The annual change in the number of people employed at San Diego companies improved again last month. Between November 2008 and November 2009, local companies shed 52,200 jobs, a decline of 4.0 percent. Which, as miserable as it may sound, is the best year-over-year job number in six months:
The above chart, as discussed earlier this week, denotes the number of people on the payrolls of local companies regardless of where those employees may live. This next chart shows that while the prior series continued to improve year-over-year, its counterpart -- the number of San Diego residents who are employed -- saw its annual change hit a new low:
It is the latter data that is used to calculate local unemployment, which came in at 10.5 percent:
The second chart indicates that San Diego's businesses suffered worse than its residents in 2008. For the time being, it appears that those roles have been reversed.
A commenter over at Piggington (my own little internet stomping ground) raised an interesting point in regard to last month's article on San Diego employment. The reader noted that the monthly data I cite measures the number of jobs held in San Diego with no regard for whether the job holders are actually San Diego residents. There is a separate data series that tracks the employment status of San Diego residents whether they are employed in the county or elsewhere. And, for the time being, the two job surveys are providing mixed signals.
The orange line on the accompanying chart displays the year-over-year rate of change in the number of employed people in the county's labor force -- in other words, the number of people living in San Diego who have jobs somewhere. The blue line denotes the annual change in payroll jobs at San Diego companies, regardless of whether the job holders live in San Diego or not.
As discussed in last month's update, the annual rate of change in the number of San Diego payroll jobs has improved for two months in a row. The chart shows that the same cannot be said for the annual change in number of employed San Diegans, which hit a new low in September.
The two series' behavior was different in 2008, as well. While the annual payroll change slid down in a more or less straight line, local labor force employment actually recovered and went positive for much of 2008. Now, on the other hand, the labor force employment series is comparing to much stronger year-ago numbers, so its annual pace is looking even worse than that of the job series.
I have used the so-called "establishment survey" data (the data reflecting local jobs without regard for where job holders live) in my monthly analyses because it breaks employment down by industry and allows for more detailed analysis. But it is the results of the "household survey" that arguably matter more for the local economy. People tend to spend money where they live, and by definition, they buy houses where they live. It seems that the employment status of people living in San Diego, not the number of people working in San Diego but potentially living elsewhere, would exert the larger effect on local economic activity. And by that measure, it seems that the annual rate of change has not gotten any better after all. Both series tell us something different and both bear watching.
All that said, I must add that I am kind of shocked by and a bit skeptical of the household survey data for 2008. It is hard to believe that employment among San Diego residents was growing at over 1 percent per year at a time when payroll jobs in San Diego were shrinking by over 1 percent and the world as a whole was pretty much plunging into economic oblivion.
That apparent discrepancy will be the topic of a future blog entry.
A while back, I noticed a funny thing about the year-over-year rate of change in the Case-Shiller index of San Diego home prices. It seemed that movements in the annual price change rate from positive to negative or from negative to positive provided a good indicator that the long-term price trend had changed direction.
Allow me to demonstrate with some graphs.
The graph below shows the home price index in blue with the year-over-year rate of change for the index in orange. It covers a ten-year period more or less centered around the early-1990s housing bust.
In the month that the annual rate of price change flipped from positive to negative territory, the price index was down just 3.4 percent from the peak. It would go on to drop a total of 17.2 percent over the duration of the downturn. So the "sign reversal" of the annual rate of change provided a fairly early indication that the long-term direction of prices had switched from up to down.
This indicator worked even better in the more recent boom. When the annual rate of change went from positive to negative, the home price index was down a mere 1.2 percent -- hardly a sliver of the overall 42.3 percent drop that had taken place as of earlier this year:
In the 1990s, this signal worked in the other direction, too. The first month that the annual change rate sustainably turned positive, the price index was only 1.4 percent above its 1996 trough. That 1.4 percent doesn't seem like all that much considering that the index would eventually rise by (are you ready for this?) 251.5 percent.
Note, however, the use of the word "sustainably." In 1994, there was a brief head-fake into positive territory, with the annual change rising at a whopping .1 percent for two months in a row. At that point, over two-thirds of the decline had already taken place, but the foray into positive annual rates was not a sign of a major trend change. So it might be better to say that the trend change was only signaled by a reversal in the year-over-year rate that lasted for several months. Generally speaking, however, the fact remains that changes in the annual rate of price appreciation from positive to negative and vice-versa have provided an interesting tell as to the direction of the underlying trend.
This makes a certain amount of sense. The housing market normally moves like a battleship -- huge and very slow to turn around. A change in the major trend could typically be expected to involve a deceleration, then a period of flattening, and finally acceleration in the other direction. The year-over-year change rate typically might not hit zero until the battleship itself had made some progress in its turn.
I bring this all up now because, if my little Case-Shiller estimation technique is to be believed, we may be just about to go positive in the annual rate of change of San Diego home prices. As of last month, the estimated index value was down just .1 percent from its value a year prior. Given that home prices were falling fast late last year, it's likely that the annual rate will hit positive territory this month.
Here's a look -- note that the estimated index has already risen 10.0 percent from its low point earlier this year:
Interesting, no?
Well, I thought so, anyway. But while interesting, it might also be fairly meaningless. Some reasons follow.
First, the market isn't really acting like a battleship right now. Maybe it's more like one of those pontoon party boats. The upturn in prices is much more abrupt than anything seen prior, as the 7-month, 10 percent rise would indicate. And since the typically battleship-like behavior of the market really provides all the reasoning behind the theory that changes in the rate signal changes in the trend, perhaps that theory doesn't apply this time around.
