The "long-term Average"

[bob2]

When people talk about house prices having increased "above the long-term average" - exactly what long-term average do they mean?

As far as I can tell, people typically mean the average up until the start of the current "bubble". But why arbitrarily exclude recent years from the average?

Excluding recent years is a good way to demonstrate the existence of a "bubble".

Including recent years is a good way to demonstrate that they're isn't a "bubble".

I guess people will chose whatever average best suits their vested interests. The graph below nicely illustrates how the choice of period over which to calculate the long term average affects how things appear.

[BandWagon]

That's a new argument I haven't seen before on this site.

You could have done the same at the top of the dot-com bubble, and shown that there was no bubble in 2000.

But you'd have a tough time getting people to believe that now.

Edited March 5, 2006 by BandWagon

[bob2]

That's a new argument I haven't seen before on this site.

You could have done the same at the top of the dot-com bubble, and shown that there was no bubble in 2000.

But you'd have a tough time getting people to believe that now.

Similarly there have been times at which, by picking a certain long-term average, it might have been possible to claim that a "bubble" in a market was about to burst - and it didn't.

If you pick the right long-term average you can show pretty much whatever you want.

One thing that concerns me about using a long term average that doesn't include recent increases in house prices is that it, to a certain extent "begs the question".

If, for example, you exclude 2000-2005 when calculating the long-term average because you consider this period to be an anomaly (e.g. a "bubble") it should not be surprising that you then come to the conclusion that, compared to the long-term average you've chosen it looks like an anomaly (e.g. a "bubble") .

[Henrik]

These are "actual" prices, right? Could we have one adjusted for inflation also (and one with average earnings / average house prices)?

Also, what happened in 2000 for prices suddenly moving away from the "turquoise" average? Cheap credit?

Edited March 5, 2006 by Henrik

[bob2]

These are "actual" prices, right? Could we have one adjusted for inflation also (and one with average earnings / average house prices)?

These are the Nationwide "actual prices".

I don't have the data to adjust these for inflation, or earnings.

But you don't need those data to appreciate the point I'm trying to make - i.e. the importance of the period you chose for your long-term average.

(Both long-term average lines would get adjusted for inflation in the same way - and would still lead to different interpretations of the current situation)

[Henrik]

These are the Nationwide "actual prices".

I don't have the data to adjust these for inflation, or earnings.

But you don't need those data to appreciate the point I'm trying to make - i.e. the importance of the period you chose for your long-term average.

(Both long-term average lines would get adjusted for inflation in the same way - and would still lead to different interpretations of the current situation)

Here's the same graph but with price / earnings (using the same data).

I conceed that you are right about chosing the sample data to fit what "you want to prove", but I'm not sure if the interpretations would still be the same if the inflation say took off (or was a lot less) in the period from 95 to 2005 compared to the period before. It doesn't really matter for your point though

However, I thought that when people on here say that the prices are above the long term average, they speak about the price to earnings ratio (or indeed the cost (repayment + interest)) as that is a good indication of what people can afford to pay? Or maybe it's the moving average?

(I've attached a graph detailing p/e from 53 onwards)

Also, if you take any year and calculate a rate of change from that year to "now", you will always end up at "today's" price if you take the original price "then" and multiply it by rate_of_change^(year_now-year_then), as that is the very definition of it. I.e that argument could be used to say that every non-burst bubble is not a bubble (as the "smoothed" price would not be above the current price "now", ever), if you see what I mean.

[BandWagon]

These are the Nationwide "actual prices".

I don't have the data to adjust these for inflation, or earnings.

But you don't need those data to appreciate the point I'm trying to make - i.e. the importance of the period you chose for your long-term average.

(Both long-term average lines would get adjusted for inflation in the same way - and would still lead to different interpretations of the current situation)

If you go to the download page you will find "UK House prices adjusted for inflation", ONS Rpi adjusted.

You will also find another spreadsheet, "UK House prices since 1952", and there you will find an adjustment for earnings inflation.

By your argument you could take the period from 1995 to 2005, see that prices have grown enormously, and that we are now well below "trend" growth. And you'd be right, for that period.

Amongst my friends (early 30's) they have all only seen this period of growth, and without doing further research, and from their experience, they expect this growth to continue.

Hence I was told I was "mad" for selling in 2004.

Yet this doesn't relate the price of a property to the economic service that it provides. This can be easily quantified by the rents the property can command.

Try creating a graph of house prices to rents, that will tell a very interesting story.

[bob2]

Sorry I could have been clearer about what I meant by long-term average.

If you look at a long enough period of time (e.g. 1973-1995 or 1973-2005) there has been an increase in house prices (both actual and inflation adjusted).

From the size of that increase you can calculate an average annual increase for that period.

For 1973-1995 this is 8.1%, for 1973-205 this is 9.1% for actual prices (obviously both much lower if you adjust for inflation).

Neither estimate is just based on a period of growth. The average increase in each case, is an average increase that has occurred in spite of any falls.

