Clear majority of people support Help To Buy

PaulF81

Graham_Devon wrote: »

Fewer than 2,000 people filling in a survey to get 50p.

Woop.

Proves the media, consumer groups, economists, politicans, business groups and journalists wrong. Infact, I'd go as far to say all these talking heads are completely irrelevant now, thanks to those people claiming their 50p.

2000 in statistical terms is significant. I dont know how exactly to work out the statistical standard deviation for such a survey with 2000 persons taking part, but I know that any survey over around 50 typically yields accurate results if executed well.

Graham_Devon

PaulF81 wrote: »

2000 in statistical terms is significant. I dont know how exactly to work out the statistical standard deviation for such a survey with 2000 persons taking part, but I know that any survey over around 50 typically yields accurate results if executed well.

Not always, especially when you are paying people to take part.

There was some really interesting research done on this. If you are paying people to take part, you can easily get the resuults you want just by ordering the answers.

So for example, if you want people to say "yes, I agree" put the answers in order of:

Agree
Don't agree
Unsure

More people will click agree just to get to the next question and finish the survey.

Got no idea what order these answers were in, just making the point really. That's why various ways of advertising have dissapeared. Paying people to click for instance just got people clicking. Paying people to watch ads just got them playing them. Same with surveys. The take up of unpaid surveys is known to be massively lower.

Some website have had to start using bogus questions simply to force people out of the survey. The question will be something like "what is 2+2". if you click the wrong answer as many do, they are screened out....much to their annoyance as they feel cheated!

PaulF81

Thats why I commented "Executed well". As you rightly say, accounting for variations is key to a good survey. There is a book out by the chap who does the polling in the USA and got 50 out of 50 states right prior to the results being announced. Was an interesting read (better than freakonomics which I also enjoyed).

Edit: just checked my library, its called "the signal and the noise" By Nate Silver.

JonnyBravo

HAMISH_MCTAVISH wrote: »

Here you go Graham.

Was doing it form memory so a percent or two out, but close enough.

Max 65.5%

Min 23.6%

Average 36.3%

Current 28.1%

http://www.lloydsbankinggroup.com/media1/economic_insight/house_price_tools_page.asp

As I said, well below long term average and at near record lows.

In fact, buying a house has only been cheaper in a handful of the last 32 years.

This data is obviously really poor.
It states 99.8% was the peak amount of income used to service mortgages in the SE.
99.8%? Really?
Not accurate IMO

ukcarper

Graham_Devon wrote: »

Yup, also worked out in two different ways.

Note how the figures Hamish linked too instantly wipe off 30% of the house price as they assume a 30% deposit. It then uses sub 70% mortgage rates to do the calculation on full time MENS wages. Sweet.

The one I linked too uses a 90% LTV, using ALL full time wages, not just mens, which seems a bit more realistic.

Little bizare that Hamish screams for higher LTV's when it comes to lending, but uses data based on the very LTV's he moans bitterly about to prove higher affordability.

I think you will find for affordablity they use all they use men for house price to earnings ratio they don't sayif they use mean or median but Nationwide use mean.
I think the difference can be explained by the the fact that the rates on 90% mortgages are a lot higher.

Graham_Devon

ukcarper wrote: »

I think you will find for affordablity they use all they use men for house price to earnings ratio they don't sayif they use mean or median but Nationwide use mean.
I think the difference can be explained by the the fact that the rates on 90% mortgages are a lot higher.

Nationwide just use "full time workers".

That suggests male, female, undecided and anything inbetween.

ukcarper

JonnyBravo wrote: »

This data is obviously really poor.
It states 99.8% was the peak amount of income used to service mortgages in the SE.
99.8%? Really?
Not accurate IMO

It looks like LLyods use average for UK while Nationwide use the region.

Both sets of figures are high for that period.

Seeing as mortgage rate was 14.5% and house price to earnings ratio was 5x I'm not surprised affordability ratio was high.

ukcarper

Graham_Devon wrote: »

Nationwide just use "full time workers".

That suggests male, female, undecided and anything inbetween.

As do LLyods for the affordablity % figures.

HAMISH_MCTAVISH

JonnyBravo wrote: »

This data is obviously really poor.
It states 99.8% was the peak amount of income used to service mortgages in the SE.
99.8%? Really?
Not accurate IMO

Of course it's not accurate for actual house buyers.

None of them are, as they all use the average income for all people, rather than the average income for actual house buyers.

What it does show is the same dataset, over time.

So whether you use a 10% deposit or a 50% deposit, or the income of actual buyers versus all people, is of not much relevance.

What is of relevance is that the calculation is the same over a long period of time, so you get the comparison using the same data in 1990 as in 2007, or in 1995 as in 2013.

And what that shows you is that there are only a handful of years in the last 32 years where buying a house has been cheaper.

If you want actual data on real house buyers, average LTV, average LTI, average price paid, average deposit size, etc, the CML provides that, but the data series is much shorter.

grizzly1911

HAMISH_MCTAVISH wrote: »

Of course it's not accurate for actual house buyers.

None of them are, as they all use the average income for all people, rather than the average income for actual house buyers.

What it does show is the same dataset, over time.

So whether you use a 10% deposit or a 50% deposit, or the income of actual buyers versus all people, is of not much relevance.

What is of relevance is that the calculation is the same over a long period of time, so you get the comparison using the same data in 1990 as in 2007, or in 1995 as in 2013.

So the richest 70% who are the only house buyers argument is irrelevant for comparison purposes?