Applying Case-Shiller Data to Price Predictions

[RottedOak]

The Case-Shiller Home Price Index is one of the most interesting sources of data on home prices available to the general public. Their methodology for tracking home resales is much more sophisticated than the commonly reported MLS numbers. And perhaps most important for this forum, their data on the 1990s San Diego housing bust may provide some useful insights into the near future. The analysis below reviews some key points from the Case-Shiller data for San Diego and applies the findings to future pricing predictions.

Before the most recent boom, the last high point for San Diego prices was in July 1990. From there, prices proceeded to decline, though not consistently, until a bottom in March 1996. That it of course just under six years later. In terms of nominal (not inflation-adjusted) prices, that bottom was 17.2% below the preceding peak. Looking backwards at the run-up before July 1990, the lowest point in the bust was equivalent to the prices from January 1989, 18 months before the peak. After March 1996, prices then began an upward trend and regained their July 1990 levels in August 1998.

Looking at this prior boom and bust as a model for the present leads to some interesting projections:

• The latest peak was November 2005. The six-year decline from the prior bust suggests that we will hit bottom sometime in 2011.
• The pullback in prices was much slower than the rise. It took almost six years to lose just 18 months worth of price gains. Applying that model suggests that the bottom for the current bust will resemble prices from early 2004. If the percentage decline for the last bust is applied, this also produces a similar result: the closest match is April 2004, when prices were 18.7% less than their future peak.
• Even allowing for a somewhat larger decline doesn’t push prices back much farther in chronological terms: a 25% decline would correspond to prices from January 2004.
• Month-to-month price increases occurred throughout the decline, usually (but not always) during spring or summer months. These did not lead into a long-term rebound until 1996.
• Anyone buying in 2007 should plan to hold their home until at least 2012 unless they want to lose money. If they put 10% or less down, they will be “upside-down” for much of that time.

This analysis also suggests some things about alternative projections:

• Claims of imminent recovery are highly implausible. We are only 15 months from the peak, and prior experience suggests we will have to wait over four times as long to reach bottom. Even if month-to-month price increases occur, it is unlikely that these signal a longer-term rebound.
• Those predicting a 20% decline in nominal prices have a firmer historical basis for their claims than those predicting declines of 30% or more.
• Those expecting “2003 pricing” are in effect predicting 30-35% declines. Such predictions do not have a firm footing in the prior data.
• Those predicting “2001 pricing” and the like are in effect predicting declines of 50% or more, predictions which have even less foundation.

I am not saying that the more dire predictions can’t come true. The boom was bigger this time, so perhaps the bust will be as well. But the evidence to support it seems to be lacking.

The Case-Shiller data for San Diego can be downloaded by following the link at the top of this post (links to the data files are at the bottom of the page that I’ve linked). Please note that the index values reflect nominal prices, as does all of my analysis above. Predictions based on inflation-adjusted “real” prices would require predicting inflation over the next several years, which I have not attempted to do.

[privatebanker]

I agree. We’re not comparing apples to apples here. We have just went through possibly the biggest credit bubble this country has ever seen. It’s starting to fizzle more and more. There are an enormous amount of ARMs resetting in the next few months. Many of these homeowners are scrambling to refi and appraisals are coming back well short. The no money down buyers are getting creamed right now. This enormous credit bubble has created several by-products. The largest being the housing bubble. The ’90s didn’t have the run up in values like we had the last few years. Many homes have appreciated over 300%. How is this a fundamental factor? Many homes in La Jolla on Soledad were selling between $300k – $500k just a few years ago. They are now selling for $1MM+. Typically, when an asset class deviates far out of it’s standard rate of return it eventually overcorrects before arriving back to it’s mean rate of return. I am a firm believer that prices will drop by a large margin over the next few years.

[Anonymous]

I agree with kev and pb.

RO, do some fundamental analysis, comparing home prices to income, rent, etc., and you’ll see that this is a much different animal than the early ’90s collapse. This is a once in a lifetime drop, upcoming.

