The Census Bureau today reported that new home sales were 3.4% higher in November relative to October, on a seasonally adjusted basis.
When I suggested two months ago that we’d seen the worst for home sales, many or our readers responded with derisive skepticism. But so far, my analysis seems to be holding up. November now marks the fourth month in a row that home sales have come in above the value from last July. Another encouraging sign is that, whereas up to now the revisions of earlier months had always been toward more pessimism, today’s news release revised upwards the estimates of home sales for August, September, and October.
To recap, the basis for my assessment was the belief that low interest rates had been the primary factor driving the housing boom, that higher interest rates had been the primary cause of the downturn, and that the fall in mortgage rates since this summer would begin to make a positive contribution to home sales by November.
Let me also recap what could still go wrong from here. Although the latest data nudge the inventory of unsold new homes down slightly, there remains a significant overhang that needs to be worked off. Ben Bernanke remains dead-on in his analysis of what this means:
[W]e should keep in mind that even if demand stabilizes in its current range, reducing the inventory of unsold homes to more normal levels will likely involve further adjustments in production. The slowing pace of residential construction is likely to be a drag on economic growth into next year.
I expect to see a drop in construction employment, and that could itself generate other problems for the economy, particularly if it appears in conjunction with other bad news such as more job losses from the struggling domestic auto sector.
And I continue to be concerned about whether the housing downturn could lead to widespread bankruptcies or default. Calculated Risk notes one assessment that 1 in 5 of the subprime mortgages originated in the last two years will end in foreclosure. It is difficult for me to be quantitative in predictions about these dynamics, other than to recognize the possibility of some nasty negative feedbacks setting in.
But so far, I think we have to say, it hasn’t happened yet. And a replay of 1994-95 looks like the narrow favorite on which to place a bet.
Technorati Tags: macroeconomics,
housing
Such an impeturbable report James: all of us derisively skeptical commenters (not led by me btprecariousw) who claimed you were not skeptical enough or derisive enough of the NAR narrative, now don’t know whether to bring on the flamethrowers or turn the fire hoses on. (Maybe it is just an isolated post and does not make a trend, so that applying either might be hasty…)
There is nothing for it but to continue with our derisive skepticism: this refinement of the housing data that purports to show some improvement?…is a wash of the underlying problem that fewer people have more money, (the mortgage rates helped a little, only) that affordabilty issues need to be resolved by some Next Big Thing and not the M&A that merely sees more money going into fewer hands.
So the DSS (derisive skeptical squad) are ready with our flamethrowers and fire hoses in the months ahead at the first sign of worsening housing stats.
This looks like the kind of up-blip that can happen in almost any downturn. Very few cycles go monotonically up or down in a perfect pattern. Monthly data is noisy. When I look at your own charts I see no disproof of a continuing, deteriorating trend. True price declines still aren’t measured, and we’ve reached a definite point of exhaustion where some people who cannot sell have withdrawn from the market.
I also suspect that end-of-year bonuses, which are highly skewed to a small sliver of the economy, are actually responsible for much of the pickup in housing stats in the past month. Look at median prices by region for instance; basically all regions are flat to falling still, except the NorthEast, which was slightly up (Goldman Sachs et. al bonuses). As a result, due to the very same volume dynamics pointed out by you on this blog recently, there was an exaggerated jump up this past month in median prices.
I sure hope people don’t make any major, long-term bets on these “soft landing” data points continuing.
Several of my colleagues and I are lurking with our bundles of cash to jump in and vulture feed if and when housing takes a hit in the more desirable areas. That in itself will place a floor under the better locations. Who knows about Modesto, Turlock and Fresno. Lots of land available so back to the fundamentals of cost to build.
Lean heavily on upward revisions to prior months. Those months are big home sales months. Tred lightly on November data. November is a light home sales month, and as yet unrevised.
And, just for balance, note that mortgage applications for purchase fell right back down into the icky range in the latest week, after 3 of 4 really nice weeks. Same problem as with the November home sales data. The past 6 weeks have been so slow for mortgage applications that we shouldn’t put much faith in the data. Better days (from a data reliability perspective) are coming.
Professor are you a homeowner?
Awright, Professor, let’s see you put a ‘happy face’ on today’s resale numbers! How about that fourth consecutive monthly drop in price. I know, I know, prices lag sales. But, in my opinion, there’s nothing to be excited about in minor blips up and down in month-to-month sales. No real trend yet, from my vantage point.
