Monday, April 28, 2014

Urbanized County Job and Wage Growth (or Loss)

Recently, I've been spending a lot of time digging into US Census data, sometimes commuting patterns, but also US Chamber of Commerce job numbers by county.  When I posted net job gains/losses by DFW area county the other day, people who were shocked to see how poorly Dallas County has faired (266,000 lost jobs in 10 years) asked how that compares to other major metropolitan areas.  It would make some sense that other areas around the country have also been de-urbanizing, so it was worth a look.

While nursing a slight hangover Saturday morning after staying up late to watch the Dallas Stars lose game 5 Friday night in Anaheim (and then watch in person :( last night), I decided to put compile and analyze that information.  Here's what I did:

I looked up the county data for the top 40 major US cities, minus New Orleans.  I left out New Orleans because 1) Katrina :( and 2) Parishes.  I like the county data because those are fixed boundaries unlike Metropolitan Statistical Areas (MSAs) which right along with sprawl.  The data also happens to be what's available, but it gives a good sense about the competition between core cities and their suburbs because often the suburbs are outside the county of the core city.

I also like job growth data because rarely can that be a bad thing whereas population growth may not necessarily be good or bad.  Job growth data can reveal whether more people are employed and a growing tax base.  It's also jobs that are officially on the books which was another reason to compare Dallas county to other urbanized counties.  Did the loss of 266,000 simply mean a number of jobs going off the books?  Did other urbanized counties experience a similar huge loss in job numbers?

Below is that table where I included 2011 job totals (latest available data), 2001 job totals (Dallas County's peak), percentage change, 2011 salaries, 2001 salaries (converted to 2011 $), and percentage change in salaries of urbanized counties.  I wanted to include this information because while job loss isn't necessarily good, it might also mean a rise in salaries from a proportionally larger labor pool competing for the lesser amount of jobs.  Ideally, you would probably want job growth and increased wages.  Lastly, I included nationwide totals to be able to compare to some sort of baseline.  How did each urbanized county fair against the mean so to speak.

Before I get to the chart how the cities layout, there is one other anomaly I have to point out.  Atlanta is split down the middle between two counties:  DeKalb and Fulton.  I included both individually in the table and then for the chart, I combined the two.  More on Atlanta later.  Now to the graphic version of the above table:

As you can see I divided the chart into quadrants.  Each of the quadrants is created by using the national mean for Job Growth/Loss (vertical line) and Wage Growth/Loss (horizontal line).  Nationally, we lost 1.42% of on the books full-time jobs from 2001-2011.  National salaries increased on average of 3.41%.  Dallas County performed below the national averages in both.

I then labeled each of the quadrants based on what was happening.  Urbanized counties that gained jobs and wages/per job will be in the upper right.  Counties that lost jobs but gained wages proportional to the national average will be upper left (where most are).  Job Growth/Wage loss is bottom right and Wage Drop/Job Loss is bottom left.  This is where you don't want to be.  This is where Dallas County is.

As you can see there is a general pattern then the anomalies.  First, the general pattern is that urbanized counties lost more jobs (-5.87%) but outpaced the nation in wage growth (4.58%). Counties: Suffolk (Boston), Los Angeles (LA duh), and Milwaukee hover right around the national mean for urbanized counties.

Let's put an urbanized county average on the chart to compare:

Now let's look at the good outliers.  Oklahoma City is right on the national average for job loss, but experienced HUGE wage growth from 01-11 (18.2%).  It's so high I feel the need to quadruple check the data.  I'm not sure I have an answer for this except for oil and gas.  We'll have to let some locals weigh in there.

DC and Houston both experienced large amounts of job growth and wage increases.  DC gained 13.03% in jobs and 10.14% in wages, presumably all on K Street.  I kid.  Sort of.  Harris County (Houston) gained 5.83% in jobs and a staggering 11.49% in wages.  Two things worth noting here are that DC has huge, highly successful suburban communities (Montgomery in MD and Arlington in VA) it is competing against and clearly holding its own.

On the other hand, Harris County is huge.  It comprises nearly all of Houston's metropolitan area and nearly 70% of the area's population.  Meanwhile, Dallas County only has about 1/3rd of DFW population. It's also about twice the size of Dallas county in land area so whatever "growth" occurs, it's likely to be accounted for in Harris' data.

For comparison's sake, Tarrant County (Fort Worth) also gained in both jobs/wages, but to a lesser degree (slightly above national average (4.77%) while gaining 8.12% of jobs.  It's dot on the chart is hanging out right next to Bexar County (San Antonio), which gained 9.39% in jobs and national average 3.4% in wages.

Now for some of the not-so-good outliers.  First, Clark County (Las Vegas) and Travis County (Austin) had quite a bit of job growth (comparatively), but they were low wage jobs as average salaries declined compared to nationwide.  Travis County ranked 2nd to last in wage growth/loss, losing -3.68%.  Clark gained 12.38% in jobs from 01-11 while it's wages were fairly flat (-0.18), but still fairly significantly below averages for both the nation and urbanized counties.