Second, there is a factor that is unique to this particular housing cycle: the all-out government effort to prop up home prices. Almost by definition, these tactics work by enticing buyers who might have waited for the future to instead buy in the present. This is all well and good for today's prices, but it drains the supply of future buyers. So the typical pattern of a slow transition from weakness to increasing strength could be replaced this time around by more short-term volatility within a long-term trend that is flat or at least less defined.
Third, we are dealing with a very small sample set. And this indicator didn't work well at all in the early 1980s, when prices stayed fairly flat while inflation did the work of reducing home valuations in comparison to rents and incomes. (This, incidentally, is a possible outcome for our market as well).
Sooooo... I don't think anyone should make too much of these graphs and relationships, but I think they make for an intriguing historical comparison. And, like everything else, they bear watching.
The number of existing homes for sale in San Diego hit multi-year lows in October, as the following graph shows:
Despite the decreased availability of homes to buy, the volume of sales was more or less the same as it had been a year prior:
The number of months' worth of inventory, which is arrived at by dividing inventory by sales, also hit new post-boom lows:
That is quite some change over the course of two years -- months of inventory dropped by 65 percent between October 2007 and October 2009. The question about what will happen to all those homes in foreclosure lingers on, but so far, shadow inventory is a no-show.
The month of October provided yet more evidence that the surge in home prices this year has something quite different than the typical spring/summer rallies that take place within longer-term housing downturns.
The median price per square foot of resale properties rose for both single family homes and condos, by 1.4 percent and 2.0 percent respectively. The following graph shows the decline and then eventual rise in this price measurement since the 2005 peak:
If the Case-Shiller proxy discussed in the prior article is correct for October, the index will have measured a 10.0 percent price rise from its April low:
By contrast, the biggest summer rally within the 1990s bust was only 2.3 percent. The Case-Shiller proxy has also hit a year-over-year decline of just .1 percent, meaning that prices were effectively flat for the year. This makes for its own interesting historical comparison, but I will save that for a future post.
The Case-Shiller index is the most accurate measure of aggregate home price changes, for reasons long-since described here. But it's been an ongoing gripe of mine (and everyone's) that the index lags so badly. The data that was just released a couple days ago on October 27, for instance, only tracks home prices through August.
So lately I've taken to using the median price per square foot data to guess, for lack of a better word, what the Case-Shiller index values for more recent months might be. An example of this estimation can be found in the graph below, which appeared in the writeup of the most recent median price data:
Here's how I arrive at these estimates. (Non-nerds may wish to fall asleep for the remainder of this paragraph). The Case-Shiller index only uses single-family home prices, and in order to increase its accuracy it is calculated based on three months' worth of data. So I just take a three-month average of the median price per square foot for single family homes. The median-based indicators are less accurate and more volatile than the Case-Shiller index, but it seems like they can at least provide a rough idea of what's happened price-wise in the last couple of months.
I wanted to get a better read on just how rough an idea we were getting. So I went back to the beginning of 2008 and plotted two numbers for each month: the estimated price change calculated as described above, and the actual price change per that month's Case-Shiller data.
Now I'm wondering whether I should have invited the non-nerds to remain asleep for the rest of the article.
Anyway, this was the result:
It looks like the proxy did a pretty good job of predicting the Case-Shiller value for each month, except for the six-month period from November 2008 through April 2009, when it did a pretty awful job. What happened there?
It turns out that was right around the time that composition of home sales started tilting overwhelmingly towards low-priced homes. In a November 2008 article I noted that in October, home sales had decreased by 11 percent on an annual basis for the most expensive San Diego zip codes while at the same time they'd increased by a blazing 186 percent in the least expensive zip codes. It makes sense that such a huge shift in preference towards lower-priced properties would result in a lower median price per square foot of homes sold.
This effect couldn't last forever, though, and it started to recede in 2009. As of an update on the July 2009 data, the gap in year-over-year sales growth had collapsed: sales in the most expensive zip codes were up 8 percent for the year compared to an increase of 21 percent for the cheap zip codes. Sure enough, that is near the time that the estimation technique started to work decently again.
So for the time being, since it's been behaving itself in recent months and since we have a culprit for past episodes of innacuracy, I am going to give my little Case-Shiller proxy the benefit of the doubt and assume that it's providing a decent preview of the what the actual Case-Shiller data will bring when it is finally released. Future disparities will be noted and their origins hopefully surmised.
Nobody should be very surprised that the Case-Shiller index rose again in August.
This time around, the long-suffering low tier turned in the best performance with a robust 2.5 percent increase. The middle tier rose 1.6 percent for the month and the high tier increased .3 percent, with the aggregate index rising 1.6 percent.
Here's a graph of the three price tiers and the aggregate index since their bubble peaks:
The longer-term view shows the subprime-driven ascent and eventual collapse of low-priced homes:
And here's a view of the tremendous improvement in the year-over-year rates of decline for all the indexes:
As of August, the aggregate index had risen 5.7 percent from its April low -- an increase that is substantially larger than the spring rallies that typically pepper multi-year housing busts.
Here, a little later than usual, is the monthly foreclosure activity update.
The following chart shows that default notices and trustee sales, respectively the initial and final stages of foreclosure, both declined last month.
But both defaults and trustee sales remain at levels that are, shall we say, elevated. The next graph shows both series since 1990, adjusted for growth in San Diego's population. The pace of foreclosure activity may be down, but it is well above anything seen in the 1990s.
We can take a slightly different view by comparing default notices to home sales, which is a way of comparing housing demand with potentail must-sell supply. The following graph shows the sale-per-defaul ratio, averaged over 12 months to remove seasonal effects, for both the current bust and the 1990s bust. Again, there is no comparison.
Foreclosure activity may have crept down of late but it's still blowing away pre-bust historical comparisons.
