More on Decelerating Home Price Declines

Submitted by Rich Toscano on February 26, 2009 - 10:46pm

People seemed to find the rate-of-change graph in the prior post interesting so I thought I'd follow up with a look at how all the individual Case-Shiller price tiers have been trending.

The results are found in the accompanying graph...

continue reading at voiceofsandiego.org

(category: )

Submitted by monaco888 on February 27, 2009 - 9:54am.

Excellent chart, Rich... these kinds of trends make the potential "bottom" much easier to visualize for each bracket.

Submitted by 34f3f3f on February 27, 2009 - 10:40am.

When I first saw this chart I thought, Oh my God! prices are going up, but was relieved that it's just the rate of decline that is easing. They've still some way to go in my view. Let's hope it's not a painfully slow process from here on.

Submitted by sdybob on February 27, 2009 - 12:15pm.

This graph by itself can be very misleading. People looking at this graph thinks things are getting better and home price is stabilizing. The fact is on the contrary, price is actually dropping at an astonishing pace of 20% to 30% a year.

Submitted by Rich Toscano on February 27, 2009 - 2:35pm.

sdybob wrote:
This graph by itself can be very misleading. People looking at this graph thinks things are getting better and home price is stabilizing. The fact is on the contrary, price is actually dropping at an astonishing pace of 20% to 30% a year.

I have a real hard time believing that anyone who actually bothered to read the title at the top of the graph would interpret this to mean that home prices are stabilizing. If they can't be bothered to do that... well, that's not my problem.

Rich

Submitted by Fearful on February 27, 2009 - 4:15pm.

Ha - I dunno, Rich, it is easy to confuse the derivative of home prices with the prices themselves!

This "decreasing rate of decrease" might well be an artifact of the various price decline mitigation measures, such as the foreclosure delay.

However, the overall curve, if it ends up being roughly symmetrical, ought to start flattening out around now.

That is, if it does not overshoot on the negative.

Submitted by no_such_reality on February 28, 2009 - 9:17am.

Rich Toscano wrote:
sdybob wrote:
This graph by itself can be very misleading. People looking at this graph thinks things are getting better and home price is stabilizing. The fact is on the contrary, price is actually dropping at an astonishing pace of 20% to 30% a year.

I have a real hard time believing that anyone who actually bothered to read the title at the top of the graph would interpret this to mean that home prices are stabilizing. If they can't be bothered to do that... well, that's not my problem.

Rich

Like others my initial reaction was low end prices are headed up. I then looked more carefully to see it year over year and this year is 30% down from last year which was 25% down from the prior year which is 2% down.

Cumulatively were 55% down and still going down at 29% a year.

For the bottom to come it, that line has to climb all the way back up to ZERO. Look at the few comments here and you see people thinking the chart shows the bottom right around the corner.

Submitted by Rich Toscano on February 28, 2009 - 10:15am.

Sure, I can see that you'd see a line turning up and one's brain will react by thinking "getting better." I'm just saying that if you look at the title (year over year rate of price change), then you look at the Y axis (having moved from -34% to -30%), it is very clear that it's not "getting better," but "still getting worse, but not as fast as before."

I was mainly objecting to the idea that the chart is "very misleading," which it isn't, because all the info required to properly interpret the chart is contained in the chart.

Rich

Submitted by peterb on February 28, 2009 - 10:51am.

Rich,
Hurry and contact NAR. They'd pay big time for this graphical analysis. It shows how things are getting "better". LOL. What'dya bet we see this from them in the next 30 days??!!

Submitted by davelj on February 28, 2009 - 11:02am.

Rich Toscano wrote:
Sure, I can see that you'd see a line turning up and one's brain will react by thinking "getting better." I'm just saying that if you look at the title (year over year rate of price change), then you look at the Y axis (having moved from -34% to -30%), it is very clear that it's not "getting better," but "still getting worse, but not as fast as before."

I was mainly objecting to the idea that the chart is "very misleading," which it isn't, because all the info required to properly interpret the chart is contained in the chart.

