Commercial Real Estate Market Stability and College Towns

If you’re looking for CRE markets that are insulated from downturns, college towns are a good place to start. The Creative Class post “Where Unemployment Is Worse Than Expected” analyzes the performance of various metro areas in this recession. Here’s one of their graphs:

Low and to the left is good (Iowa City), high and to the right is bad (Detroit, Kokomo and Elkhart). An excerpt from the post:

College towns number among the best performers, doing much better than predicted: Champaign-Urbana, Illinois, home to University of Illinois (-2.2); Iowa City, University of Iowa (-1.81); Manhattan Kansas, Kansas State University (-1.82); College Station, Texas, Texas A&M (-1.74); New Haven, Connecticut, Yale University (-1.54); State College, Pennsylvania, Penn State University (-1.47); Boulder, Colorado, University of Colorado (-.93); Austin, Texas, University of Texas (-1.0); Ann Arbor, Michigan, University of Michigan (-.94); and Ithaca, New York, Cornell University (-.97), among others.

The correlation isn’t perfect; for example, the metros with the major Oregon universities (Eugene and Corvallis) have both underperformed. However, the relatively stable employment base and demand for services created by large universities tend to buffer these markets. And, since employment is the most important determinant of CRE performance, CRE in these markets tend to do better.

The Problem With Medians

Prices, vacancy rates, rents - most pronouncements of real estate trends are the medians of a set of data. These reports are largely meaningless, or, even worse, actively misleading. Lansner on Real Estate’s post “Is Median Price Giving Bum Signals?” quotes John Burns, and Orange County housing consultant, on the subject:

We are extremely concerned that policy makers, banking and real estate industry executives, investors and others will use misleading home price data to conclude that home prices have stabilized. They have not. These same influencers used this data in 2006 and 2007 to make decisions, many of which have proven to be poor decisions. It was a tough lesson, and hopefully one that won’t be repeated. This is a complex issue. Here is why: Reported home prices and home price indices rely on a small sample of transactions that represent far less than 1% of the owned homes in an area.

The data problem is serious for homes, and exponentially worse for commercial real estate because there are even fewer transactions. The problem is an ongoing theme over at Calculated Risk; see “Misleading Housing Price Data” and “Median Price Mix Example”, for example.

An alternative is the “same store” trend, which compares sales prices and operating data from the same property over time. The problem with using this approach for commercial real estate is that property and price performance are idiosyncratic. For example, land values can change radically as a result of permits being obtained or a local building moratorium being imposed. The loss of a major tenant can have a big impact on the value and operating data of an office or retail property. A new, incompetent property manager in a multifamily property can cause a big spike in vacancy. There are so many micro factors which can influence prices and performance that in my opinion it’s very difficult to draw macro market conclusions from the relatively small samples we have to work with. 

Certainly we can identify major sustained trends, but if some told me their data showed office values are down 30% in a market, my conclusion would be prices are down somewhere between 15% and 45%.

What’s the solution? When it comes to commercial real estate, I don’t think there is one – there are just not enough transactions to draw meaningful conclusions.  Employment trends provide a good proxy for commercial real estate performance; I’ve yet to see a real estate market getting better when the market is losing jobs. I’ll believe CRE and residential markets are getting better when we start seeing employment growth again.

Bend, Oregon, and Elkhart, Indiana: Employment

I was in Bend, Oregon on Friday. For those not in the Pacific Northwest, Bend is famous for its unemployment rate. From the AP, on June 3:

The Labor Department said Wednesday that unemployment in April rose from a year earlier in all 372 metropolitan areas it tracks. Indiana's Elkhart-Goshen's rate jumped to 17.8 percent, up 12.7 percentage points from a year ago. The Indiana region, which posted the largest increase from last year, has been pounded by layoffs in the recreational vehicle industry.

The second-highest jump occurred in Bend, Ore. Its rate rose to 15.6 percent, up 9 percentage points from last year.

So how much is Bend like Elkhart? If you look at unemployment, they’re pretty similar:

Bend:

Elkhart:

However, the change in employment is a much different picture. Absolute numbers:

And percentage change:

 

Elkhart has lost almost 15% of its employment base – Bend less than 5%. Unemployment figures are interesting, but in evaluating the health of a local economy I think the number of people with jobs is much more relevant.

Employment data from this BLS site.