Let's take a slightly different look at the construction, finance/real estate, and retail job sectors. I have long highlighted these three industries in my analysis because they were directly involved in the housing bubble, benefitting from the respective frenzies for building homes, lending funds, selling homes, and spending all the money that issued forth from the regional home equity ATM.
Sunday's chart on year-over-year employment declines shows that in the year leading to September, more jobs were lost outside the three housing bubble beneficiary sectors than within them. But over the course of the real estate bust, it is the housing-related industries that have take taken the big hit.
The accompanying chart displays the various job sectors' changes, in percentage terms, over the three years leading up to last month. This time period begins before the start of the nationwide recession, which officially commenced in December 2007, but after the air had started to leak from the housing bubble.
The cumuluative losses over this time period were far worse for the housing beneficiary sectors than for the economy at large. Retail was down over 9 percent, finance declined by over 11 percent, and construction shrank by a brutal 30 percent. This compares to a decline in all other sectors of the economy of a little over 1 percent for that three-year period.
In job terms instead of percent terms, retail lost 13,900 jobs, finance lost 9,500 jobs, and construction dropped 27,900 jobs. This is a total of 51,300 jobs shed in these three sectors, compared to 13,400 jobs lost in the rest of the economy over that same period. The housing bubble beneficiary sectors have accounted for 79 percent of all San Diego jobs lost over the past three years.
The rate of year-over-year job losses in San Diego declined again last month, according to the latest estimates from the EDD. Between September 2008 and September 2009, the region lost 52,000 jobs. This is not great, obviously, but it's an improvement over recent months.
The accompanying graph provides a look at trailing-year job losses since January 2008, with the sectors that were at the epicenter of the housing-fueled economic boom broken out separately.
When discussing the June job numbers, I hazarded a guess that the year-over-year numbers would soon start to look "less dismal." This wasn't because I expected that things would necessarily improve, but because the monthly job numbers would start comparing to progressively lower year-ago figures once we started measuring against late 2008, when the jobs picture really started to get ugly. One reader (that I know of) mocked this forecast but so far it's looking like it may have been correct, as the annual rate of job loss peaked at 57,300 jobs one month later in July and has declined since.
Employment is actually estimated to have increased by a meager, but at least positive, 900 jobs between August and September. One shouldn't make too much of a month's data, especially when seasonal factors are involved. But month-to-month employment shrank in both September 2007 and September 2008, so it's not as if September is a particularly strong month for job growth.
There has been so much volatility over the past year that inflection points could be missed by focusing only on the annual rates of change. So monthly changes are something I'd like to keep an eye on, with all due disclaimers about seasonality and the random fluctuations that can occur in a single month of data. This is especially so when the data in question is very much an estimate -- a topic I've often discussed in relation to the EDD's job numbers.
Bearing all that in mind, we've had one month of decent employment data here in San Diego. Let's see what happens next.
On Monday, voiceofsandiego.org colleague Liam Dillon offered up some good analysis (with charts, my most beloved visual aid) suggesting that the current economic and market downturn probably accounts for about half of San Diego's recently-announced $179 million budget deficit. This contrasts with the language coming from Mayor Sanders that seems intended to heap all the deficit blame on the recession.
That San Diego seemed to suffer from a structural deficit even during flush economic times is an important point. I would take it a step further and suggest that the incidence of a recession is not all that valid of an excuse anyway.
First, recessions happen. It seems reasonable that budget planning, whether at the household, business, or city level, should account for the economic ups and downs that life inevitably brings.
But this recession is a particularly poor scapegoat. While its severity shocked most people, myself included, it should have been clear that a period of economic weakness for San Diego was very likely at some point. Back in the good old days of 2006, to cite just one example of how this outcome wasn't entirely unpredictable, I put up an article showing that housing-related job sectors had accounted for 49 percent of San Diego employment growth in the entire decade to date. The numbers made it quite clear that San Diego's seemingly robust economy was heavily dependent on an unsustainable housing bubble.
Throughout those pre-recession days I also wrote about the magnitude and riskiness of debt being taken on by San Diego home buyers, the correlation between home prices and retail activity, and the historical precedent for a housing bust being a major cause of recession -- to say nothing of the housing bubble itself.
The data that fed into this analysis was for the most part freely available to anyone who cared to analyze it. And it all pointed to the conclusion that weak employment and consumer (which is to say, taxpayer) activity was, as I put it in a refutation of some particularly horrid analysis by the Union-Tribune editorial staff, "a likely if not inevitable result of the excesses of the housing bubble."
So while it may have been difficult to foresee just how bad this recession would be, having been blindsided by the fact that there was a recession at all isn't much of an excuse for anything.
Here are a few quick graphs to portray demand and supply for San Diego resale housing in September.
Sales increased last month and remained slightly above the levels from the same time last year. While sales are still on the weak side compared to San Diego's real estate history, they've soared from the record-setting low levels in which they were mired in late 2007.
The availability of housing for sale, meanwhile, remained at the modest levels seen since this spring:
We can compare how supply stacks up against demand by dividing the number of homes for sale in a given month by the number of sales that same month. The combination of decent demand and tight supply has made for a very low months-of-inventory figure throughout 2009, and last month was no different.
The median price per square foot of San Diego resale homes rose yet again in September. Prices by this measure rose 2.2 percent for detached homes, 5.5 percent for condos, and 3.1 percent for a volume-weighted aggregate of the two.
The following graph, displaying the median price per square foot since it peaked in 2005, shows that the condo series is quite volatile and that the series for detached homes gives a much better read of what's going on.
Here's an update of the graph showing the Case-Shiller index since its peak with estimated values for August and September. As I explained last month, this graph uses the detached home median price per square foot to predict what the Case-Shiller index will show for recent months when it finally comes out.
