Technology Sector Employment Statistics for the 12 major U.S. Technology Sector geographic regions (CSAs) are presented below for the period from January 1992 to December 2011. The data presented is derived from the most recent U.S. Bureau of Labor Statistics’ Quarterly Census of Employment and Wages (BLS/QCEW) released on June 28 2012.
An earlier version of this analysis was presented in April 2010 for the period of January 1990 to September 2009. At that time, the Great Recession was causing a notable drop in employment across all sectors of the U.S. economy including the Technology Sector. As we see below, the actual bottom of U.S. Technology Sector employment occurred four months later in January 2010.
Since the January 2010 bottom, Technology Sector employment overall has been gradually improving. Nonetheless, the summary observations from the April 2010 analysis were still largely valid. With the recent BLS/QCEW release, some regions are recovering to pre-recession employment levels, and thus, this updated look at the numbers is now warranted.
Note: These Summary Observations are more easily read and understood given the context of the charts and tables below. Thus, initially, you may want to skip this Summary, scan the data below, then return here.
• U.S. technology sector employment reached its low point in January 2010, down -7.8% from the pre-recession peak in July 2008. Since January 2010, U.S. technology sector employment has been recovering; by December 2011, it was down only -3.2%. If the 2011 employment trajectory continues, U.S. technology sector employment should recover to pre-recession levels by Q3 of 2012. Individual CSAs differ from the U.S. in their degree of losses during the recession and in their employment growth trajectories since January 2010.
• Technology sector employment in the San Jose/SF (encompassing Silicon Valley), Seattle, Atlanta, and Denver CSAs have recovered, and now exceed their pre-recession levels. The recovery in the Atlanta and Denver CSAs may be somewhat surprising given their relatively flat employment levels in the earlier 2000s.
• The San Jose/SF CSA showed the most rapid decrease in employment at the beginning of the recession across the CSAs; it was also among the deepest declines, -9.1% in January 2010 compared to July 2008, a loss of 31K jobs. Perhaps, this rapid decline was due to worries highlighted in Sequoia Capital’s presentation, titled “R.I.P. Good Times”, given to its portfolio companies at the beginning of October 2008, three weeks after the collapse of Lehman Brothers.
• Then, from January 2010 to December 2011, the San Jose/SF CSA showed the most rapid increase in employment, recovering to and exceeding pre-recession technology employment numbers, +2.5% compared to July 2008, adding 39K jobs since January 2010. Because this recovery began from a greater decrease in employment, there may appear to be more frenzy than would otherwise be apparent. Regardless, the sustained job growth and the large number of jobs created likely provides greater opportunity for those entering the technology sector workforce.
• The San Jose/SF CSA continues to have the highest concentration technology sector employment, in both absolute measure, 346K jobs, and relative measure, 3.1 LQ.
• As noted in the earlier analysis, for the full 20 year period, 1992-2011, the Washington DC and Seattle CSAs showed significantly greater technology sector job growth, +114K and +77K respectively. These are greater in both absolute and relative measures among the 12 CSAs. Both Washington DC and Seattle CSAs had the smallest declines in technology sector employment during the recession, -2.8% and -2.9% respectively. The Washington DC CSA has largely recovered to but not exceeded pre-recession employment levels. The Seattle CSA has exceeded its pre-recession employment levels, by +2.7%.
• Amazingly, the Seattle CSA has not only exceeded its pre-recession employment levels, it is the only CSA to exceed it’s peak dot com employment levels! Seattle had 10% greater technology sector employment in December 2011 than at its peak during the dot com bubble, December 2000. The Washington DC CSA is close to its dot com bubble technology sector employment, only -5% below, and is on trajectory to exceed those levels. In contrast, other CSAa are significantly below their dot com bubble levels; for example, the San Jose/SF, Boston, and New York CSAs are well below their dot com bubble peaks, by -29%, -28% and -26% respectively.
• The Minneapolis and Boston CSAs more closely tracked the U.S. technology sector trajectory of job losses and recovery. The Boston CSA lost fewer jobs relatively from the July 2008 to the January 2010 bottom, -6.3% for Boston compared to -7.8% for the U.S. The Boston CSA technology sector recovery lagged slightly behind the U.S. from July 2008 to December 2011, -3.9% vs -3.2%, and if the current job growth trajectory continues, it will be 2014 until the Boston CSA recovers to pre-recession July 2008 technology employment levels.