The real fun will come in Jan.-Feb., when folks who’ve been unsuccessful in attempts to sell their homes, and pulled them off the market over the holidays, put them back on the market. Inventory will go up big time (and it’s already happening: per bubbleinfo.com, which comments on north San Diego County real estate, active listings have increased just over 20% in two days, Dec. 26 compared to Dec. 24). Should be an interesting 2007.
Quite interesting that you have a optimistic perspective regarding the data.
However, the data itself is incomplete and as a result, very misleading for the two reasons that have been discussed in abundance:
1. Does not include contract cancellations
2. Does not include incentives
Yes, it can be seen as a leading indicator, but without these two components, its reliability is very questionable.
Better leading indicators are mortgage applications (which have fallen), and building permit issuances (which also have fallen).
Today’s existing home sales numbers are a better indicator as well. Volume pretty much stayed flat, but prices have gone down yet again, in record sequence.
“And a replay of 1994-95 looks like the narrow favorite on which to place a bet.”
Perhaps- but somehow, when I try and swallow the magnitude of this past housing boom cycle (hard to swallow since it’s unimaginably and frighteningly large), I have a hard time believing that a repeat of 1994-1995 will occur.
We are in uncharted territory here, and I really don’t think a 1994-1995 replay is the narrow favorite.
In other words:
Do you believe the BIGGEST appreciation in history will result in one of the SMALLEST contractions?
-OR-
Do you believe the BIGGEST appreciation in history will result in one of the LARGEST contractions?
So, it’s interesting for me to see that you, professor, believe in the former.
Personally, I have a feeling the latter has a better chance of materializing.
MMAfia
JG, my prediction all along has been about home sales, not home prices.
Charts, yes, we are fortunate to own our home. But if I have a bias here, it would be because of my attachment to my research rather than to my home. My research suggests that mortgage rates have a big effect on home sales, but that it takes quite a while for that effect to show up.
We are close to the bottom. The truth is you won’t know until 6 months after it has passed. James, stick to your guns, I loved your original post on calling the bottom but then you almost retracted a little while later after just a few new data points. Your first instint is usually the best right?
Also, there are some really strong correlations between the UK and CA housing market. After a some miserable months in 2004 through 2005, they have had a great 2006 (Apprecation of over 10%). I have posted info from Credit Suisse on this analysis on my blog where they say 99% correlation from 1982 to 2005 with a lag. Perhaps you could review with your trained eye and give some analysis. Sure seems to me we have trailed the UK market by about 15 or 16 months.
Professor, thanks for responding. Ken Fisher, who I think is a smart guy that’s prone to grandiose, flawed statements does bring up a good point about testing research claims against international markets. For example, if something works in the US stock market, but doesn’t work in the Czech and British stock market, that something might not be that valuable as an indicator. If it works in all markets, then it’s probably a great indicator.
This relates to your mortgage rate research because it’s my opinion that the Japanese housing market throws a gigantic monkey wrench into the low rates price support equation. Japanese housing: Sine 1990, nearly 20 years of constant decline of national property values amid super low mortgage rates. Yet we only want to look at American data from that starts after the great depression. We seem to only pick data sets we like. I think there’s huge bias in the mortgage rates are low therefore housing will be supportes argument.
Charts, unquestionably it is relevant that the period I studied (1988-2006) was one of low and relatively stable inflation, lacking the deflation and complete economic bust that are present in your Japanese data. If I thought we were moving into a Japanese-style regime, I wouldn’t want to use such a model for forecasts.
The Chicago Merc trades housing price futures based upon the OFHEO index for selected cities. The most recent quotes indicate the following predicted changes in nominal housing prices between February and November of 2007: San Diego +1.2%; Los Angeles -0.8%; and San Francisco -4.4%. These chnages are quite different from recent historical trends. And these markets are in Coastal California!
Professor,
I’m surprised you look at the seasonally adjusted numbers. Looking at the raw numbers as presented on Calculated Risk’s blog, the numbers don’t look convincing at all. Year-over-year the trend looks terrible since new-home-sales were lower than those seen in 2003. Are you really putting so much faith in a slight seasonally adjusted uptick during the slow sales season?
Professor,
You say that your research shows mortgage rates have a big effect on home sales and that it takes quite a while for that to show up. It would be common sense that declining mortgage rates would cause some people on the side lines to get into the market.
But there is credible evidence (from the data available via ziprealty.com, zillow.com, etc.) that to the extent that there is a slight decrease in inventory, it was caused by sellers pulling properties off the market during after getting frustrated (I followed about 50 such properties where sellers took them off the market).
It looks like that your research is ignoring other factors that affect house prices. Such as the psychology (mania) among consumers between 2001–2005.