Durch - that graph is interesting but not really relevant to the point I'm trying to make.

I'm not trying to argue that a crash is not imminent, or that a crash is imminent. Rather I'm just trying to show how your choice of average determines your conclusion.

For example, the first graph in the article on Why you should NOT buy a house In the UK now looks like it has been constructed using an average increase that was calculated by excluding recent increases.

[Henrik]

If you look at a long enough period of time (e.g. 1973-1995 or 1973-2005) there has been an increase in house prices (both actual and inflation adjusted).

But between 73 and 95 there was not an increase in p/e ratio, in fact there was a significant reduction. I personally think that it's what people can afford to pay what matters, but this is totally irrelevant to the point you are making

Neither estimate is just based on a period of growth. The average increase in each case, is an average increase that has occurred in spite of any falls.

...

I'm not trying to argue that a crash is not imminent, or that a crash is imminent. Rather I'm just trying to show how your choice of average determines your conclusion.

But my point is that the whole idea of using the "average increases" method is flawed, as the "then to now" graph will always end up at the price "now" (by definition), so it could be used to say (at any point in time) that the prices "now" are not in a bubble.

I guess what I'm trying to say is that you are right about chosing the time period for average increases is important for the outcome, but that the whole idea to prove bubbles or not using such a method is questionable (but I'm happy to be proven wrong )

For example, the first graph in the article on Why you should NOT buy a house In the UK now looks like it has been constructed using an average increase that was calculated by excluding recent increases.

To me it looks like the p/e ratio? (We're talking about http://www.firsttimebuyerhelp.co.uk/upload...ngs19822005.gif, right?)

[theChuz]

When people talk about house prices having increased "above the long-term average" - exactly what long-term average do they mean?

As far as I can tell, people typically mean the average up until the start of the current "bubble". But why arbitrarily exclude recent years from the average?

Excluding recent years is a good way to demonstrate the existence of a "bubble".

Including recent years is a good way to demonstrate that they're isn't a "bubble".

I guess people will chose whatever average best suits their vested interests. The graph below nicely illustrates how the choice of period over which to calculate the long term average affects how things appear.

good first post bob and welcome

[Warwickshire Lad]

As far as I can tell, people typically mean the average up until the start of the current "bubble". But why arbitrarily exclude recent years from the average?

If you look at average prices from 1952-2001 which is a good 50 year sample of prices, then it's not unreasonable to think that prices should fluctuate along the average baseline established over that 50 year period.

Everyone knows prices are almost wholly disconnected from their fundamentals and they will come down one way or the other.

[LazyDay]

I guess people will chose whatever average best suits their vested interests. The graph below nicely illustrates how the choice of period over which to calculate the long term average affects how things appear.

I don’t think you did the right averaging here. What you actually did, you calculated the average rate of growth over 1973-2005 and plotted the resulting exponential smoother to the actual price graph, that’s fine. However, this does not have anything in common with the moving average graph which is usually referred to when somebody tries to compare actual prices and some form of average.

The way you do your graph, exponential smoother 1973-2005 will ALWAYS meet the actual prices in the beginning and in the end of the period, by construction. This is the nature of this smoother. This is not the way to gauge for the over/undervaluation, I am afraid.

What you have shown with your graph is the gap in prices that has occurred due to the accelerated growth in recent years. Recent growth was above the long-term average, hence the bubble.

P.S. Henrik, I think, is making the same point here.

[Henrik]

I don’t think you did the right averaging here. What you actually did, you calculated the average rate of growth over 1973-2005 and plotted the resulting exponential smoother to the actual price graph, that’s fine. However, this does not have anything in common with the moving average graph which is usually referred to when somebody tries to compare actual prices and some form of average.

The way you do your graph, exponential smoother 1973-2005 will ALWAYS meet the actual prices in the beginning and in the end of the period, by construction. This is the nature of this smoother. This is not the way to gauge for the over/undervaluation, I am afraid.

What you have shown with your graph is the gap in prices that has occurred due to the accelerated growth in recent years. Recent growth was above the long-term average, hence the bubble.

P.S. Henrik, I think, is making the same point here.

Ah, this is exactly what I was trying to say, but put into human readable form

[bob2]

I don’t think you did the right averaging here. What you actually did, you calculated the average rate of growth over 1973-2005 and plotted the resulting exponential smoother to the actual price graph, that’s fine. However, this does not have anything in common with the moving average graph which is usually referred to when somebody tries to compare actual prices and some form of average.

Yes - there are probably better ways to model the trend. Using the average annual increase over a time period is perhaps simplistic. But that is exactly what seems to have been used in the second graph in the article Why you should NOT buy a house In the UK now.

I'm just trying to show you can paint a different picture.

Lazyday wrote...

Recent growth was above the long-term average, hence the bubble.

My point is - it all depends how you define the long-term average.

Whichever definition you use - yes, recent annual % increases have been larger than the long term average % increase.