Figure out how to survive the next five-seven years so that you’re in a position to buy at rock-bottom prices thereafter.

[kev374]

the early 90s model cannot be applied to the current downtrend because this one has it’s own unique set of circumstances. The subprime and exotic loan meltdown is a HUGE factor in this decline and it was hardly present in the last bust. The price appreciation from 2003 onwards is purely due to exotic lending and a large majority of these homeowners will lose their homes when their exotic loan resets. Also we have a much higher percentage of cash-out refinancings in this cycle compared to the early 90s. This is another aggravating factor.

I subscribe to “the higher it goes the harder it falls” point of view. I think this downturn will also be the same timespan of 5-6 years but we will probably see a much sharper contraction earlier in the cycle due to the over leveraged post 2003 buyers foreclosing en masse.

Just my $0.02

[RottedOak]

Although it is certainly possible that the drop will be bigger this time, what I’m hearing in the responses so far is just “this time it’s special,” with no strong basis for believing it. I do not consider it convincing when someone makes general statements about ARMs, how much prices have risen, etc., and then moves from those disconnected references to a specific prediction about future SD home prices. It sounds rather similar to all the special pleading from those who argued that prices wouldn’t drop because everyone wants to live in SD, we haven’t had big job losses like in the 90s, etc. In this case the exceptionalism is just aimed in the other direction. I don’t give either side a free pass on making a case that is backed up with sound analysis.

Any analysis based on a faster decline (as suggested by Kev) requires some type of assumptions about what the decline curve would look like. Any future projection based on price/rent or price/income ratios (as suggested by jg) requires making an additional projection of how rents and/or incomes will change. I don’t have any methodology I consider especially trustworthy for projecting these items. However, I did an off-the-cuff analysis (borrowing some assumptions about income from Rich’s own analysis in “Risks of a Serious Home Price Decline“) of a price/income-based model with a somewhat accelerated decline. This analysis suggested a bottom in 2015 at about 30% below peak, which corresponds to mid-2003 prices.

This latest analysis is a bit fast and loose for my tastes and I’ll see if I can come up with something more supportable. But even this approach comes up much less drastic than some predictions, such as those calling for 40-60% declines, pre-2003 pricing, etc. If someone who believes in these types of predictions has a well-grounded methodology for supporting them, I’d love to hear an explanation. But if you want to convince me of something, please heed what I said above: a wave of the hand in the direction of ARM resets, etc., is not a well-grounded methodology.

I want to be clear that I’m not arguing in favor of buying now, “soft landing,” etc. What I am doing is challenging those who declare a belief in a historically unprecedented drop to back up those predictions with more than “it’s different.”

[lurkor]

“Although it is certainly possible that the drop will be bigger this time, what I’m hearing in the responses so far is just “this time it’s special,” with no strong basis for believing it.”

This boom was way bigger than the 80s boom… so why should the bust not be bigger than the post-80s bust? It seems like common sense; an assertion to the contrary requires backing. The second analysis you did makes more sense because it gets back to actual fundamentals as opposed to an arbitrary percent decline.

[blackbox]