Assuming the projections are more or less correct, we can compare recent change to the index with those of the past. And that comparison is interesting. I estimate that the Case-Shiller index has risen 8.8 percent from its April low point -- an increase that blows away any seasonal rally that took place in the 1990s downturn.
Whether the price increase will endure is a subject for debate, but it must be acknowledged that this is something quite a bit more than the typical spring/summer bounce.
For a long time I have been discussing, with various degrees of rantiness, government intervention in the housing market. When I first touched on the subject in early 2007, before any bailouts had begun, some of the potential interventions I envisioned seemed kind of far-fetched. By late 2007, as I noted in a Manimal-referencing followup, many of these same interventions were already underway.
And now? The lengths to which the government has gone to prop up the housing market have surpassed even my own cynical expectations. By a long shot.
The most widely-discussed intervention right now is the $8,000 first-time homebuyer tax credit, but that's nothing. Far more significant is the fact that the Federal Reserve has committed to create money out of thin air in order to finance the purchase of $1.25 trillion worth of mortgage-backed securities. (Purchasing mortgage-backed securities provides funds to be lent out as individual mortgages.) The Fed is also buying $200 billion worth of bonds issued by government-backed mortgage giants Fannie and Freddie, also to presumably be lent into the mortgage market.
This amounts to intervention on an immense scale as nearly $1.5 trillion of money is directly shunted into the financing of home purchases.
Clearly, dumping all this money into the mortgage market results in mortgage rates that are significantly lower than they would otherwise be. It's tough to say how much rates have been suppressed, but the following comparison gives a rough idea: in October 2008, the month before the $1.25 trillion purchase program was announced, the 30-year fixed conforming mortgage rate averaged 6.20 percent. Last month, the rate on the same mortgage averaged 5.06 percent. This translates to a savings of nearly 12 percent for the monthly mortgage payment on a given purchase price.
Other efforts to prop up home prices are out there as well. Off the top of my head, and not counting the generalized bailout of the entire financial industry, these include: the aforementioned $8,000 credit; the Fed purchase of $300 billion worth of US Treasuries (which also drives down mortgage rates because mortgage rates are determined partially based on Treasury rates); the explicit guarantee of $5 trillion worth of Fannie and Freddie mortgages; government-sponsored FHA loans that require only 3.5% down; and assorted foreclosure moratoria.
My opinion on how much good all these programs will do in the long run can be found in the second article linked above; I will spare readers of this particular blog entry. The point here is that the government has pretty much declared war on the housing bust.
So it's no wonder that we've seen improvement in the housing market in the form of an unusually strong summer rally, increasing sales, lower inventory, and a rapid improvement in the year-over-year rate of change in home prices. But when I write about these improving conditions, I get a lot of pushback from readers arguing that the improvements are not "real" because things would be a lot worse without all the government efforts to support the market.
These readers are absolutely right in the sense that the housing market owes much of its quasi-recovery to artificial and, in the long-term, unsustainable factors.
Where I part ways with them lies in their belief that this distinction matters all that much in terms of forecasting near-term outcomes. The fact is that the government is massively intervening in the housing market. That's the world we live in today. And to the extent that we are trying to figure out what's going to happen in the housing market over the next year or two, it's just not all that useful to put a lot of thought into what would be happening in that non-existent world where the government isn't going to such lengths to prop up housing.
It's true that some of the intervention programs are scheduled to end soon. The $8,000 tax credit expires this coming December and the $1.25 trillion mortgage backed security monetization program is set to end in early 2010. That's the theory, anyway. But already, certain congresscritters are angling to get the homebuyer tax credit extended or even increased to $15,000 and expanded to benefit all buyers instead of just first-time buyers. As for the Fed's mortgage- and Treasury-buying spree, they stated in August that they "will continue to evaluate the timing and overall amounts of [their] purchases of securities in light of the evolving economic outlook and conditions in financial markets." In other words, if they feel like extending the program, they will go ahead and do so.
So these predetermined deadlines don't seem to be very meaningful. As it has done since the beginning of the crisis, the government changes the rules on the fly and as it sees fit. Perhaps some programs will expire and the housing market will weaken again. In this case, I would guess that the real estate-boosting programs would be fired right up again and probably increased at that point. Or perhaps the programs will be preemptively renewed as some of our legislators are already trying to do with the buyer tax credit. I don't know.
What I do know is that heavy-handed government intervention on behalf of the housing market will be with us for a while to come. Housing market analysts who ignore that fact do so at their own risk.
The July update of the Case-Shiller San Diego home price index is in. The index increased by 2.5 percent from June -- a substantial (if expected) one-month bounce.
As usual, Kelly has done a nice writeup on month-to-month changes with and without seasonal adjustments. I'll supplement her piece with some visual aids.
First up is a look at the three price tiers and the aggregate price index from their respective peaks in the 2005-2006 region:
While the recent price increase has been unusually strong in the context of a preceding decline, it appears less significant compared to the mighty boom-bust cycle that has played out this decade:
Here is a two-decade look at prices:
This next chart tracks the year-over-year rate of change for the three tiers and the aggregate index.
This figure has improved vastly over the past year. Overall prices were still falling at about a 12 percent rate as of July, but look at how much this rate has slowed: the year-over-year rate of decline was 25 percent a year prior and 16 percent just a month prior. If the current rally can continue for a while, prices could actually soon be rising on a year-over-year basis. Of course, that's a fairly big "if."
In response to last week's article on historical home sales, a reader requested charts expressing the sales data in terms of dollars' worth of homes sold instead of just the number of homes sold.
