Seattle Bubble

There are many arguments to be had between the Seattle Bubble housing bulls and bears, but one belief that seems to be commonly shared is that one of the primary drivers of demand and pricing in our market is a steady stream of rich Californians moving up and driving prices up. The bulls argue that our booming market is fueled by a never-ending flow of California equity, and that this situation is unlikely to end because Seattle is just-so-special. The bears point out that, as the California market slows down, so will the flow of emigrants, which will bring this gravy train to an end.

I counted myself more in the latter camp than in the former. But as someone who is always on the lookout for data to support my viewpoint, I was excited to find out that the State of Washington Office of Financial Management publishes an annual population estimate for the state. In this report, they show immigration trends based on license surrenders by major source of population. The chart below, showing the long-term migration trend from California and Oregon, is the one that caught my eye.

Click to enlarge

The dips and peaks of California immigration shown on this chart looked suspiciously like the oscillations of our local real estate market. So, I put in a couple of calls to Theresa Lowe, the State Demographer - and soon found myself in possession of the data from which this chart was derived.

From there it was a quick effort in Excel to compare this data to the Case-Shiller Index, the longest running and most accurate gauge of price available for this market. I matched the 12 month rolling total of immigration to the year-over-year appreciation in the index, because looking at annualized data should smooth out seasonal variations. The results of this comparison surprised me.

The left hand side of the graphic below shows California immigration versus home prices. To the naked eye, there does appear to be a relationship. Major peaks and valleys roughly align. However, the right hand side of the same graphic, which uses the same data, tells a very different story. This chart shows a scatterplot of the two time series. As you can see, there is almost zero correlation between the two. The coefficient of correlation is negative 3%, and the R-square is practically zero. Based on this data - there isn’t a slight relationship between the two data series, there is effectively no relationship at all!

Click to enlarge

Actually, if you look closely back at the chart on the left- you can see why this is the case. At the beginning of the time series, immigration is still climbing right through the biggest drop in home appreciation. Then for the next 7 years (through 1998), immigration tails off, while the rate of appreciation climbed steadily. Both trended the same way from 2001 to 2006, but more recently immigration has held steady while price appreciation has drastically eroded. The two series don’t move together at major points of change, and for a good part of the series they are moving in opposite directions altogether.

Not ready to give up yet, I thought that perhaps most people are like me - and don’t go to get their licenses immediately after moving. They do important things first, like buy houses. So I experimented with shifting the moving-date data data out further - between one and 12 months. That modification only served to make the explanatory value worse - and if anything, ended up showed a greater negative correlation. (in other words, more Californians equals lower home prices!)

So the relationship between the number of Californians moving here and home price appreciation appears to be a bust. What about immigration at large? I ran the same analysis for all immigration to the state versus home prices to see if that showed any relationship. Here, I found a greater correlation. This time, I got a coefficient of correlation of postive 34% and an R-square of 0.116 - which can be interpreted as “moderate” correlation. There is definitely a relationship, but with such a low R-square -the explanatory value of immigration as a driver of home prices is very low.

Click to enlarge

So it appears, at least based on drivers license data (which should include most, if not all, potential home buyers), there doesn’t appear to be that great of a relationship between immigration and home price appreciation - and that old standby that California Equity is driving our market doesn’t appear to have much substance behind it at all.

The “California Equity Locust” appears to be a mythical beast, whose powers are greatly exaggerated.

Edit: The graph below shows how CA immigration compares to a blended CS index for California, built from weighting SF, LA, and SD in the same way they are in the CS 10 city index. As you can see, there is a very strong negative correlation between these two time series, as many readers have commented.

Tacoma: Your mission, should you choose to accept it, is to sell 1,500 high-end condos in 14 months:

Hundreds of new, pricey condominiums exclude young singles needed for a thriving city core, according to the author of a study analyzing the downtown housing market.

Builders and developers say land costs and water views push prices beyond the mid-$200,000 range generally considered doable for the first-time buyer. And even though the number of unsold units remains high in some neighborhoods, they say demand is strong for high-end condos. Tacoma’s average new condominium price, according to the study, was $348,893 at the end of last year.

The 149-page report, finished last week, identified three kinds of future condo buyers: female baby boomers, young professionals, and married folks with no children at home. It recommended adding edgy lofts and more small spaces that Generation Y buyers can afford.
…
As of December, all six neighborhoods surveyed averaged 14 months of condominium inventory, which measures how long it would take to sell everything built and approved.

A healthy market for new construction tends to be in the six- to 12-month range, said Deanna Sihon, the study’s author.
…
Since 2004, nearly 400 condos have been sold downtown with another 525 for sale and about 1,500 proposed, according to the study.
…
A year ago, a hot market meant condo shoppers had to make rapid buying decisions, said RE/MAX real estate agent George Pilant.

Not so now.

