actual figures show that house purchase lending for September was 48,000 not 50,000, hence a 4% drop is 46,000."

Yeah, right. If your numbers are likely to be subject to revisions of 5% or so (as we’ve been seeing recently, both month-on-month and annually), then you shouldn’t be making authoritative statements in your press releases.

I’ll buy the occasional revision which works in their favour, but not what we’ve been seeing recently. At the very least they should knock off 750 loans each month because that’s what the average annual revision has been for the past 22 months. That’s a very hefty upward bias, and in my view that makes their statisticians either incompetent with their estimates or – how shall I put it? – overly optimistic (cough).

Next month’s figures should be interesting though. Note the footnote in today’s release:

“New reporting arrangements for our Regulated Mortgage Survey will take effect in January 2011. As a result of this, next month’s figures, including all historic data back to April 2005, will be based on a new version of the dataset, and may show revisions over this period.”

[koala_bear]

CML say

"In September, the data we released (grossed up to reflect all of the market) showed that there were 50,000 house purchase loans in the month but the figures are always an estimate as we do not always get every lender's data in on time. A month later, the actual figures show that house purchase lending for September was 48,000 not 50,000, hence a 4% drop is 46,000."

Good result in getting that from them.

So it looks if they don't get the data from a particular lender they just use the previous months figure for the lender as an estimate, which would be relatively sensible especially in a rising market...

So if there was another late arrival from one of the lenders for this month, the figure could be worse than stated when they are "real" next month.

[Democorruptcy]

Good result in getting that from them.

So it looks if they don't get the data from a particular lender they just use the previous months figure for the lender as an estimate, which would be relatively sensible especially in a rising market...

So if there was another late arrival from one of the lenders for this month, the figure could be worse than stated when they are "real" next month.

This is also what they said:

When we release our monthly mortgage figures, we use the data we have been given by our sample of lenders and we gross the figures up to reflect total market size. Our figures are therefore estimates and the figure for a particular month can sometimes fluctuate due to data revisions, as you have seen with September’s number of house purchase loans. This is standard practice in the reporting of data and those in the industry, including financial journalists, to whom these press releases are aimed, are aware of this.

It sounds like a silly how many people have been watching TV figure. Just a sample that is then multiplied upwards as if that really reflects a countrywide figure.

So who are in their sample and are they swapped about month to month when it suits the CML? i.e. one when one firms figures is more suitable to the CML than another?

[koala_bear]

Next month's figures should be interesting though. Note the footnote in today's release:

"New reporting arrangements for our Regulated Mortgage Survey will take effect in January 2011. As a result of this, next month's figures, including all historic data back to April 2005, will be based on a new version of the dataset, and may show revisions over this period."

I've been wondering what the changes are and what they will actually do to the numbers.

Include 100% of lending not just 90% (much easier now as there are far fewer lenders)

Include the BTL (currently not included)

Include additional loans the 100 to 125% part of NRocks infamous together mortgages (currently not included)

Include MEWing

Include mortgage type (i.e. split out Interest only types according to repayment vehicle)

Include broker or direct from lender data

The CML blurb on the change to the current format from the previous in 2005

1. The regulated mortgage survey (RMS) is based on lenders providing us with a copy of their statutory product sales data (PSD) which they also submit to the FSA. The RMS has much wider market coverage than the old SML (90% compared to 50%) and so the data is more robust. Because of this, there are discrepancies between the data we previously reported in the SML and the new RMS data. Here are the most notable changes:

• total lending for house purchase is lower than previously estimated;
• loans to first-time buyers are significantly higher;
• typical loans to purchase price ratios are slightly higher and income multiples lower;
• average loan sizes are higher, particularly for movers and remortgagors;
• volumes of fixed-rate lending are higher, and non-discounted variable-rate leading lower;
• fewer loans are taken out on a capital-plus-interest basis, and more are interest only; and
• Scotland has lower lending volumes.

The differences in data between the RMS and SML are explained fully in the attached note, 'New mortgage market data: key changes and better information'

link to attchement http://www.cml.org.uk/cml/media/press/680 (pdf at the bottom)

[FreeTrader]

Good result in getting that from them.

So it looks if they don't get the data from a particular lender they just use the previous months figure for the lender as an estimate, which would be relatively sensible especially in a rising market...

So if there was another late arrival from one of the lenders for this month, the figure could be worse than stated when they are "real" next month.

I don't think what we're seeing is reasonable, koala.

As you note, the CML’s mortgage data is compiled from their Regulated Mortgage Survey (RMS), which started in late 2005. This in turn is based on Product Sales Data returns that lenders are required to submit to the FSA.

Some lenders submit data monthly, but the FSA only requires a minimum of quarterly reporting. It’s therefore at the end of each quarter that the CML’s numbers should be more accurate because they have a more complete dataset, and in the other two months they have to make estimates.

If you look back through the annual revisions the CML has made for the past 22 months, every revision that falls on an end-quarter month has been upward or zero with an average annual revision of +400 loans.

Every revision which falls on months where estimates are made has been downward except for one (and that was zero). The average revision is -1400 loans, and that represents a 2% upward bias on the estimates.

[koala_bear]

This is also what they said:

It sounds like a silly how many people have been watching TV figure. Just a sample that is then multiplied upwards as if that really reflects a countrywide figure.

So who are in their sample and are they swapped about month to month when it suits the CML? i.e. one when one firms figures is more suitable to the CML than another?

Just done a bit of digging, the CML figures are based on lenders statutory data submissions to the FSA (effectively cc'd to CML) which most lenders do monthly but are only required to do quarterly by the FSA.

Some lenders only submit quarterly so they have to fudge it for those lenders (more subtle than /3?) then guestimate how much of the market the CML lenders cover (~90%) and then scale up to what they think is 100%. Quarterly to monthly adjustment is easily done by sitting on the data for a month comparing it with next months data release (they could easily have had the data this time last month) and using the behaviour of the lenders' who submit data monthly to estimate/model the behaviour of the lenders who submit quarterly, the revise a month or 2 later when you get the real data...

The improvements they mention starting for the next data release may be monthly reporting for all lenders which might increase the accuracy but could lead to a sudden extra drop for the next month as a drop in a currently quarterly reporting lender no longer happens as a data revision later...

EDIT: Just seen FreeTrader's comment

Edited December 10, 2010 by koala_bear