• The New York CSA also was among the CSAs with the deepest technology sector recession declines, down -9.5% on January 2010 compared to July 2008, losing 28K jobs, which is the second largest absolute loss among the 12 CSAs. The New York CSA technology sector recovery was on par relative to the U.S., down -3.2% in December 2011 compared to July 2008, but adding 19K jobs from January 2010 to December 2011, the second largest number of jobs added among the 12 CSAs. If this trajectory continues, the New York CSA technology sector will recover to pre-recession employment levels by the end of 2012. Thus, the New York CSA is similar to the San Jose/SF CSA in that the large number of recession job losses were followed by a large number of recovery job increases, by which the New York CSA technology recovery may similarly appear more frenzied.
• The Dallas and Chicago CSAs had the deepest relative technology sector recession declines; the bottom of their job losses came later than the U.S.; compared to July 2008 Dallas was down -10.5% in May 2010, and Chicago was down -9.8% in March 2010. Both have been recovering since their lows; in December 2011 compared to July 2008, Dallas is still down -4.7%, and Chicago is down -5.3%. If each maintains its 2011 job growth trajectory, each will recover to pre-recession technology employment levels during 2013.
• The Philadelphia and Los Angeles are the last 2 of the 12 CSAa. Both continued to lose technology sector jobs well beyond the U.S. bottom; their technology employment levels have stabilized somewhat in 2011, but recovery is not yet apparent. Philadelphia did have a noticeable increase in December 2011, but little can be extrapolated from the one data point. If recent trajectories continue, it seems that neither CSA will recover to pre-recession technology sector employment levels. If nothing changes, this may be a structural change to a lower concentration of technology sector employment in the Philadelphia and Los Angeles CSAs.
Chart 1 shows technology sector employment for the 12 Combined Statistical Areas (CSAs) that have technology sector employment greater than 80,000 jobs; these are the same 12 CSAs analyzed in April 2010. Only private sector employment numbers are used.
Chart 1 covers the period from January 1992 to December 2011, which includes the most recent data available in Quarterly Census of Employment and Wages from the U.S. Bureau of Labor Statistics. The order of CSAs listed in the legend is the rank order of the CSAs’ technology sector employment in December 2011, which is the final data point in the chart.
Demographic data for the 12 CSAs is shown below.
Click on chart to see full size.
As before, the most notable feature in the chart is the burst in technology sector employment during the dot com bubble from 1996 to 2003; this bubble distorts employment and industry trends. Nonetheless, time has progressed far enough from the bubble that notable trends have emerged.
Also apparent is the drop in technology sector employment due to the current Great Recession. For the entire U.S., Technology Sector employment peaked in July 2008 prior to the recession, and bottomed in January 2010 during the recession. Since the bottom, technology sector employment has increased gradually. Some CSAs have seen better recovery than others; this is explored further below.
The San Jose/San Francisco CSA (dark blue) has the greatest number of technology sector employees. Followed by the New York CSA (pink) the Washington D.C. CSA (orange) the Los Angeles CSA (violet) the Boston CSA (turquoise) and so on.
A few notable trends or characteristics. The Washington D.C. CSA and Seattle CSA show significant sustained job growth over the 20 tear period. The Washington D.C. CSA went from 5th rank in 1992 to 3rd rank in 2011, adding ~114K technology sector jobs, overtaking the Boston CSA and Los Angeles CSA. The Seattle CSA went from 12th to 7th rank, adding ~77K technology sector jobs, and more than doubling the area’s number of technology sector employees.
With respect to total population, the San Jose/San Francisco CSA and Boston CSA are roughly comparable in size, and the Washington D.C. CSA is only slightly larger (+15%). The New York CSA and Los Angeles CSA have much larger populations (>2.5x) than the other CSAs. They do have substantial technology sector employment, but at a lower relative density to several of the other CSAs. Discussions of CSA demographics and of the relative density of technology sector employment is below.
Technology Sector employment is highly concentrated in these 12 CSAs. While the 12 CSAs represent ~34% of the U.S. population, they represent ~49% of all U.S. technology sector employment. This proportion has been fairly consistent during the 20 year period, varying between 48% and 51.5%.
After the top 12 CSAs, technology sector employment drops off very rapidly. As of December 2011, CSAs ranked 12-24 together add ~492K technology sector employees or an additional ~12%. And, the next 100 CSAs, ranked 25-124 add another ~512K technology sector employees or an additional 12%. Thus, the top 12 CSAs dominate technology sector employment.
