Saturday, February 1, 2014

Average Home Prices and the CPI

A first look:

Graph #1: House Prices (blue, left scale) and the CPI (red, right scale)
Quick and dirty, Graph #1 is, with the two datasets on two different scales. I prefer 'em on one scale. I multiplied the CPI by 1200 and put it on the left scale together with house prices:

Graph #2: House Prices (blue) and a Scaled-Up CPI (red) on the Left Scale
Aha! The red line goes up more slowly on Graph #2 than Graph #1. The second graph presents a more honest picture. Both datasets on the same scale gives a more honest picture. But we can do better. We can get more interesting. (Well, I think interesting.)

For Graph #3 I divided new house prices by the CPI. Where the line is flat, the two are increasing at the same rate. Where the line is going up, house prices are going up faster than the general price level. And like that:

Graph #3: House Prices relative to Consumer Prices
Some time around 2005 you can see the housing bubble on Graph #3 just as you saw it on Graphs #1 and #2. But #3 shows similar increases in the late 1970s and again in the late 1980s. We didn't see increases like that on the earlier graphs.

Now that's interesting.


Jazzbumpa said...

What you have in graph 3 is inflation adjusted new house prices.

I've become increasingly skeptical of inflation adjusted anything. Who's to say CPI is the right index?

Compare it with the GDP deflator. Not huge differences, but if the data set went back to the 50's, I'm sure you see an expanding gap.

Looking at YoY % Change tells quite a different story.

Would you have imagined a down-sloping trend channel? But looking at the log of the data really tells the same story, with the slope dramatically decreasing over time.

Bottom line - I don't know how to look at data any more.

JzB the puzzled trombonist

Jazzbumpa said...

Left out the log scale link

The Arthurian said...

"What you have in graph 3 is inflation adjusted new house prices."

Yeah, that's one way to look at it: inflation-adjusted or "real" house prices. I don't particularly like that way of looking at it. I prefer the idea of simply comparing house prices to prices in general. There's less Friedmanesque baggage attached.

The difference is not in the calculation, but in the way one chooses to use or abuse the numbers.

I still think it's interesting that Graph #3 shows what looks like three housing bubbles.

And yes, I did imagine a downsloping trend. I thought the increase of the third peak, relative to its starting level, was less than the increase of the first & second peaks, relative to their starting levels.

But I didn't think of a way to show it. You did.