“Buyers have so many choices they don’t feel a sense of urgency,” he said.
…
As in any type of residential real estate, demand is driven by population and job growth, said Paul Turek, an economist with the state Employment Security Department.

But condos are a niche product that at higher inventory levels, he said, raise this question: Will good-paying jobs needed to sell such downtown housing continue to be created?

“I suppose that’s where the gamble is,” he said. “In the Tacoma area, we have some high-paying jobs. Whether there’s enough to support the building of the condos remains to be seen.”

Good luck with that. Seriously.

(Devona Wells, Tacoma News-Tribune, 03.25.2007)

An article in the Times yesterday about the Puget Sound’s job recovery following the dot-com bust got me thinking again about the oft-claimed jobs to home prices correlation. The usual assertion goes something like this: “Prices are justified because our economy is strong (i.e. - lots of jobs), and as long as we keep adding more jobs, home prices will not stop increasing, because more jobs equals more demand.” It’s certainly a comforting belief for inflated housing enthusiasts to hold when the local job situation is on an upswing:

Even though it ended on a somewhat muted note, 2006 was still the best year for job creation in Washington in nearly a decade, according to figures released Tuesday by the state Employment Security Department.

The state averaged nearly 2.87 million nonfarm payroll jobs last year, a gain of 3.2 percent, or 91,500 jobs, over 2005’s average. That was the most nonfarm jobs added in a year since 1997, when 98,600 were created.

The year-end figures also show that 2006 conclusively marked the Puget Sound region’s full recovery, in terms of total jobs, from the dot-com collapse and subsequent recession of the early 2000s.

The article included a nice graph showing the number of jobs in the four-county region since 2000. It has been well demonstrated in other markets that job growth or loss does not directly relate to home prices, but I thought it would be interesting to compare the Seattle Times graph with some data from the Seattle Bubble spreadsheet to see how well the more jobs = more demand = rising prices claim has held up in King County over the last five or six years. To obtain data about the number of jobs in King County I went to Workforce Explorer, the source cited in the Times article.

First I present you with the graph that comes closest to supporting the view that jobs are the primary source of demand.

Click to enlarge

When you compare the percent change year-over-year in both the number of jobs and the median (residential only) home price, the curves actually almost line up, with both job and home price changes being increasingly positive from about early 2003 to the end of 2005. However, the total change in home price increases during that time went from +7.3% to +20.0%, while the total change in jobs went from -1.9% to +2.3%. Despite the similar curves, I think it would be difficult to argue that such a slight change in the job situation drove the major price increases seen over the same period.

Here is same data presented in a slightly different way:

Click to enlarge

At the end of 2006 there were roughly 4% more jobs in King County than January 2000, yet home prices had increased a whopping 85%. I’d like to hear the logic that tries to argue that such a paltry increase in jobs will cause that large of a price increase in homes.

Now let’s take a look at home sales. Supposedly the improving job situation is driving demand, and demand is measured by sales, so let’s see how the two compare.

Click to enlarge

Hmm. It would appear that sales were experiencing the strongest growth during a time when the number of jobs was actually declining. In the summer of 2003 sales were up by as much as 45% over 2002, and yet the job market was still declining by roughly 1.5%.

Looking at the raw numbers of jobs versus sales the disparity becomes even more clear:

Click to enlarge

If jobs are supposed to drive demand, why is it that once the number of jobs began to increase in early 2004, sales actually leveled off? And why have sales been dropping off so steeply in the last year despite what the Times reports as “the best year for job creation in Washington in nearly a decade”? Could it be, perhaps, that the number of home sales and prices of sold homes in fact have very little to do with the number of jobs in a region?

I challenge anyone out there that still believes more jobs = more demand = rising prices to show me the data that supports any sort of correlation between those data sets. Lacking that, I hope we can finally put this dead argument to rest.

All of the above graphs—and the data behind them—can be downloaded in Excel format.

(Drew DeSilver, Seattle Times, 01.17.2007)
(Workforce Explorer, Industry Employment: Historical Series, 01.2007)

Why have residential real estate prices experienced an unusually rapid increase in last few years? That’s the big question that we all want the answer to, right? There’s one argument that goes something like this:

There just aren’t enough homes for everyone. People are moving to the Puget Sound at a rapid pace, and homebuilding just isn’t keeping up. Furthermore, even as more people move here, the size of households keeps shrinking, meaning that demand is increasing even faster! So it makes good sense for home prices to soar and rents to increase, because people have far less choices about where they will live than they did ten or twenty years ago.

Indeed, this would be a pretty compelling argument, if it were backed up by the facts… but is it? I dug through the Census archives to find the answer.

As it turns out, most elements of the above argument are true. Population is indeed rising at a fairly rapid pace. From 1960 to 2000, King County population surged from 935,014 to 1,737,034—an increase of 86%. During that same time period, the average household size dropped 21%, from 3.04 to 2.39. These two statistics combine to give us a 136% net increase in the total demand as measured by the number of households (307,759 to 726,792).