In Northern Virginia, in highly sought-after areas like Vienna/Oakton, houses that sold for $300 in 1988 are listed close to $650K now, but houses built in 2000 and sold at $370K are listed at $780K–in other words, in the same area, older houses took approximately 18 years to double (4% appreciation over 18 years) even with all the bubbly (20–30%) appreciation post-2001.
This shows that there had been serious price declines during the housing recession of 1991–1997, a fact that people with short memories don’t bother to look at.
At the current rate of price declines, it will take at least another 5 years for the real estate market to reach a bottom by any fundamental analysis. A good professor should ignore the month-to-month noise and look at all factors that affect real estate on a long-term basis.
D. Rumsfeld, we’ve talked quite a bit about the seasonally unadjusted numbers here as well, and those are in fact what I always use in my own statistical analysis. But if you try to interpret the seasonally unadjusted data on the basis of year-on-year comparisons, you’re guaranteed to be quite late recognizing the turn-around– the year-on-year will stay negative long after the month-on-month turn positive. We certainly can all agree that Nov 06 was much worse than Nov 05. But the question is whether Nov 06 shows the same deterioration relative to Nov 05 as, say, July 06 showed relative to July 05. Here are the year-on-year percentage changes for the seasonally unadjusted data: July -29%, Aug -20%, Sep -16%, Oct -26%, Nov -16%. So, I think the seasonally unadjusted numbers are consistent with the claim that there has not been a significant deterioration since July.
TedK, I’m not making predictions about house prices but rather about the number of houses sold. And what I teach my students (which may or may not be your own philosophy) is that the more variables you put into a statistical model, particularly when those added variables are poorly measured, the more there is that can go wrong. I think there’s a lot of evidence that the more modest your statistical model is, the better it’s likely to perform out-of-sample.
In previous cycles the decline in the number of homes sold has been of greater duration and magnitude than what we have seen so far.
In previous cycles the decline in home prices has also been of greater duration and magnitude than what we have seen so far.
drbrightside, the Belgian economist Daniel Gros has discovered that European housing lags U.S. housing by an average 18-24 months. He thinks this also applies to the U.K. and Australia:
http://today.reuters.co.uk/news/articleinvesting.aspx?type=propertyNews&storyID=2006-11-17T094326Z_01_NOA734602_RTRUKOC_0_EUROPE-HOUSING.xml
JDH, please tell me what do you think of chart 21 in this letter from Gary Shilling:
http://www.safehaven.com/article-6329.htm
Zephyr, take a look at a graph like this one. Evidently you’re not counting 1967, 1984, 1994, and 2001 as “cycles”, in which case your argument is basically tautological. I agree that if we see the level of overall income fall on the magnitude of an economic recession, that will bring house sales down further. But it’s not a foregone conclusion that we’re beginning an economic recession.
Charlie, there’s a lot to respond to in that piece by Gary Shilling. I note that he seems to be primarily talking about prices, whereas I am focusing on sales. But as far as prices are concerned, I think it’s misleading to compare house prices with the CPI. In long run equilibrium, house prices should grow like nominal GDP rather than like the CPI. I also think the very low interest rates of 2002-2005 are a very relevant fundamental. But I share Shilling’s concerns about subprime mortgages.
The cycle patterns are not perfectly smooth – there are occasional off years that reflect temporary volatility rather than a cyclical decline. While that could be happening now, I believe we are in the early phase of a typical cyclical real estate downturn. And these typically have several years of declining transaction volume and declining real prices for existing single family homes. I expect prices to hit bottom in 2008 or 2009. Only then will I start buying again.
Quick Links: 01-02-2007
Happy New Year. Let’s start 2007 with a couple quick links from very late 2006:
Blogging Stocks’ Zac Bissonnette wants you to keep an eye on excessive executive pay. The money going into a CEO’s pocket might be better kept in the com…
Ok, This posting is a bit late….But in any case I think you’re missing the big picture here. Lower interest rates won’t necessarily help housing out of it’s slump. There’s only a finite number of people who can buy a house in this country, and now that number is shrinking with the tighter lending standards and subprime fallout. Add this to the fact that the Affordability Index to own a home is still near all time highs. Real estate historically appreciates 3-4%, but the prior 10 years saw this rate double. There has to be reversion to the mean. This means that housing prices are 30-40% overvalued. It’s estimated that 20% of new home sales in ’06 were speculator purchases. These guys are gone. This is not the “Pause that refreshes”, this is a bursting bubble. The high inventory levels will take YEARS to burn off, not months. THese little blips of improvement here and there are just that. I agree with the first post by Zephyr, but my guess is the bottom will be more like 2009-2010.