Housing Was Not Massively Overbuilt

It’s widely taken as a given that because we have too many empty housing units now and because prices have collapsed, that housing was overbuilt. For example, from Unnatural Rent:

In addition, the recession and rising unemployment have slowed down new household formation, encouraging people to live with roommates. In many markets, apartment rents are unlikely to post any growth during this year, and some may even see declines.
This drop in demand has been combined with a massive increase in the supply of housing (both single family and multifamily) over the past decade. While office and industrial did not experience a huge wave of overbuilding, that isn't quite the case for retail and multifamily.

This is true in a sense – if we had fewer housing units now the situation would be better.  However, throughout the bubble years supply and demand were balanced. My argument is premised on the idea that additions to housing supply should roughly correspond to additions to employment:

1 new job = 1 additional unit

Obviously, not every person who gets a job creates a new household, but households are also created without jobs, and in my experience nothing too bad happens to housing markets where job growth exceeds new housing additions. The data for job creation and residential permits issued since 2004 is summarized below:

Supply and demand were in synch until 2007. In 2008, demand went off a cliff, which goes to show that jobs can be lost faster than residential development can wind down.

I think this data also supports the notion that the bubble price escalation was driven by easy financing, and not fundamental demand.

Employment data is from this BLS website, permit data at this Census Department website.

The Geography of Job Losses

There’s an extremely cool animation of job gains and losses since 2004 at Tip Strategies. Here’s a couple of screen shots:

The animated version is amazing. This reminded me of Jim Hamilton’s powerpoint which shows the regional propagation by quarter of recessions between 1969 and 2004, saved here. Did you know that the recession starting in 1990 actually began in Arizona in the fourth quarter of 1988?

You have to click through the slides yourself; the action starts on slide 14.

These are both good reminders that there are huge variations in a recession’s impact among metro areas.

Economy and Real Estate Post Picks: Week of May 3, 2009

Wholesale Sales Continue to Slide: Commodities are the hardest hit

The Latest Employment Report: Not as bad as previous months, but still not good

Does a Decline in Initial Jobless Claims Signal the Recessions End? A discussion of the impact of initial, continuing, and net jobless change

Regional Disparity in Unemployment Rates: The West Coast is faring poorly in this recession

CMBS Loan Performance: Transfers to Special Servicing are up sharply

Economic and Real Estate Post Picks: Week of April 13, 2009

Is the Cutback in Consumer Spending Abnormal? Consumer spending patterns in prior recessions

Job Loss Patterns in this Recession: An interactive map showing the geographic pattern of job losses in this recession over time

General Growth Property Bankruptcy: A set of stories exploring various aspects of the bankruptcy of the nation’s second largest mall owner.

What Direction is the Economy Headed? Russell Investment’s Economic Dashboard shows most indicators are headed in the right direction

Retail Sales Down in March: Declines in gasoline and motor vehicle sales were big contributors

Economic and Real Estate Post Picks: Week of April 6, 2009

Wholesale Sales Up, Inventories Down: Good news on these indicators

Why This Recession is Different: Unlike most recessions, this one is balance sheet driven

Has the Housing Market Bottomed? Builder stocks and the spread between mortgage and treasury yields are both hopeful signs

Why is Consumer Debt Declining So Sharply? An explanation for the sharp decline in credit card debt

Initial Unemployment Claims and the End of Recessions: Does the recent peak in initial unemployment claims signal we are nearing recovery?

Economic and Real Estate Post Picks: Week of March 16, 2009

Post Recession Employment Trends: How long does it take for employment to recovery after a recession ends?

Cap Rate Closing/Asking Gap: The spread between asking and closing cap rates is widening.

Maturing Loans Are Coming Home to Roost: The looming problem of maturing income property loans with no exit strategy

Reflation: The risk of deflation has diminished.

Has the Economy Hit Bottom Yet? Probably not, but the rate of decline is slowing.

Economic and Real Estate Post Picks: Week of March 9

Recession Proof Industries: Government and health care do best

Layoffs versus Quitting: Despite the economic downturn, more people quit than are laid off in most industries

Has the Decline in Retail Sales Stabilized? Charts showing retail sales trends

Wholesale Sales Decline: Wholesale sales by sector; durable goods show the biggest declines

Inventories are Declining Rapidly: A major inventory correction is underway

Economic and Real Estate Post Picks: Week of March 2, 2009

Price Stickiness and the CPI: The components of the CPI change at very different rates

A Long Recession Ahead?: The decline in household wealth could mean this recession will be a long one

Employment Decline, Recession, and Depression: A comparison of employment declines between this recession, 1981, and the Great Depression.