In the charts and tables presented here, technology sector jobs are those classified by the following NAICS codes:
|NAICS 334||Computer and electronic product manufacturing||25.5%|
|NAICS 518||Data processing, hosting and related services||4.8%|
|NAICS 519||Other information services||5.0%|
|NAICS 5112||Software publishers||8.3%|
|NAICS 5415||Computer systems design and related services||38.9%|
This list of 6 NAICS codes is purposely a more focused definition of technology sector than is often used, but it does broadly represent electronics, computers and computing, communications, and Internet technologies and business activities, and it does capture, as much as possible, employment related to the formulation and growth of these industries.
The percentage in the right column is the portion that the single NAICS industry represents of employment for the total of the 6 NAICS industries across all 12 major CSAs. Individual CSAs may vary from these aggregate proportions.
Before delving into more detail on the employment numbers, the next three tables (1a, 1b, 1c) provide more context for the 12 CSAs by listing most recently available selected U.S. Census Bureau statistics. For easier comparison, the last two rows of each table list the technology sector employment and the location quotient as of December 2011 derived from Chart 1 and Chart 10 data.
Location quotient (LQ) is discussed more fully below. Briefly, technology sector LQ is the concentration of technology sector employment in a CSA relative to the concentration of technology sector employment across the U.S. as a whole. An LQ greater than 1.0 means that the CSA has proportionally greater technology sector employment than the U.S. as a whole and that the technology sector is more important to the CSA’s economy; an LQ less than 1.0 means proportionally less employment concentration and less economic importance.
A few additional items to note. The San Jose/SF CSA is roughly comparable to the Boston CSA, roughly the same population, roughly the same land area, etc. The Washington DC CSA has a similar but slightly higher population. Seattle CSA population is one of the smaller of the 12 CSAs.
One notable difference is that San Jose/SF has the highest median house value, significantly higher than Boston, Washington DC, and Seattle, and higher than the 2nd and 3rd most costly CSAs, Los Angeles and New York. Though housing costs may become a consideration for later stage expansion of companies in the San Jose/SF area, housing costs have not been much of a deterrent to starting and growing major successful technology companies over the past 20 years, see S&P 1500 data.
Dallas, Washington DC, Seattle, and Denver CSAs have the highest relative number of new housing permits.
|San Jose-SF||Seattle||Washington DC||Boston|
|Tech Empl, Dec 2011||345,879||127,209||263,454||210,059|
|Tech LQ, Dec 2011||3.11||2.07||2.03||1.70|
|Population est July 1 2011||7,563,460||4,269,349||8,718,083||7,601,061|
|Education BA+, 25+yr, %, 2006-10||41.3%||35.1%||42.3%||37.5%|
|Housing units, 2010||2,908,294||1,803,069||3,461,848||3,170,897|
|Owner-occupied housing % 2006-10||58.1%||63.8%||66.6%||64.2%|
|Median house value 2006-2010||646,012||345,138||383,295||344,916|
|New house building permits 2011||10,261||13,275||27,185||9,073|
|Median household income, 2006-10||76,600||64,035||80,795||66,236|
|Private nonfarm estblmt 2009||194,655||115,257||212,389||198,561|
|Private nonfarm employment 2009||3,047,403||1,663,283||3,557,236||3,390,892|
|Land area, square miles, 2010||8,741||9,888||9,999||8,796|
|Population per square mile, 2010||854.5||424.7||857.4||859.4|
|Tech Empl, Dec 2011||84,595||151,715||105,542||84,649|
|Tech LQ, Dec 2011||1.76||1.49||1.36||1.34|
|Population est July 1 2011||3,157,520||6,887,383||5,712,148||3,655,558|
|Education BA+, 25+yr, %, 2006-10||38.4%||30.4%||33.3%||36.1%|
|Housing units, 2010||1,302,189||2,670,033||2,308,439||1,493,637|
|Owner-occupied housing % 2006-10||66.6%||63.6%||67.7%||72.8%|
|Median house value 2006-2010||253,733||142,627||193,020||235,440|
|New house building permits 2011||8,223||25,106||8,999||5,630|
|Median household income, 2006-10||61,400||57,037||57,820||64,350|
|Private nonfarm estblmt 2009||90,461||147,196||137,520||99,587|
|Private nonfarm employment 2009||1,263,691||2,637,451||2,181,147||1,775,645|
|Land area, square miles, 2010||13,060||14,059||10,376||9,523|
|Population per square mile, 2010||236.7||478.8||541.5||379.7|
|Los Angeles||Philadelphia||New York||Chicago|