Hey RO,
Unless you came out with a prediction using your model 6 to 10 ten years ago on how high real estate prices would go…..your current prediction is useless. You want to use logic, math, and fundenmentals to predict the downside when all that took a back seat on the upside. Market psychology and easy credit caused the real estate market to overshoot on the upside, and market psychology and tight credit will cause the real estate market to undershoot on the downside.
That’s my prediction, and it’s just about as good as yours, because your so-called model didn’t call the upside, and I don’t think anyone called the upside because earnings, affortablity, and logic stopped existing for at least 3 to 4 years…..
Nasdaq went to 5000 without any fundementals, and it went down to 1100 without any fundmentals. It finally settled at around 2500 because fundementals took over. Let me ask you something. Had any american market index ever dropped almost 80%? Nope, why…because a market index had never jumped so high in such a short period, and fundementals like PE, cash, margins, and inventory turns stop being taken into account.
That is why it is special, and your predictions are pretty much a guess backed with by normal market historical stats and data points. That’s pretty much useless until the undershoot occurs. It might end up going down only 20% from the peak when it settles, but it will go lower percentage wise if only for a short while. That is my guess, and that is exactly what it is… a guess. Too bad you can’t add market psychology as a mathematical constant in your equations. On the way up, part of the market psychology was that people had to buy this month because pricing would go up by next month. The reverse will not take place. Market Psychology will cause people to think that they can hold off a purchase (if they qualify for a mortgage) this month because next month prices will be lower……
In conclusion, your prediction is a useless as anyone’s else. Mine included. Only a look back after the correction has occured will the percentage of the downside be known. Go ahead and use all the historical stat models you wish. I am sticking to
“it’s special this time”. Only time will tell who was right.

[RottedOak]

lurkor,

It is a bit cliche to pull out the old saying, “Common sense is not common,” but I’ll do it anyway. During the boom, lots of people’s common sense told them “real estate always goes up,” “San Diego is a highly desirable place to live,” etc. Common sense may suggest a point to consider — and in fact I acknowledged your own point in my original post — but it isn’t necessarily reliable.

I also want to take exception to your comment about “an arbitrary percent decline.” I did not apply an arbitrary percentage. I applied the historical pattern from the last bust. This might or might not be the most accurate approach, but it is not arbitrary.

[RottedOak]

blackbox,

Agnosticism is an entirely valid stance in situations where there is not enough information. But I do ask for consistency. If someone says, “we don’t have a good basis for making predictions, and therefore I refuse to make one,” then I have no objection. They are being consistent with their agnosticism. If someone says, “we don’t have a good basis for making predictions, so I predict X,” then I object. It looks like the supposed agnosticism is really just an excuse for not having a better justification for the prediction.

By the way, to answer your (presumably rhetorical) question about the stock market: in the stock market crash at the start of the Great Depression, the Dow bottomed at about 89% below the peak. So in fact there was a historical example to use. If you have historical data about a bigger housing bust, I’d be happy to take a look at it. I’m just using the best data I have handy.

[renterclint]

Rotted Oak,

I like what you’re trying to do here. And I agree that there are a lot of folks here who make unfounded predictions throwing out huge % of decline without having a firm argument or explanation of how they quantified such declines.

Maybe those who site the re-setting ARMs & lack of equity can provide a way of applying the impact of these factors. It would be great to hear “3000 SD households or 35% of all current home-owners have less than 5% equity (95% financed), and of this group 85% are facing interest rate reset during 2007.” I would guess this type of info is not available, but if it were we could come a lot closer to having reference data to serve as inputs into your model, and come up with a slightly more meaningful prediction of the housing decline.

BTW, I do not believe that you can compare stock-market busts to our housing situation. Prices are much stickier on the way down for housing. People dump loser stocks pretty quickly – obviously they will not slash the price of their homes in the same manner. Stocks & houses are definitely apples & oranges.

Must-sell housing inventory will certainly put pressure on the whole market, but to what degree really? So many of us are banking on the foreclosures & short sales to drag the whole market down, but what if 90% of the home-owners actually have the ability adjust their personal budgets to hang on and keep their homes for the long haul? They will not sell if they do not have to. Will the other 10% really create a 40-50% in value? I am not so sure.

[kev374]

RO, everything we say right now is obviously a hypothesis based on educated guesswork by looking at historic affordability metrics such as rent/mortgage and prices/income ratios. If we look at previous bust cycles we will see that the ratios have always reverted back to their base values of somewhere around 3-3.5.

Given the theory that history repeats itself we can argue that this time it will not be any different, the relationship between prices and income will revert back to somewhere around 3-3.5.

