Monday, May 26, 2008

Sacramento Regional Real Estate Trends for May 25, 2008

No big changes to report this holiday weekend, so just the graphs. If you are desperate for some reading material this Memorial Day, may I suggest the following:

The Taking of Pointe-Du-Hoc on D-Day
The Battle At Qala-I-Jangi in Afghanistan
Operation Phantom Fury in Iraq
Battle of Mogadishu in Somalia












11 comments :

Anonymous said...

CR posted a correlation between months of inventory vs price movement. Any chance there is enough data for the sac area to plot a similiar graph?

Max said...

I could, but this is a rare instance where I disagree with CR. I don't think an R2 of 0.5 means anything.

The problem is he (and I) just don't have a large enough dataset to do this kind of analysis.

DrDoom said...

Max:

Why do you disagree with CR on inventory vs price? The graph speaks for itself...looks correlated to me.

CR shows an R2 of .665 which is a 66% correlation. Fairly strong for the multiparameter equation we call the real estate market. Where did you get CR saying R2 was .5?

Also I sent you local Sacramento data on similar stuff with an R2 of .85 (using absorption which is better parameter when inventory is high and sales are low).

Shouldn't we post the local data and see what the comments and feedback are?

Max said...

CR shows an R2 of .665 which is a 66% correlation. Fairly strong for the multiparameter equation we call the real estate market. Where did you get CR saying R2 was .5?

Sorry, I meant to say 0.54. Statistically speaking, 50-66% correlation means 34-50% non-correlation. The data just doesn't mean that much.

I'm not saying there isn't a relationship, just that it's not quantifiable with the limited available dataset.

DrDoom said...

Max:

A 66% correlation says the single most important parameter in describing price is inventory supply. Not the only parameter just the most important. It is 2/3 th of the story. Everything else adds up to around 1/3.

Can you show any other parameter that has a better correlation to price? How about interest rates, unemployment rates, average age of state employees?

Please show me some data or link to anything that has a stronger correlation to real estate prices. I would love to see it.

Also I still don't get the .54. When I went to calculated risk it should a r2 of .665. Where is the .54 number coming from?

Sippn said...

Of course there's a relationship... but is it a cause-effect relationship? NO.

Poor lending practices ultimately caused both.

Ice cream and crime rise and fall together also, which should we ban?

Correlation just tells you if they rise and fall somewhat together, not cause and effect.

Max said...

Where is the .54 number coming from?

On the post that I linked to, he compared months of supply to CPI adjusted Case-Shiller with a 0.54 correlation. It's right there in the post.

Can you show any other parameter that has a better correlation to price?

We we're talking about price changes vs inventory, and a 0.54 correlation is just not that meaningful. You can compare anything in the world to anything else, but Correlation does not imply causation.

That being said, I think you could say in general terms that the higher the inventory, the more likely prices will drop. But you can't be sure they will drop, and you certainly don't know by how much.

This is basic supply and demand theory, which you don't need graphs or R2 values to prove.

Max said...

Sippn said it much more succinctly than I did. :)

DrDoom said...

Sippn:

Who said anything about cause? Inventory level is result of other factors in existing home sales. In new home sales it is a free parameter (for the builders) but not in existing home sales. The point of studying inventory levels is it can tell you what is going to happen down the road. However, we can't learn the lessons if we don't work on the data.



Max:

Current CR (calculated Risk) link (today) has these words and corresponding graph about 4 paragraphs down the post:

"The best linear fit has been added to the graph (plus the formula with an R2 of 0.665)."

You must be on another link or post. Can you point me there?

I implore you to spend some time on the inventory question. I believe it can shed some light on the future.

DrDoom said...

To clarify, there are several CR postings from 5/27. Here is a link to the CR analysis on MOS and Case-Shiller (prices off 21%):

http://calculatedrisk.blogspot.com/2008/05/case-shiller-real-prices-off-21-from.html

Sippn said...

"Sippn: Who said anything about cause?"

You did...

"A 66% correlation says the single most important parameter in describing price is inventory supply. Not the only parameter just the most important. It is 2/3 th of the story. Everything else adds up to around 1/3."


My boys get up in the morning to a 90% correlation. The cause is me yelling at them. That is the indicator you would need to watch.

Correlation is not exclusive. More than several items can have 66% correlation.


Just giving you a hard time :)

HEY, none of us really knew except the one guy who sold CDOs short and made billions. Finding causal (not casual) relationships IS tough, ask all those other PHDs in DC why they didn't see it coming.

Something interesting.... why oil price is climbing... weeks of explanation about tight future oil supply vs growing China and India, etc... today, one analyst said ... its the flight of funds from real estate lending looking for new returns until real estate lending settles down again. The investor that figures out the real reason for the oil price climb will be the one that gets out with their gains...