I had to cobble a few things together to make these charts. (Non-nerds may skip the rest of this paragraph). Ideally, the dollar volume of homes sold would be calculated by either just adding up individual prices of every sale or multiplying the average price times the number of units sold, which would both come out to a precise total of dollar volume. But I didn't have historical average prices. What I did have historical data on was the Case-Shiller index, which I rebased to the median price for all homes sold in San Diego in August 2009. This isn't exactly the same as the average price for any given month, for various reasons, but I suspect it will be close enough to get a general idea of what's going on.
Okay, everybody back? Here is the chart of existing home sales in terms of dollars since 1990, with the requisite 12-month average to smooth out the seasonal ups and downs:
Given the declines in both prices and volume since the height of the boom, it should not be surprising that the dollar volume of homes sold since that time has dropped hard. The twelve-month average dollar volume hit its bubble-peak high in June 2005 at over $1.83 billion per month of activity. At the low point in August 2008, the twelve-month average had swan-dived by a full billion dollars, falling to $.83 billion per month. The average monthly dollar volume has since rebounded to $.98 billion, but as one would expect this is still substantially below peak levels.
There are a couple of reasons why we really can't use the above chart to make valid comparisons to the more distant past. (Incoming -- more nerdy stuff). One was discussed in the prior article: since San Diego population and housing supply has been growing that whole time, one would expect sales to increase accordingly. The use of dollars as a measuring stick induces another problem. As a result of inflation, the dollar's purchasing power has been steadily eroded away since the start of the data in 1990 (and long before that, of course).
A proper comparison of today's dollar volume with that of times past must account for both San Diego's population growth and the declining purchasing power of the dollar. For a single data series that encompasses both effects and is available going back to 1990, I chose total San Diego income. This is not the per-capita income I often use to compare with home prices, but rather the total income earned by all San Diegans in a given year. This number will rise both with inflation and with the increasing San Diego population, so we can adjust for those two elements by calculating home sale dollar volume as a percent of total income.
For those who are still awake, here is the resulting chart:
There is still a mighty plunge from the bubble days, of course. But home sale dollar volume as a percent of income looks to be right in the middle of the range that prevailed throughout the 1990s. Notably, it is not far below the volume that took place during the healthy, post-bust but pre-bubble market of the late 1990s.
What does it all mean? Well, it's a little late on a sunny Friday afternoon to think too hard about such things. The main thing that jumps out to me is that the housing market's share of the economy as a whole isn't really very low in an absolute sense, but just in comparison to the go-go bubble years where all things housing-related (prices, volume, employment, and cocktail-party chatter, to name a few) grew unsustainably oversized. The trauma since then has largely resulted from things getting back to normal.
But that's not really anything we didn't know already.
According to the Employment Development Department's latest estimates, San Diego's year-over-year rate of job losses slowed for the first time in 2009. The region's employment decreased by 55,600 jobs between August 2008 and August 2009, a decline of 4.3 percent.
That is good news -- if the latest numbers turn out to provide an accurate picture of San Diego's employment situation.
The numbers that comes from the EDD each month are estimates that are subject to future revisions. The reason I note this is that the apparent improvement between the July and August figures resulted from a downward revision in July's original employment number. The report that came out last month estimated a year-over-year loss of 55,100 jobs in July, while the latest EDD release revised July's loss to 57,300 jobs. So August's seemingly improved loss of 55,600 jobs is better than the revised July figure but worse than the original estimate.
We will only know if August was actually an improvement if the August numbers are not themselves subject to a big downward revision as happened with the July figures.
But at least the initial estimate shows an improvement. That hasn't happened this year until now.
The unemployment rate was revised upward for July but estimated to have held steady in August at 10.4 percent.
San Diego unemployment continues to fare better than California (12.1 percent) but worse than the nation as a whole (9.6 percent).
Home sales may have recovered strongly from the depths they plumbed in 2007 and 2008, but they are still anemic when compared to San Diego sales activity over the past couple of decades.
You wouldn't know it from looking at a chart of historical home sales. The number of sales, indicated by the blue line in the following graph, has recently vaulted up to the upper part of the historical range:
But it's not quite that simple. (Is it ever?)
The most recent month charted was July, which is typically one of the fastest-paced months for sales. And the spikes up and down in the chart make it pretty clear that seasonal swings in sales activity are pretty huge. So the most recent data point should be compared not to the overall range of the blue line's ups and downs, but rather to the vicinity of the line's annual peaks.
Another way to account for the seasonal changes is to use an average of the past twelve months' sales. Since this moving average will always include all twelve calendar months, seasonal effects have no impact. The twelve-month average is indicated by the orange line in the above chart.
Even the moving average appears to be fairly robust compared to the past range of sales. But there's another issue to consider. Since 1990, when my data set begins, a lot of people have moved to or been born in San Diego. So to really do a fair comparison of current and past sales, it is necessary to adjust the figures for San Diego's population. I have done so in the next chart by increasing prior sales in proportion to the subsequent increase in the San Diego labor force (used as a proxy for population because, unlike population data itself, labor force stats are available on a monthly basis like the sales data).
Adjusting for population growth provides a pretty different picture. Recent months appear to comprise one of the lowest peak seasons since the data began. And while the twelve-month average has rebounded in a big way from the unheard-of lows reached in 2008, it is still barely above the levels set during the slowest periods of the 1990s bust.
Here's one other way to look at it: labor force-adjusted sales during the most recent July were 11 percent lower than the average July sales figure for the entire period.
Sales have certainly come back from the dead to some extent, but they are still pretty weak compared to what could be considered normal. The recent rally seems to have been caused more by limited supply (notwithstanding the shadow inventory of foreclosed properties I'm so fond of writing about) than by robust demand.