Credit Crunches and Small Business Finance: How small businesses are financed, and what happens in a crunch

Is the Pace of Layoffs Declining? Trend data from October, 2008 says maybe

Economic and Real Estate Post Picks: Week of February 9, 2009

How Bad is the Employment Picture, Really? Recession comparisons using Payroll Employment versus Household Employment data (hint: payroll employment is a better data source)

How This Recession is Different: Consumer, bank, and business balance sheets are much more leveraged then previous recessions

Real Disposable Income Up: The savings rate also improved in December

Significant Fall in Domestic Demand: The worst decline post-WWII

Retail Store Opening and Closings: Good information on trends in openings and closings by retail sector

Economic and Real Estate Post Picks: Week of February 2, 2009

Sharp Contraction in Trade: Both imports and exports have gone off  a cliff

Jobs Forecast by State and Sector: Interactive map and charts of a Moody's Economy.com forecast of job loss/gain by sector and state through 2012

The Housing Market: 1982 versus 2009: A comparison of our current situation with the situation in 1982

The Behavior of LIBOR in This Economic Crisis: Very detailed discussion of LIBOR and its recent movements

Upcoming Economic Indicator Releases: A useful calendar of upcoming economic indicator releases, with links directly to the data sites.

Will We have a Commercial Real Estate Crisis?

Casey Mulligan thinks probably not. From his New York Times piece:

For months now, experts have been predicting that commercial real estate will be “the other shoe to drop.” But in fact, non-residential building fell far behind housing construction during the housing boom. This shortage of commercial buildings relative to housing suggests that a commercial real estate crisis will not occur, or that at worst it will occur with much less severity than did the housing crash.

Here is the chart purporting to support this argument:

(Click on image for a larger version in a new window)

The error Dr. Mulligan makes is the belief that the housing bubble and future CRE performance was/is primarily a function of inventory. The chart suggests housing prices have collapsed because too many residential structures were built. That’s like saying Citibank’s stock price has collapsed because too many shares have been issued. Home prices have dropped because the financing that people used to buy homes at an inflated price is no longer available, not because there are more homes than people are willing to occupy. To the extent CRE inventories were tight, values were inflated, which won’t help us now if the deals were leveraged based on the higher values.

For example, look at Miami, a residential bubble market. The graph below shows the number residential permits issued in relation to the number of new jobs created on a rolling 12 month basis. The secondary axis is the OFHEO Housing Price Index year over year change.

(Click on image for a larger version in a new window)

Home price increases began decelerating in late 2005, but at the time Miami was creating twice as many jobs as new units, so if anything the market was undersupplied. Something else was clearly dragging prices down, and in retrospect we know it was the withdrawal of aggressive lending parameters.

Now, of course, most markets are losing jobs, and most markets are still adding units (and commercial real estate) as projects work there way through the development pipeline. We won’t see a recovery until the employment situation turns around.

What does this mean for CRE? We don’t know for sure how many deals were done with aggressive underwriting during the peak years, but we know there were quite a few and so we can expect some decline in values related to the withdrawal of aggressive leverage similar to what’s happened in the residential market. We also know that CRE is sensitive to employment trends, and those are very negative. The CRE situation may not become as bad as residential, but if it doesn’t it will be because the underwriting was better and employment improves. It won’t be because there was a lack of inventory.

The Landes Apartments Project has the Best Multifamily Location in the Whole World

OK, I don’t know this for sure, because I haven’t visited every multifamily site in the whole world. But, I think this location (901 8th Avenue, Seattle, WA) is a contender. Here is my logic:

1. The best apartment location should perform well in difficult market conditions.
2. In difficult market conditions, the sectors which perform best are government, health, and education.
3. The Landes location is ideally suited to appeal to government, health, and education workers.

With regard to the second premise, there is a helpful post at Macro and Other Musings titled "Where are the Safe Jobs?". Here is a chart from that post:

By far the most jobs have been created in the government and education/health services sectors. This is not a fluke of this recession – Eric Janszen put together charts of every sector showing data back to 1940 (posted here) and reaches the same conclusion.

So here is the location of the Landes Apartments (“A” on the Google Map below):

Easy walking distance to Seattle University, three major medical centers, and the Seattle/King County government buildings (shaded in red at the lower left).

There might be better locations, but I don’t know of any.