Rich

The graph is crystal clear. Apparently a few folks here don't understand calculus.

Submitted by Fearful on February 28, 2009 - 11:11am.

Rich, I would not change a thing. I was just hoping someone would pick up on the "derivative" comment, and they did. So I am happy.

Here's a thought: Add the longer term, to show how the earlier years had repeated year-over-year price increases, and now we are in year-over-year price decreases. There are many of these sorts of YoY charts out there.

On another point: Is there an easy source for CPI less shelter? I keep my own archive of the Case-Shiller data, and have used overall CPI deflator in the past. However, doing so is a little circular, as shelter (OER?) is part of the overall.

Thanks,

Eric

Submitted by paranoid on February 28, 2009 - 3:12pm.

the graph is perfect.
for those who don't understand what "year-over-year% change" means, I have pity for them. they don't deserve to read the chart.

The only thing I'd ask for is to also include a graph with longer time range, so tha we can see what happened in the past and what's happening now.

Submitted by Nozferat on February 28, 2009 - 3:38pm.

Uhhh..Rich. Think about it for a moment. Look at the guttural reaction of people in the past 5-6 years that brought us into this mess.

So it's no wonder people jump at such data and think things may be getting better.

Needless to say...I think this is going to get worse. There are reasons behind this uptick and it's not because people are in better shape. As prices have been coming down, people have not been getting wealthier poised for purchase. If anything this data is probably getting alot of noise from extraneous sales such as foreclosures, short sales, etc...

We can also think about this graph like what you'd see in a vibration response graph say..when you hit a tuning fork. You tap the fork, it starts ringing. Its response is going to be a diminishing sine wave...and I think this is the same. I'll venture to say we're going to follow with big dip.

Submitted by Nozferat on February 28, 2009 - 3:40pm.

no_such_reality wrote:

Like others my initial reaction was low end prices are headed up. I then looked more carefully to see it year over year and this year is 30% down from last year which was 25% down from the prior year which is 2% down.

Cumulatively were 55% down and still going down at 29% a year.

For the bottom to come it, that line has to climb all the way back up to ZERO. Look at the few comments here and you see people thinking the chart shows the bottom right around the corner.

That depends on what that ZERO is. For it to climb up BACK to zero indicates that you believe the high point was normal.

Submitted by NotCranky on February 28, 2009 - 4:45pm.

The way I understand it, climbing back up to zero means we could potentially be somewhere around the trough region of the cycle ....if we are still talking about realtively normal, other than the amplitude, cycles.

Submitted by all on February 28, 2009 - 4:47pm.

Fearful wrote:
Rich, I would not change a thing. I was just hoping someone would pick up on the "derivative" comment, and they did. So I am happy.

Second derivative would look even better :)

Submitted by madcow on February 28, 2009 - 5:22pm.

Very interesting, we are past the inflection point for the low end and near it for the high end. I have been waiting for this signal, for me there was no point to even look at MLS before this signal. In the case of a sinusoidal-like function, we would be right in the middle right now...

Submitted by Rich Toscano on March 1, 2009 - 11:40am.

Russell wrote:
The way I understand it, climbing back up to zero means we could potentially be somewhere around the trough region of the cycle ....if we are still talking about realtively normal, other than the amplitude, cycles.

Yes, if this hits 0 (assuming it stays there) it probably means the bottom is already in the past. At the peak (see the longer term chart from the prior post), things were still rising at a yoy rate because yoy is lagging. At the trough, things will still be falling at a yoy rate.

Re. the question on inflation, this is the primary source: http://data.bls.gov/PDQ/outside.jsp?surv...

Rich

Submitted by no_such_reality on March 1, 2009 - 12:27pm.

Nozferat wrote:
no_such_reality wrote:

Like others my initial reaction was low end prices are headed up. I then looked more carefully to see it year over year and this year is 30% down from last year which was 25% down from the prior year which is 2% down.

Cumulatively were 55% down and still going down at 29% a year.