The number of San Diego homes officially going into default declined to 2,906 in August, down pretty sharply from July's 3,543 defaults. The number of homes actually being repossessed by lenders also fell, dropping to 1,358 in August from 1,607 in July.
But while foreclosure activity may have declined from its recent levels, it was still quite brisk:
And the pace of foreclosures in August still towered over what was seen in the region's 1990s housing bust:
The disposition of all these foreclosed homes is a hot topic. Will they eventually hit the market and put downward pressure on prices, or will they continue to remain in the shadows as they have thus far? Until that question is answered convincingly one way or the other, forecasting the future of San Diego's housing market will be a tough business.
As in recent months, the supply of San Diego homes for sale in August continued to be quite low compared to the number of people buying homes.
I typically show a chart measuring the number of months' worth of inventory -- available inventory divided by the number of sales in a given month. This time, let's break that figure down into its components.
Sales activity was down for the month of August, but this next graph shows that this is a typical seasonal pattern. Home sales in August were up about 6 percent from a year prior.
Now look at the amount of housing inventory for sale. While the pace of sales changed only mildly from a year ago, inventory has plummeted 30 percent over the same period.
Supply and demand combined to keep months-of-inventory near the unusually modest levels of recent months.
Shadow inventory may or may not be waiting in the wings (we'll check up on foreclosures next), but for now there just isn't a whole lot of housing to choose from in San Diego.
Summer may be winding down, but the summer rally in the size-adjusted median price of San Diego homes continued another month. From July to August, the median price per square foot rose .7 percent for detached homes, 1.3 percent for condos, and .8 percent in aggregate.
This was not much of a month, relatively speaking, but it turns out that the mid-2009 rally as a whole has been unusually powerful.
A while back I wrote about the fact that spring-summer price rallies tend to take place even during multi-year price declines. As the following graph from that article shows, a spring or summer increase in home prices doesn't rule out the possibility that a longer-term price decline is still underway:
What's interesting is that the price increase that's taken place since earlier this year has been unusually strong compared to spring rallies past.
The chart directly above utilizes the Case-Shiller index, a measure of aggregate home prices that is superior to the size-adjusted median price (not to mention available all the way back through the last housing bust). Comparing apples to apples, the Case-Shiller index has risen by 2.0 percent from its low point in April of this year. This is on par with, though not quite as big as, the largest spring rallies of the 1990s.
But because the Case-Shiller index is only available through June, that 2 percent increase ignores whatever price rises may have occurred in July and August. To address that issue, I have attempted to estimate the July and August Case-Shiller readings based on changes to the three-month average of the detached home median price per square foot. If the movements in the size-adjusted median are pretty much in line with changes to actual home prices, this should provide a decent idea of what the July and August values for the index will be.
According to this technique, the Case-Shiller index will show an increase 6.8 percent between April and August of this year:
That increase is nearly three times as large as any spring rally that took place during the 1990s. We will see how accurate my little estimation technique turns out to be, but for now it's looking like this rally is substantially more powerful than anything we saw during the last bust.
For a while I've been tracking a set of statistics that highlighted the great disparity between homes sales in higher-priced and lower-priced areas of San Diego. What we've been seeing for quite some time now is that compared to the expensive areas, the cheap areas had fallen a lot more in price but had experienced drastically higher sales volume on a year-over-year basis.
As of July, this disparity was still in place to some extent -- but the gap had closed substantially.
Here's how this particular study works. I take all the San Diego zip codes, throw out any that had fewer than 15 sales in July 2008 or July 2009, and order the remaining ones by how expensive they are based on the July 2009 median price. Then I average the year-over-year change in sales and median price for 20 most expensive zip codes, and then do the same thing for the 20 least expensive. (We all know that the median price has its problems, but I figure that by averaging together 20 zip codes we can at least remove some of the noise and get a decent read on what prices are doing).
The results are shown in the table below. To sum it up, the expensive zip codes saw a mild increase in sales and a minor decrease in price between July 2008 and July 2009. Over the same period, San Diego's least expensive zip codes experienced a substantial rise in sales alongside a big drop in price. (More detailed data can be found here).
What's interesting is that while the differences between areas are fairly dramatic, they are much less so than they were as recently as April 2009, the last time I updated this data. Have a look at the April version of the above table:
Between April and July, the year-over-year price change improved 6 percent for expensive homes (from negative 12 percent to negative 6 percent) while it improved 10 percent for cheap homes (from negative 34 percent to negative 24 percent). That brought the respective rates of price change a little closer together.
But the change in year-over-year sales volume was a lot more dramatic. The annual sales change increased 13 percent for expensive areas (from negative 5 percent to 8 percent) but declined a gigantic 58 percent for low-priced areas (from a 79 percent annual increase in April to a 21 percent increase in July). Here's a visual take on the annual sales changes:
So while cheap areas are still selling faster on a year-over-year basis than expensive areas, they are doing so to a far smaller degree than they were earlier in the year. I'll keep tabs on this data in the months ahead to see if the trend continues.
As with every month so far in 2009, more existing San Diego homes went into foreclosure than were sold. Just barely, though -- the ratio of home sales to default notices (the initial stage of foreclosure) was just gnat's eyelash below one-to-one. The ratio was .997, to be exact. That's the best sales-per-default ratio all year.
But it's still terrible. The following graph shows that while the sales-per-default ratio is above the lows set earlier in this downturn, it's still well lower than it was at any time during the two decades or so that preceded the current housing crash.
Here is a look at the same data series using a twelve-month average to smooth out seasonal variations. This gives a better big-picture idea of what's going on. And what's going on is that while the twelve-month average sales-per-default ratio has been moving upward since late 2008, it's still quite low compared to times past.