|Tech Empl, Dec 2011||216,014||88,312||290,473||119,707|
|Tech LQ, Dec 2011||0.95||0.93||0.92||0.83|
|Population est July 1 2011||18,081,569||6,562,287||22,214,083||9,729,825|
|Education BA+, 25+yr, %, 2006-10||28.0%||31.1%||35.2%||33.0%|
|Housing units, 2010||6,276,022||2,654,272||8,820,997||3,890,941|
|Owner-occupied housing % 2006-10||55.8%||69.6%||55.4%||67.8%|
|Median house value 2006-2010||484,162||232,534||464,269||253,887|
|New house building permits 2011||19,551||7,467||25,096||7,798|
|Median household income, 2006-10||59,891||61,317||66,752||60,979|
|Private nonfarm estblmt 2009||418,604||157,105||614,903||242,156|
|Private nonfarm employment 2009||6,285,022||2,671,180||8,594,442||3,988,657|
|Land area, square miles, 2010||33,955||5,942||11,793||8,472|
|Population per square mile, 2010||526.5||1,099.5||1,872.8||1,143.4|
Table 2 shows technology sector employment numbers from Chart 1 for 5 specific dates:
|January 1992||begins the 20 year data set used here|
|February 2004||minimum U.S. tech sector employment after the dot com bubble|
|July 2008||peak employment prior to the decline due to the Great Recession|
|January 2010||minimum employment during the recession|
|December 2011||ends the data set showing a beginning of job recovery|
The dates for the minimums and maximums are determined by using the total U.S. technology sector employment numbers. The dates also happen to coincide with the minimum and maximum of the 12 CSAs in aggregate. The actual minimums and maximums for an individual CSA may differ from the aggregate minimums and maximums; these differences generally are not notable; any notable differences are discussed when appropriate.
Table 2 also shows technology sector employment for the 12 CSAs combined and for the entire U.S.
|CS488 San Jose-SF CSA||285,936||314,831||337,517||306,645||345,879|
|CS408 New York CSA||291,901||283,860||299,956||271,518||290,473|
|CS548 Washington DC CSA||149,662||237,619||265,068||257,552||263,454|
|CS348 Los Angeles CSA||271,703||239,873||238,600||217,506||216,014|
|CS148 Boston CSA||221,759||204,548||218,492||204,820||210,059|
|CS206 Dallas CSA||120,241||154,336||159,222||143,768||151,715|
|CS500 Seattle CSA||49,855||96,208||123,809||120,214||127,209|
|CS176 Chicago CSA||120,823||121,147||126,420||114,688||119,707|
|CS122 Atlanta CSA||62,094||101,300||102,369||98,260||105,542|
|CS428 Philadelphia CSA||84,965||87,994||95,661||87,867||88,312|
|CS378 Minneapolis CSA||68,192||82,801||87,021||80,566||84,649|
|CS216 Denver CSA||54,076||81,101||83,363||78,925||84,595|
Trends and changes across CSAs might be a little more apparent in Table 3 which shows the difference in technology sector employment, jobs added in black and lost in red, between several pairs of the above dates:
|January 1992||– December 2011||from start to end of the 20 year data set|
|January 1992||– July 2008||from start to the pre-recession employment peak|
|February 2004||– December 2011||from post bubble low to the end of the data set|
|July 2008||– January 2010||from pre-recession high to the recession low|
|January 2010||– December 2011||from recession low to the end of the data set|
|July 2008||– December 2011||from pre-recession high to the end of the data set|
|CS488 San Jose-SF CSA||59,943||51,581||31,048||-30,872||39,234||8,362|
|CS408 New York CSA||-1,428||8,055||6,613||-28,438||18,955||-9,483|
|CS548 Washington DC CSA||113,792||115,406||25,835||-7,516||5,902||-1,614|
|CS348 Los Angeles CSA||-55,689||-33,103||-23,859||-21,094||-1,492||-22,586|
|CS148 Boston CSA||-11,700||-3,267||5,511||-13,672||5,239||-8,433|
|CS206 Dallas CSA||31,474||38,981||-2,621||-15,454||7,947||-7,507|
|CS500 Seattle CSA||77,354||73,954||31,001||-3,595||6,995||3,400|
|CS176 Chicago CSA||-1,116||5,597||-1,440||-11,732||5,019||-6,713|
|CS122 Atlanta CSA||43,448||40,275||4,242||-4,109||7,282||3,173|
|CS428 Philadelphia CSA||3,347||10,696||318||-7,794||445||-7,349|
|CS378 Minneapolis CSA||16,457||18,829||1,848||-6,455||4,083||-2,372|
|CS216 Denver CSA||30,519||29,287||3,494||-4,438||5,670||1,232|
Very apparent are the notable and sustained job growths over 20 years in the Washington DC CSA (~114K jobs added) and in the Seattle CSA (~77K jobs added). The Seattle CSA has approximately half the population of the Washington DC CSA, making its job growth that much more significant. The data here only counts private sector jobs, not government jobs. Though many of the DC area jobs likely directly or indirectly support government activities and programs, they are private sector jobs nonetheless.