Inflation adjusted income growth has been nonexistant in the last 5 years. If people are not making more money how are they suddenly richer to pay more for their homes? Especially considering that some fixed costs like education and health care have been skyrocketing. It doesn’t make sense.

As for making assumptions that income growth will be stagnant or will even decline we can make this assumption due to the trend of the last few years. Companies are already aggressively cutting costs or moving jobs offshore, it’s in the news almost every day. Given the increasing availability of talent in Asia it doesn’t make much sense for companies to suddenly start paying MUCH more in the future for US workers.

Now increases in rents are dependent on income, if income cannot rise substantially then I don’t see how rents can go up either.

If you’re saying that by 2015 prices will come down 30% nominal then that is a 60% drop in inflation adjusted terms! (inflation at 4.3% for Orange County). My guestimate was a 40-50% drop in inflation adjusted terms.

[CAwireman]

RO,

I don’t know that I’ve seen your posts here before.
(maybe I missed them)

I welcome thoughts which run contrary to ones presented on this forum.

Also, since Poway Seller stopped posting, its become somewhat boring around here.

Please keep your valuable and interesting comments coming!

[LA_Renter]

I agree with most of what you are saying renterclint. Something you just posted just struck me.

“but what if 90% of the home-owners actually have the ability adjust their personal budgets to hang on and keep their homes for the long haul?”

I notice we talk a lot about NOD’s, foreclosures, REO’s but we don’t spend much time talking about that next category of homeowner that will be able to make it barely. IMO if 90% of the homeowners can adjust their personal budgets to keep their homes it will still have an overall negative impact on the local economy. As these loans reset higher people will have the choice to walk away with blemished credit or hold on with greatly reduced discretionary income. That means no vacations, no new flat screen TV, no new luxury car etc. Even if they hold on the economy will still suffer. The key to determine price declines is demand. If we slip into recession then the possibilities of the large price drops are greater. I guess my point is that we face that risk if these people hold on or not.

[lurkor]

I think it IS arbitrary to say that prices will decline by x% just because last time around they declined by X%. You have to understand why they declined by X% last time, and then apply the same “whys” to this situation given the different inputs this time around. That’s why I thougth the second approach (returning to a fundamental measurement) made more sense.

[LA_Renter]

Irvine as an example. I have to admit I feel giddy about using a city in the heart of the “it will never go down” OC. I have taken a lot of slack form my friends down there. Here is a quote from a recent Bloomberg article I’m sure we have all read about Irvine.

“In Irvine, where just nine months ago office vacancies approached a three-year low, home prices were at an all-time high, and unemployment was less than the national average, at just 3.6 percent, the unraveling subprime mortgage market is ruining the recent prosperity.

Hometown lenders including New Century and Ameriquest Mortgage Co. already have fired more than 3,000 people, house and condominium prices are down 17 percent since June and office vacancy rates are poised to double this year, said John McDermott, regional manager for Orange County at commercial real estate broker Sperry Van Ness.

“It’s a huge engine that has been shut off,” McDermott said. “I don’t know where the new influx of jobs are if you take the lending market out of the equation.”

O.K. this is an acute example of what a Southern California city looks like when you take RE related jobs away. It appears from this reporting you do get the beginning of steep price declines. The city of Irvine looks like it is in localized recession. I’m getting that this is a harbinger of things to come in So Cal. “It’s a huge engine that has been shut off” I think that statement is true of all RE and RE related jobs right now. “I don’t know where the new influx of jobs are if you take the lending market out of the equation.” If you take the word lending and replace it with “RE and RE related jobs” I think that describes the situation facing California’s economy right now. The impact on Irvine is more immediate given the large Corp presence of subprime lenders but it gives a telling view of what things will look like if SD and other So Cal cities go into negative job growth. I don’t know if this is the methodology that RO is looking for but it suggest strong anecdotal evidence of how this will play out to the more severe side. It’s worth watching.

[Author]

Posts