Here's a closeup of the twelve-month sales-per-default ratio during the current bust alongside the same figure for the early-1990s housing downturn. The ratio is still less than half of what it was this far into the prior bust.
While sales volume may be healthy, shadow inventory continues to mount.
What's still unclear is when -- or even if -- the bulk of this inventory will come out of the shadows. The rumors and speculations are all over the map: that the foreclosures are about to be dumped onto the market en masse, that they will never hit the market at all, and everything in between. The entire topic is quite murky, at least to me. But the concept of shadow inventory remains very much on the radar screen, and it will continue to do so until the numbers being tracked in the above charts approach normalcy once again.
The Case-Shiller index of San Diego home prices notched up its second monthly gain in June. Wait -- wasn't June, like, two months ago? And given that the index is based on the preceeding three months' worth of data, doesn't this give a better idea of the price movement in May (the middle month of the three) than June?
Yes and yes. Such lagginess is what we hate about the Case-Shiller index. What we love about it is the fact that it is an apples-to-apples comparison based on subsequent sales of the same homes. This provides a much more accurate view of home price changes than indicators like the median price, which measures how much the typical buyer is paying but doesn't account for what he or she actually got for the money.
As an aside, I've now seen two different real estate boosters write in two different publications that the Case-Shiller index is flawed because it "only counts homes that have been sold twice in a short period of time" (or something to that effect -- I'm paraphrasing but the intended meaning is accurate). This is simply and completely untrue -- there is no time limit between home sales. So, apparently you can just make stuff up. I wish I'd known that sooner; writing this blog over the past three and a half years would have been a lot less time-consuming.
Back to the June data. The aggregate San Diego index was up 1.6 percent for the month. My pal Kelly notes that even if you back out the seasonal influence of the typical summer rally, the index still rose for the month.
The following graph shows the decline in the three price tiers as well as the aggregate index since their respective peaks. (Please see this article if you'd like to know more about how the tiers are calculated or why a $407,000 San Diego home is considered "high-priced"). After showing some relative weakness, the high tier has been strongest for the last couple months, with the low tier pulling up the rear. Just like old times. Well, 2008, anyway.
For some perspective on the post-boom bloodshed in the low tier, consider that tier's spectacular rise during the bubble:
The subprime lending frenzy ran up prices at the low end more, percentage-wise, than in the middle or high tiers. The housing crash has seen the low tier give back those artificial gains and return to something approaching parity with the other two tiers.
The next chart provides an updated look at the year-over-year changes to each tier. Prices continue to be firmly down from a year prior, but their annual rate of decline has been steadily shrinking for a while now:
The summer rally lives on in the Case-Shiller index. And based on more recent data on prices and inventory, I expect the Case-Shiller rally will continue for at least another couple of months.
The California Employment Development Department released the latest job estimates today. According to these estimates, July saw San Diego hit its highest year-over-year rate of job loss in the downturn to date. Between July 2008 and July 2009, the region lost 55,100 jobs, a decrease of 4.2 percent.
The following graph shows how many jobs were gained or lost in the three housing bubble-related sectors I like to highlight -- construction, finance, and retail -- along with all other sectors. While the trouble started in the housing-beneficiary sectors, the growth of that green bar shows that losses have mounted outside those sectors over the past year.
Next up are a couple of somewhat homely graphs showing which industries lost (or, in a couple cases, gained) the most. This time around I've separated out education and health care, which for some reason have always been lumped together by the agencies that compile these statistics. Health care and government were the only year-over-year gainers last month.
Here's a look at the same sectors but terms of percent of the sector's size instead of the number of jobs gained or loss.
Boy, that financial sector doesn't seem to be doing so badly. You know, the same financial industry that was at the heart of the crisis, and later got bailed out with massive infusions of taxpayer money? Yeah, that one.
Finally, here is a graph of the unemployment rate over this recession and the last two recessions for comparison. Unemployment nudged up last month and also hit a new high for the downturn to date (not to mention the prior two as well).
Remember, unemployment is not a leading indicator. At least, it never has been, and that's been true even at higher unemployment rates than we are experiencing right now. As I showed in a graph a couple months back, unemployment has never peaked until recessions either had ended or were just about to. This was even true for the Great Depression (yes, I checked... and while it was tougher to tell because I only found yearly instead of monthly data, it did appear that unemployment peaked right around the very end of the Depression).
So it really doesn't fit with historical precedent to cite the high rate of unemployment as a justification that the economic downturn is going to get worse (even though I hear exactly that reasoning all the time). Of course, this doesn't mean that the downturn won't get worse. It just means that the unemployment rate is not a reliable leading indicator as to what will happen next with the economy.
Earlier this year, I advanced the case that San Diego home prices had, as a whole, become "reasonable." For all its seeming lack of ambition, this statement launched a slew of (often slightly combative) emails my way.
Half a year or so later, it's time once again to check in on how local home prices stack up against incomes and rents.
The first graph compares San Diego-wide home prices, as measured by the Case-Shiller home price index, divided by local income per San Diegan. It turns out that not a whole lot has changed since the last snapshot of this series in December. Prices declined earlier in the year, but a combination of the summer price rally and declining incomes has pushed the ratio back up to a point where it has hardly changed since the beginning of 2009.
The second graph compares prices with San Diego rents as recorded by the inflation-measuring Bureau of Labor Statistics. As with the price-to-income ratio, the relationship between home prices and rents is just about where it was at the beginning of the year.
(Data nerds may wish to note that the 2009 local income figures in the above graph are estimated based on changes to national incomes, and that home prices in both graphs are estimated for the past couple of months based on the size-adjusted median single family home price).