Also worth looking at is the curious case of Ohio, where the three largest urbanized counties Hamilton (Cincy), Franklin (Columbus), and Cuyahoga (Cleve-o) all lost significant amount of jobs, but each performed fairly well with regard to average salary increases.  Let's look at them individually:

County      City           Jobs           Wages
Hamilton    Cincy       -17.28%      7.60%
Franklin     Columbus -10.71%      3.68%
Cuyahoga   Cleve-o    -16.53%      4.66%

Unfortunately, we have to look at the terrible outliers.  I'm just going to come right out and tell you these are Wayne County (Detroit), Dallas County (sigh), and Jackson County (Kansas City).  DeKalb County in Atlanta would be on this list if it was kept separate from Fulton County.

It's likely not going to surprise you that Wayne County (Detroit) ranks last of all urbanized counties in both job gain/loss as well as wage gain/loss, losing 22.73% of its jobs while wages fell -4.78%.  Of all the urbanized counties, Dallas County ranks 2nd to last in job growth in this period, having lost a total of -266,195 jobs, which works out to -17.68% of its jobs.

As I've pointed out before, all of the other DFW area counties experienced job growth (as well as huge population growth).  The area gained 1.2 million people during this period, but Dallas County lost a near Detroit level of jobs.  Meanwhile, the suburban counties (who the cities compete with more than other cities, it's worth noting) gained huge amounts of job numbers.  Parker County went up 62.33%.  Rockwall went up 70.63%.  Denton up 47.07%.  And last, but certainly not least Collin County (where Toyota is relocating to) gained 46.68%.

In terms of wages, Dallas County grew by 1.63%, which is at least upward.  However, it's still well below the national average, even further below the urbanized county average, and ranks 6th to last of all urbanized counties in wage growth.  This is why I interpret the above chart to show Dallas County as the 2nd worst performing urbanized county in the U.S. with Jackson County (KC) close behind.  DeKalb is up there too (lost 22.17% of jobs but gained 2.38% in wages), but as it only represents half of ATL, Fulton County's positive performance ameliorates the bad.  Overall, Fulton/DeKalb gained significantly in average wages (7.32%).

Let's revisit that list:  Detroit, Dallas, Kansas City, and (half of) Atlanta.  Those are all areas with extremely high highway lane miles per capita.  Coincidence?  I think not.


Now for some more census graphics to visualize the What, Where, Why, and How of Dallas' job loss.

Below is a heat map for where Dallas urban core workers commute from in 2002 and 2011.  If you look closely you can see what I'm defining as the urban core with the little orange polygon.  I selected that area approximately to include uptown, downtown, all three major hospital areas, and Jefferson Street in Oak Cliff.  In that area in 2002 there were 226,000 jobs.  You can see that population density forms a natural gradient, densest towards the core and getting less and less dense as where downtown workers live radiates away.

Now below is 2011.  Interestingly, the urban core actually gained 5,000 jobs from 226k to 231k.  This is easier to see in animation (which I'll be creating), but you can still tell that the downtown commuter has dispersed as Dallas County has de-densified, or perhaps dis-urbanized.

Also worth noting the commute lengths to these 230,000 jobs has increased by 17.4% (which coincidentally also happens to be right about the percentage in Dallas county job loss). The average urban core worker in 2002 commuted 18.9 miles.  Meanwhile, in 2011 that number climbed to 22.2 miles.

Simply saying this is by preference is missing the larger point.  Yes, people are making the rational decision for themselves, but this pattern isn't repeated elsewhere (except for Detroit).  We're actively subsidizing disinvestment and decay in Dallas.

People left, then jobs follow.  That's the pattern.  All because we're suburbanizing the core when it wants to urbanize.  It's also the only way for the core to compete.  DC didn't try to be more like the suburbs.  It rejected proposed highways throughout.  Now, its suburbs are urbanizing to compete with it (see: Arlington and Bethesda).  That's a race to the top while DFW is racing to the bottom (with a few exceptions).

There is a misperception that Texas cities, DFW in particular, have sprawled because of the availability of land.  This is untrue.  The reality is that growth has several factors limiting what we might as well call carrying capacity.  They are water (or just basic needs for life, which might as well include air quality if we bring China into the equation), money or debt (because fewer people will want to relocate into a place with huge amounts of debt, which also has implications on maintaining lifestyle and what made the growth occur), and lastly, infrastructure.  All of these are interrelated.

Land isn't viable if it can't be reached.  Leaves don't grow without branches.  The suburbs only grew (often at the expense of the core cities) because policy makers, and by proxy, taxpayers agreed to build infrastructure out to the land.  The question is, can we afford to subsidize growth in the suburbs when we're actively hurting the core cities?  Can those suburbs even afford themselves long-term when all of that infrastructure has to be re-built with low density tax base?  I think the answer is no.