Which Markets Have Lost the Most Jobs?

Employment has suffered the most in Detroit (no big surprise there). However, there are some surprises in the other rankings, including which markets have held up the best.

A word on methodology. I looked at the highest employment level in each market since January, 2000, and compared it to the latest level. All data is from BLS Local Area Unemployment Statistics.

Here are the results:

Employment as of November, 2008	Peak Since 1/2000	Current	Change from Peak	% Change from Peak
Detroit	2,217,186	1,899,782	(317,404)	-14.32%
San Jose	962,408	845,417	(116,991)	-12.16%
San Francisco-Oakland	2,250,832	2,138,050	(112,782)	-5.01%
Chicago	4,721,131	4,542,407	(178,724)	-3.79%
Los Angeles	6,307,149	6,098,378	(208,771)	-3.31%
Riverside-San Bernadino	1,711,443	1,658,533	(52,910)	-3.09%
Atlanta	2,650,838	2,569,010	(81,828)	-3.09%
Washington DC	2,967,601	2,882,203	(85,398)	-2.88%
Miami	2,739,126	2,668,358	(70,768)	-2.58%
Orlando	1,070,271	1,048,644	(21,627)	-2.02%
Denver	1,346,897	1,323,378	(23,519)	-1.75%
San Diego	1,485,911	1,468,666	(17,245)	-1.16%
Sacramento	1,003,441	994,697	(8,744)	-0.87%
Dallas	3,023,034	3,005,173	(17,861)	-0.59%
San Antonio	906,335	902,089	(4,246)	-0.47%
Austin	831,555	829,083	(2,472)	-0.30%
Houston	2,680,121	2,675,806	(4,315)	-0.16%
Las Vegas	936,369	934,956	(1,413)	-0.15%
Phoenix	2,022,781	2,022,725	(56)	0.00%

Detroit employment peaked in June, 2000, and has lost jobs ever since. Here is a chart showing year over year job loss for this market:

(Click on charts to open larger versions in new windows)

The market with the second worse performance is San Jose. It, along with San Francisco (to a much lesser extent), has never fully recovered job losses sustained in the dotcom bust. Here is the year over year chart for San Jose:

The best performing markets are also something of a surprise: Las Vegas and Phoenix. Both of these markets have severely distressed housing markets, and the conventional wisdom is housing difficulties drag down employment. Here are the charts for these two markets:

Finally, let’s discuss Riverside-San Bernardino for a minute. There is no question this market is hurting – I’ve discussed it previously here and here. But, a Bloomberg story with Calculated Risk commentary suggests a parallel between Detroit and this market because both have the same high (9.5%) unemployment rate. I think it’s wrong to suggest Detroit’s situation, which has had sustained job losses for eight years totaling 14.3% of it’s peak employment base, is similar to Riverside-San Bernardino, which has only lost jobs for a little more than a year and is down a little more than 3% from it’s peak. I’ve previously argued unemployment is not a good measure of market distress, because it’s possible to have very high unemployment rates and still have positive employment growth.

You can download a free report which provides similar employment charts on many other markets here.

Who Cares About Unemployment?

Obviously, a lot of people, and not just those that are unemployed. But is the unemployment rate a good measure of how a local economy is doing?

Here is a chart of year over year employment change and the unemployment rate for McAllen, Texas:

(Click on chart for a larger version in a new window)

In January, 1999 the unemployment rate was pushing 20%, but between January, 1998 and January, 1999 the local economy added nearly 5,000 jobs. In fact, McAllen has added jobs in every year over year period for more than 10 years, during which time the unemployment rate never dipped below 5%. It seems pretty clear employment can be increasing despite a relatively high employment rate. And, I would argue the change in employment is a better indicator of an area’s financial health than it’s unemployment rate.

Data from BLS Local Area Unemployment Statistics

Cliff Diving: Riverside – San Bernardino

The Wall Street Journal has a piece on the softening commercial real estate market in Riverside-San Bernardino. The money quote: "California's Inland Empire, the two-county region that stretches east of Los Angeles, has gone from a booming development smorgasbord to a basket case in a few short years."

Here is year over year employment change for this metropolitan area:

(Click on image for a larger version in a new window)

Employment growth in this market has been decelerating since 2004, and has been negative for more than a year now. CRE can’t do well in such an environment.

You can download a free report which provides similar information on 18 other markets here.

Is Overbuilding Responsible for Excess Housing Inventory?

