The greatest bubble in UK history is now bursting. To me, the fact that the market was a bubble was pretty obvious, but identifying when it was going to go pop, was nigh on impossible (as is the case with all bubbles). I have been blogging on both the UK and US housing market for a couple of years
http://www.noelwatson.com/blog/CategoryView,category,Housing%2Bmarket.aspx
and it is interesting to go back and see how accurate the betting exchanges such as Spreadfair have been - these were pricing in falls well before an incredulous general public realised what was about to happen.
http://www.noelwatson.com/blog/CategoryView,category,Housing%2Bmarket.aspx
The latest Nationwide numbers came out yesterday showing the (now expected) falls.
http://www.nationwide.co.uk/hpi/historical/Sep_2008.pdf
The thing that interested me was on page three - "Long Term Real House Price Trend". It shows that after the recent falls we have been having, we are almost back to the long term mean, so on first impression this would imply that we haven't got much further to fall. However, this ignores two key points. Firstly, house prices are cyclical, and secondly, I believe the long term trend of 2.9% over RPI has been dragged up by recent rises.
So, how much further have house prices to fall. Up until now I have been reluctant to give predictions, feeling that the betting exchanges (Spreadfair) and property derivative sites were the best way forward
http://www.tfspropertyderivatives.com/pdf/RISK&MANAGE/2008/Sep-08.pdf
but I thought that it would be fun to have a punt and see how it turns out in a couple of years
The most recent graph can be found here (doesn't include Q3, but we can fix this)
http://www.nationwide.co.uk/hpi/downloads/UK_house_prices_adjusted_for_inflation.xls

We can get our Q3 number from the latest report £165,188
http://www.nationwide.co.uk/hpi/historical/Q3_2008.pdf
next we need to decide what is a more accurate real rate of return. For this we can look at the Nationwide archive
http://www.nationwide.co.uk/hpi/archive.htm
and in particular April 2000, as the figure of 1.5% p.a. reflects (in my opinion) a more accurate reflection of growth, as it occurs soon after a big fall preceded by an equally big rise
http://www.nationwide.co.uk/hpi/historical/MPR0004.pdf

Comparing the two graphs, it is interesting to see the trend line for Q1 2000 is around £25k lower than for Q2 2008.
We can easily extend the current 2.9% trend into the future (each cell in column D is 1.007 larger than the one above - 2.84pa - this doesn't include the Q3 numbers to make it 2.9%, but it is close enough)
Next we add our own 1.5% trend. For this, we add multiply each cell in column E by 1.00375 multiplied by the cell above. Note that our starting figure for the 1975 Q1 1.5% trendline is not the same as that for the 2.9% trendline - we need to adjust for the fact that the data series started before 1975. Note how our modified trendline now crosses at around £75000 for Q1 2000, as it did with the April 2000 report
Next is the tricky part. We have to discount our house prices for inflation. The calculations are shown in the spreadsheet. I am assuming an inflation rate of 1.2% a quarter which gives 4.7% per year.
To attempt an accurare looking graph, I have included two variables, initial quarterly fall and decrease on the quarterly fall per quarter. I have estimated these at 10500 and 700 respectivelly.
So, what is the fall from peak to trough. The peak was round £185k, the low on my graph is £83k (2013 Q1) - so we are looking at falls of around 55% - so around 45% to go (from peak).

This may seem like an extreme few, and I may have made a fundamental error in my calculations. However, in the latest nationwide report, it states
“House prices now are over 60% higher in real terms than they were at the start of the decade, even taking into
account the falls since last October"
So if we are around 165k now, this would take us down to 100k. As I said earlier, these things tend to undershoot, so maybe 83k doesn't look so unrealistic.
Spreadsheet attached:
Nationwide2008Q2.xls (121.5 KB)
EDIT: There are still some areas of concern
1. Trend line start dates differ
2. Real house prices are discounted from the present time - therefore they are not comparable for different time periods
Attempt 2:
Changes
1. Trendline starts at same point
2. Monthly trend increased to 1.005 (2% pa)
3. Starting quarterly fall 10000
4. Quarterly decrease 800
We now have a bottom of £97588 in 2011 Q4

Note that this is effectively a hack of a hack, but it is best guess so far.
Spreadsheet attached:
Nationwide2008Q2v2.xls (141 KB)