Given the lack of any substantial change to these valuation ratios, I will once again declare aggregate San Diego home prices to be "reasonable" in comparison to their historical relationship with incomes and rents. Homes are no longer overpriced, but they aren't quite cheap yet, either. (And please note the use of the word "aggregate." As often discussed here, different areas of San Diego have behaved quite differently, meaning that while the aggregate home price is reasonable compared to the traditional fundamentals, certain areas may be genuinely cheap while other areas may still be quite overvalued.)
San Diego homes do look underpriced, on the surface, if we use monthly payments instead of purchase prices. The payment-to-income and payment-to-rent ratios are just barely above the multi-decade lows set earlier in the year.
However, the low payment-based ratios are less a result of inexpensive homes than they are of rock-bottom mortgage rates. If mortgage rates are to rise in the future (and I believe they will, a lot) then these ratios don't give such a good read on whether homes are overpriced or underpriced on a longer-term basis. If you'd like to read more about why this is, I went a lot deeper into the topic in the prior entry on monthly payments.
So I'm sticking, still, with "reasonable." And with that, I'll wait for the next batch of emails to come in.
The month of July saw an additional 3,543 San Diego homes enter the foreclosure process when their owners were sent notices of default. An additional 1,607 homeowners received trustee sale notices, indicating that their homes had entered the final stage of the foreclosure timeline before repossession.
Both figures are slightly down from the prior month, as the accompanying graph shows, but not too far off all-time highs.
So despite the dearth of currently available housing inventory, there is still reason to be concerned that all these foreclosed properties will eventually hit the market. The shadow inventory problem -- most recently discussed here -- remains.
I've been talking for months about the low availability of San Diego housing for sale. In July, as the accompanying graph shows, that scarce inventory got even a little scarcer.
The graph measures the number of existing homes available for sale in comparison to how quickly homes are selling, generating a single measure of supply versus demand. Last month, there were just 4 months' worth of inventory on the market -- clearly quite low compared to prior years.
On top of that, 29 percent of homes for sale were "contingent," meaning that they were short sales or the like with offers awaiting approval from the lenders. In other words, this 29 percent wasn't really for sale in the traditional sense. (See this blog entry for a primer on the "contingent" category of for-sale homes).
It's the same story as in recent months. "Right-now" inventory remains very limited (and prices are rising accordingly) but plenty of "maybe-future" inventory waits in the wings. I'll be back to check in on that second type of inventory next week.
There was a bit of divergence in the size-adjusted median price of San Diego homes sold in July. By this measure, single family home prices were up a robust 3.7 percent from June, whereas condos gave back most of June's explosive gain with a subsequent 6.0 percent decline.
I don't make too much of the drop in the condo size-adjusted median. The accompanying graph shows that the condo figure has tended to be much more volatile than the single family figure and has been subject to violent one-month swings in either direction. This is largely because the sample size for condos is much smaller (fewer than half as many condos sold as did detached homes last month). The size-adjusted median price for detached homes gives a better month-to-month read on what's happening.
In any case, a volume-weighted aggregate of the single family and condo price figures rose for the month by .8 percent, so I'm scoring July as a continuation of the home price rally that has been going on since early 2009.
I noted earlier in the week (and incessantly before that) that home prices have a seasonal tendency to rise in the spring and summer even during the midst of a multi-year price decline.
That sounds like a good enough excuse to make a chart.
Below you will find a look at San Diego home prices, as measured by the Case-Shiller index, from 1990 through 1996. These years encompassed the entirety of the five-or-so year home price decline visited upon San Diego after the late-1980s housing bubble.
As you can see, the Case-Shiller index measured a rise in aggregate San Diego home prices for every single year of the long price decline. (Although 1993 just squeaked in there with a one-month, .1 percent increase).
I'll bet that during each of those spring rallies, a lot of people became filled with hope the housing bust was finally over. But through five of these head-fakes, it wasn't.
Here for comparison is a look at the current housing downturn:
After an anemic spring bounce in 2006, the Case-Shiller index fell unceasingly through 2007 and 2008, only to finally register an uptick again earlier this year. The lack of spring-summer rallies is a testament to the brutality of this housing crash.
The next graph overlays the two housing crashes in order to compare duration and magnitude:
The price decline this time around has been substantially larger -- an outcome that was unsurprising based on the comparatively vast overvaluation of homes coming into the 2005 bubble peak. But while it may feel to some like this price decline has gone on forever, it has not yet endured nearly as long as the 1990s version.
Whether it eventually does so remains to be seen. Either way, the first graph should make it clear that a spring-summer home price rally should not be taken as evidence that the housing bust has come to an end.
E-MAIL POST
The orange line on the accompanying chart displays the year-over-year rate of change in the number of employed people in the county's labor force -- in other words, the number of people living in San Diego who have jobs somewhere. The blue line denotes the annual change in payroll jobs at San Diego companies, regardless of whether the job holders live in San Diego or not.
The accompanying chart displays the various job sectors' changes, in percentage terms, over the three years leading up to last month. This time period begins before the start of the nationwide recession, which officially commenced in December 2007, but after the air had started to leak from the housing bubble.
The accompanying graph provides a look at trailing-year job losses since January 2008, with the sectors that were at the epicenter of the housing-fueled economic boom broken out separately.
The month of July saw an additional 3,543 San Diego homes enter the foreclosure process when their owners were sent notices of default. An additional 1,607 homeowners received trustee sale notices, indicating that their homes had entered the final stage of the foreclosure timeline before repossession.
I've been talking for months about the low availability of San Diego housing for sale. In July, as the accompanying graph shows, that scarce inventory got even a little scarcer.
There was a bit of divergence in the size-adjusted median price of San Diego homes sold in July. By this measure, single family home prices were up a robust 3.7 percent from June, whereas condos gave back most of June's explosive gain with a subsequent 6.0 percent decline.