The President of the National Association of Home Builders says “The excess housing inventory in today’s market is the result of unprecedented foreclosures, not overbuilding.” Paul Jackson, Housing Wire, suggests this statement “borders on the certifiably insane ." I may be certifiably insane, but I think the NAHB position is closer to the truth.

Obviously, we have excess inventory. The amount is subject to debate, but arguments can be made for between 1.75 to 4 million excess units (see this Calculated Risk post, for example). Obviously, many of the excess homes are newly completed builder inventory. You can read the story behind the pictured subdivision here.  So, in a sense builders are responsible for at least a portion of the excess inventory. They built it, it’s empty, end of story.

But, of course, it’s not that simple. There are a lot of people who are living in substandard housing, in apartments, with their parents, with roommates, etc. who would be delighted to be living in these “excess” units. The problem is much of the excess is located in places people don’t want to live or can’t find jobs (read, for example, these depressing posts about Detroit in The Big Picture and The Weekly Standard). And, much of the excess is not affordable even at today’s depressed prices to the people who want the units.

I think Miami is a good example of what actually occurred. Here is a chart of residential permits issued in Miami between 1999 and November, 2008:

(click on images to open larger versions in a new window)

On it’s own, this is about as clear a case as you can get of overbuilding – permits obviously spiked between 2004 and 2006, which nicely dovetails with the peak of the subprime craziness. But, consider employment growth in Miami during the same period:

At the same time permits were peaking at around 45K per year, Miami was adding jobs at 100K a year. Can you really say builders were overbuilding when there are twice as many people with new jobs as units being added to supply? If anything, the numbers imply a housing shortage in the peak period. Here is a chart showing the ratio between new jobs and residential permits:

From mid-2002 through 2007 Miami was adding more jobs than housing units, and for most of this period it was adding around two jobs for every new housing unit. This was not an overbuilt market during that period.

In contrast, here is an equivalent chart for Houston:

Housing prices have held up relatively well in Houston, and most people do not consider it to have been one of the bubble markets. But, note Miami had substantially more jobs added per new unit than Houston did during this period. The data suggest Houston was relatively overbuilt compared to Miami.

In fact, the data suggest that maybe part of the problem in the bubble markets was builders didn’t build fast enough to keep pace with the demand created by new jobs (you can see similar charts for Los Angeles, San Diego, Las Vegas, and many more markets here). I’m not ready to go so far as to suggest they should have done so – had lenders stuck to reasonable underwriting standards more of that demand would have shifted to the rental market and we would have seen higher rents and less vacancy in that segment, which I think we all agree in hindsight would have been better than putting people in houses they couldn’t afford.

Does the Housing Market Benefit When Investors Buy Foreclosed Homes and Rent Them to Tenants?

Yes, it does. I wouldn’t have thought this question worth posting about since it seems so obviously true, but since Nobel laureate economist Joseph Stiglitz and Yale University Professor Robert Shiller apparently disagree (see this Bloomberg article), maybe I should explain my reasoning. Calculated Risk agrees with me for some good reasons, but I have a couple more.

The Schiller and Stiglitz argument is that the speculators will sell the homes when prices recover, and the reentry of these homes into the for sale market will be a drag on price recovery. There’s no data in the Bloomberg piece, and the anecdotes all involve buyers who are renting out the houses they’ve acquired. Apparently, we would be better off if lenders held the properties vacant until owner occupant buyers can be found rather than sell the properties to landlords.

Everybody including me loves owner occupants, but the day when residential REO can be absorbed by owner occupant purchasers is a long way away. Employment is falling sharply in all the distressed markets: for example, here’s what’s happening in LA:

You are not going to have much residential demand in LA until employment is trending up again no matter what you do to incentivize owner occupant buyers (we could waive down payment and credit requirements, of course, but we know where that got us). It does neighborhoods no good to have lots of boarded up houses for years (just ask someone from Detroit what 60,000 vacant units have done for them).

My second objection is more subtle. I have previously argued that a relatively high percentage of single unit rental housing correlated with the size of the housing bubble in that market. For example, of the markets tracked in the Case Schiller Price Index, Los Angeles, San Francisco, and San Diego had the highest percentage of single unit rentals in 2000.

I believe the investors that owned those units were probably sellers during the bubble days, and that the purchase of REO by investors is a return back to the previous equilibrium rather than a new direction. Unfortunately, we’ll have to wait a while for data and there are a lot of moving parts so we may never know conclusively.