On the supply side of the equation, the number of “housing units” also experienced a greater than two-fold increase (122%), from 333,959 in 1960 to 742,237 in 2000. Of course, 122% is not as large of an increase as 136%, so you can see that from 1960 to 2000, home building did not in fact keep up with demand. This caused the percentage of occupied housing in King County to steadily increase from 92.15% in 1960 to 97.92% in 2000.

This is all very interesting, and so far would appear to back up the “not enough housing” argument. Of course, it is said that the best lies are those that contain the most truth. The real boom in King County home prices didn’t start until after the year 2000. So let’s compare 2000 to 2005*, using numbers readily available directly from the Census website.

In 2000, there were 742,237 housing units available to 726,792 households, for an occupancy rate of 97.92%. In 2005, there were 792,682 housing units available to 747,157 households, dropping the occupancy rate to 94.26%, a level not seen since 1980. Whoa. It would appear that during the five years of most aggressive home price growth, home building has more than kept up with increased demand.

Here is the complete data table for 1960 to 2005:

It should be noted that there are many different sources available for current (2005) population estimates. However, even under the most aggressive of these estimates, the occupancy percentage still declines from 2000 to 2005 (down to at least 1990 levels). I chose to use the data on the Census website since it was most directly comparable to previous Census data, and it is the only source I have been able to locate that contains an estimate of the average household size and number of housing units for 2005.

So what does this all mean? I think at the very least it shows that home building in King County has kept up with demand during the recent housing boom. It seems most likely that building has even surpassed demand by a non-trivial amount. If you have data that shows otherwise, I would love to see it. However, after considering the available data, I believe we can safely bury yet another unfounded argument that attempts to justify today’s housing prices.

*2005 data based on the 2005 American Community Survey, which “is limited to the household population and excludes the population living in institutions, college dormitories, and other group quarters.” Therefore, while total population is likely to appear low when compared directly to Census data, the number of housing units is also scaled down accordingly. Since this post is about housing supply for “households,” the exclusion of group quarters does not affect the final “percent occupancy” calculations.

(US Census Bureau, 2000, 2005)

Here’s a familiar song, courtesy of Tom Kelly at the Everett Herald.

It used to be a popular notion among local real estate agents that the Northwest housing market lagged behind the California market by about six months.
…
I thought about that idea recently when I read that home sales decreased 30.1 percent in August in California from the same month in 2005, the largest sales decline since August 1982.
…
Things are a bit different here, and will continue to be. According to the Northwest Multiple Listing Service, home sales were down about 15.7 percent in September from the same month last year yet prices were up 9.4 percent, marking the first time in two years that year-over-year price growth has not been in double-digit territory in Western Washington.

The premise of this article appears to be that the Northwest only lags California on the way up, but we won’t have to worry about following California down. Let’s see how well the author backs up that claim.

While the past 24 months have been crazy, the long-term outlook for the Puget Sound housing market continues to be bright. Here’s why.

Availability of jobs props up the housing market, and the job outlook for Western Washington continues to be extremely healthy, according to data compiled by Stewart Title Company. In fact, the Seattle-Tacoma-Everett area is expected to add jobs at a rate of double the national average for at least the next three years. While homes might take longer to sell and sellers again are considering offers contingent on the sale of the buyer’s home, local prices are not headed backward or even close to a “soft landing.”

Okay, so our housing market will remain strong because there are plenty of jobs available. But wait, what happened to the California comparison? What does the job situation look like in California? Are jobs not plentiful there? Tom doesn’t say.

Instead, he totally drops the original point he seemed to be making, and closes the article with a series of bold assertions.

“No housing market has ever collapsed unless the underlying economy went sour,” [real estate economist John] Tuccillo said. “Short of recession, this means that virtually every housing market in the U.S. will hold up even though sales may slump and prices decline.” He did note, however, that home prices may slump in upper-Midwest rust belt areas.

What about a worst-case scenario — mass foreclosures and rising inventories?

“If the United States undergoes a recession in 2007, the housing market will do much worse than we anticipate, but so will autos and retail,” Tuccillo said. “Exotic mortgage instruments will have an impact in increasing the foreclosure rate, but in any loan made before 2005, the consumer is in a positive equity position and will weather financial distress.”

So, when your friends in California swear the sky is falling and real estate will no longer be the same, remind them that property is cyclical and that their neighborhood will rebound when the “down” period ends late next year.

And, the down period in the Puget Sound will mean slower, not negative, appreciation.

Sweet. Home prices definitely won’t drop significantly unless there’s a recession, but even if there is one, every pre-2005 loan will be totally safe, and worst case, all the pain will be over by the end of next year. Those are good things to know. I’m glad Mr. Kelly let us in on this reassuring absolute knowledge that he and his real estate economist friends are in possession of.

(Tom Kelly, Everett Herald, 10.15.2006)