However, whether or not you consider current prices to be far above where you'd expect them to be depends on what long-term average annual % increase you use to generate your expectation.

This will determine whether you view the 1989-1995 fall as a "correction" and the 1996-2004 rises as a bubble - or the 1989-1995 fall as an anomaly (what's the opposite of a "bubble"?) and the 1996-2004 rises as a "correction".

Edited March 5, 2006 by bob2

[Guest Riser]

Yes - there are probably better ways to model the trend. Using the average annual increase over a time period is perhaps simplistic. But that is exactly what seems to have been used in the second graph in the article Why you should NOT buy a house In the UK now.

I'm just trying to show you can paint a different picture......

I agree that the long term trend line is open to some interpritation depending on what period you select. However, Nationwide use an inflation adjusted Trend of 2.4% per annum from 1975 Q1 to present which gives a current trend price of £112,500 indicating that properties are currently overpriced by around 40% which is consistent with the following chart and agrees with IMF warnings that property in the UK is overvalued by at least 30%.

UK property overvalued by 40% - Negative HPI sell signal for property.

Last years prediction still on track

[bob2]

I agree that the long term trend line is open to some interpritation depending on what period you select. However, Nationwide use an inflation adjusted Trend of 2.4% per annum from 1975 Q1 to present which gives a current trend price of £112,500 indicating that properties are currently overpriced by around 40% which is consistent with the following chart and agrees with IMF warnings that property in the UK is overvalued by at least 30%.

Thanks for clarifying that. I was wondering what trend they were using.

[RightToExistInASpace]

Does anyone out there have a graph showing monthly mortgage payments on an "average" house as compared to average salaries?

I know "affordability" means nothing if interest rates go up, but this would be interesting to see in relation to the affordability argument that some put forward.

[bob2]

Does anyone out there have a graph showing monthly mortgage payments on an "average" house as compared to average salaries?

I know "affordability" means nothing if interest rates go up, but this would be interesting to see in relation to the affordability argument that some put forward.

One other small point about averages.

People often talk about whether someone on an "average" income can afford the "average" house.

But "average" does not equal "typical".

Average incomes are inflated by high earners with the consequence that most people earn less than the average.

Similarly, average house prices are inflated by expensive properties with the consequence that most houses are cheaper than the average.

Edited March 5, 2006 by bob2

[munro]

One other small point about averages.

People often talk about whether someone on an "average" income can afford the "average" house.

But "average" does not equal "typical".

Average incomes are inflated by high earners with the consequence that most people earn less than the average.

Similarly, average house prices are inflated by expensive properties with the consequence that most houses are cheaper than the average.

In principle this is true, but I think it's true to say that Rightmove et al take out the very high-end properties (anything more than 3 SDs up - ?) whereas no such adjustment is applied to wage figures. This is why the ONS say specifically on their website that recent average wage figures are increasingly skewed upwards because of a small number of very high earners.

Thinking about it this is a good reason for being even more wary of average wage/average house price comparisons!

[AteMoose]

Welcome bob

unfortunatly i cant accept the higher line is the average, as you can see your average line is exponential, which means house prices at some point will be infinatly high against incomes. Extrapolating your average line up the average house price will be 250k -> 300k very shortly, but wages will continue to be 22k.. in another few years average price will be 500k but average wage will only be 25k, an exponantial average line isnt sustainable....

[geneer]

Indeed, why not take an average for the last year and show how house prices have never increased!

Huh.....what do you mean thats not long term?

[LazyDay]

One other small point about averages.

People often talk about whether someone on an "average" income can afford the "average" house.

But "average" does not equal "typical".

Average incomes are inflated by high earners with the consequence that most people earn less than the average.

Similarly, average house prices are inflated by expensive properties with the consequence that most houses are cheaper than the average.

Yep, that's why, as I have noticed, they prefer to follow the median rather then the mean figures for house prices in the US, for example. "Typical" = "median" in skewed samples.

[bob2]

Welcome bob

unfortunatly i cant accept the higher line is the average, as you can see your average line is exponential, which means house prices at some point will be infinatly high against incomes. Extrapolating your average line up the average house price will be 250k -> 300k very shortly, but wages will continue to be 22k.. in another few years average price will be 500k but average wage will only be 25k, an exponantial average line isnt sustainable....

Errrr...both lines are exponential functions. Price inflation is an exponential process. As is wage inflation.

[Ngugi]

Rather than look at the average, I have seen some pundits look at the line which joins up the level of the troughs. This apparently increases at a similar rate to the growth in GDP which suggests that house prices have always naturally corrected to some level related to GDP.

[AteMoose]

Errrr...both lines are exponential functions. Price inflation is an exponential process. As is wage inflation.

I agree, wage inflation is exponential 2.5%, your average house price graph is ALOT higher, the result is houseprices accelerating away from wages exponentially ad infinitium, which obviously wont happen... Otherwise according to your graph average houseprice will be 1 million within 5 or 6 years, and average salary will only be 25k......

Edited March 6, 2006 by moosetea