For the bottom to come it, that line has to climb all the way back up to ZERO. Look at the few comments here and you see people thinking the chart shows the bottom right around the corner.

That depends on what that ZERO is. For it to climb up BACK to zero indicates that you believe the high point was normal.

No. No. No. No. For that line to climb back to zero means prices are neither going up nor going down from wherever they were the year before. In other words, when that chart reads ZERO, the price this year is the same as the price last year.

So in two years, when that chart crawls back to zero, it'll be where the prices are next year at this time which have all the hallmarks of being another 20-30% down from here.

Submitted by sdybob on March 1, 2009 - 2:03pm.

Rich,

I want to say thank you for doing a great job in putting all these data together. I have really enjoyed reading them. It is obvious now that my comment on the graph came out wrong, to the least it was not what I meant to say. I apologize.

Submitted by Rich Toscano on March 1, 2009 - 9:16pm.

sdybob wrote:
Rich,

I want to say thank you for doing a great job in putting all these data together. I have really enjoyed reading them. It is obvious now that my comment on the graph came out wrong, to the least it was not what I meant to say. I apologize.

No worries, sdybob, it led to an interesting conversation. Thanks for the note.

rich

Submitted by EJ on March 2, 2009 - 10:11am.

Another way to think about the peaks and troughs for this type of chart: each peak and trough indicates a point of inflection in the home price vs time chart.

For the data shown, a trough is pretty well defined (at least for the low end). Prior to this trough the corresponding home price curve was concave-down and negative slope, after the trough the home price curve is concave-up and still negative slope. Prior to trough, rate of home price decline (first derivative wrt time of home price) is increasing. After the trough, rate of home price decline is decreasing.

Either way, until it is above zero, it is all price declines.

Submitted by Fearful on March 3, 2009 - 10:35am.

Rich, thanks for the link to the BLS data. It looks like it is quarterly to semiannually; do you know if there exists monthly data for this region and without shelter? I just interpolated between the data points but was wondering if there is monthly data out there.

FYI, I applied the two different deflator data sets to the C-S data, and found there is not much difference between the resulting curves. With steady house prices, the difference, at several percentage points, would be significant, but the actual underlying change is much larger than any difference between deflators.

Submitted by Eugene on March 3, 2009 - 4:59pm.

We can expect the line to cross zero into positive land approximately 6 months after the bottom.

Submitted by SD Realtor on March 3, 2009 - 5:05pm.

Esmith that is only if the behavior is a single wave or is parabolic. If we have a schizophrenic market with waves of depreciation where each wave is multiple years then we can cross 0 and then hump back down below it again.

Submitted by Eugene on March 3, 2009 - 5:13pm.

Absolutely.

However, we should encounter a local minimum before the curve crosses into the positive land. This fact is known in mathematics as Rolle's theorem.

Submitted by Colombo on March 3, 2009 - 7:15pm.

How does a graph of this type account for the fact that *over time*, as housing prices decline, a home that was in one particular band will find itself in another particular band. In other words, do the definitions of high, middle, and low change? I'm guessing there will always be some fuzziness at the dividing points between the three bands.

Submitted by Fearful on March 4, 2009 - 10:19am.

esmith wrote:

However, we should encounter a local minimum before the curve crosses into the positive land. This fact is known in mathematics as Rolle's theorem.

The shorter the time interval, the nearer you are to the minimum when it turns up. Obvious, if you think about it.

Submitted by Fearful on March 4, 2009 - 10:22am.

Colombo wrote:
How does a graph of this type account for the fact that *over time*, as housing prices decline, a home that was in one particular band will find itself in another particular band. In other words, do the definitions of high, middle, and low change? I'm guessing there will always be some fuzziness at the dividing points between the three bands.

I am not sure how or if this is handled within the Case Shiller methodology. This was discussed here when the tranches first came out, and there was no consensus. I think there is very possibly an error in the historical data, because the historical data are not revised when the tranche cutoff points are changed.

However, the aggregate is not affected, and the various tranche curves do not seem to move that much relative to the aggregate, so the error does not appear to be large.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.