The Atlanta and Denver CSAs also show job growth over the 20 year period, but their growth mostly occurred in the 1990s, and has been relatively flat in the 2000s. However, Atlanta and Denver CSAs do show surprising recent job growth from the recession low at January 2010 to December 2011 and from pre-recession peak at July 2008 to December 2011.
All CSAs, except Los Angeles, added jobs from the current recession minimum in January 2010 to December 2011.
Only 4 CSAs, San Jose/SF, Seattle, Atlanta, and Denver have recovered to pre-recession employment levels, with more jobs in December 2011 than at the pre-recession peak employment in July 2008.
Only 4 CSAs, Los Angeles, Boston, New York, and Chicago have net job losses for the 20 year period from January 1992 to December 2011.
Only 3 CSAs, Los Angeles, Dallas, and Chicago have fewer technology sector jobs in December 2011 than at the post dot com bubble minimum of February 2004.
The Los Angeles CSA stands out as the only CSA of the 12 that is consistently showing job losses across all of the time periods and a downward trend in total technology sector employment.
Table 4 shows the annualized growth rate (CAGR) of technology sector jobs during the 6 time periods for each of the 12 CSAs, the 12 CSAs in aggregate, and the U.S. as a whole. Of course, a begin-to-end calculated growth rate averages out monthly and yearly fluctuations, but it does allow comparing growth between CSAs for these time periods.
|CS488 San Jose-SF CSA||0.96%||1.01%||1.21%||-6.19%||6.48%||0.72%|
|CS408 New York CSA||-0.02%||0.17%||0.29%||-6.42%||3.58%||-0.94%|
|CS548 Washington DC CSA||2.88%||3.52%||1.33%||-1.90%||1.19%||-0.18%|
|CS348 Los Angeles CSA||-1.15%||-0.78%||-1.33%||-5.98%||-0.36%||-2.87%|
|CS148 Boston CSA||-0.27%||-0.09%||0.34%||-4.22%||1.33%||-1.15%|
|CS206 Dallas CSA||1.17%||1.72%||-0.22%||-6.58%||2.85%||-1.40%|
|CS500 Seattle CSA||4.82%||5.67%||3.63%||-1.95%||2.99%||0.80%|
|CS176 Chicago CSA||-0.05%||0.27%||-0.15%||-6.29%||2.26%||-1.58%|
|CS122 Atlanta CSA||2.70%||3.08%||0.53%||-2.69%||3.80%||0.90%|
|CS428 Philadelphia CSA||0.19%||0.72%||0.05%||-5.51%||0.26%||-2.31%|
|CS378 Minneapolis CSA||1.09%||1.49%||0.28%||-5.01%||2.61%||-0.81%|
|CS216 Denver CSA||2.27%||2.66%||0.54%||-3.58%||3.69%||0.43%|
The growth rate numbers in Table 4 reinforce the observations made for Table 3 jobs added/lost numbers.
The Seattle CSA (4.82%) and the Washington DC CSA (2.88%) show high sustained growth across the 20 year period.
Not surprisingly, all CSAs show negative growth (loss) from the July 2008 pre-recession peak to the January 2010 recession low. However, the Washington CSA and the Seattle CSA show relatively lower losses, indicating relative job resilience during the recession. Though not as dramatic, Atlanta, Denver, and Boston CSAs also show lower relative losses compared to the U.S. as a whole and the 12 CSAs in aggregate. Dallas, New York, Chicago, and San Jose-SF CSAs show more severe relative losses compared to the U.S. as a whole.
Of the 4 CSAs that show positive growth from the July 2008 pre-recession high to December 2011, the Atlanta, Seattle, and Denver CSAs are in the group showing relatively lower recession losses. Whereas the San Jose CSA moved from severe relative recession losses to positive growth.
Chart 2a shows technology sector employment for the 12 CSAs from January 2010 to December 2011, from the low of technology sector employment during the current recession to the end of the available data. The chart’s data is indexed to January 2010, so that the relative growths can be compared. Also, shown is a white line for the U.S. technology sector as a whole, so that CSAs can be compared to the overall U.S. technology sector.
Click on chart to see full size.
A very notable one month dip in employment is apparent in August 2011 across the New York (pink), Boston (turquoise), Washington DC (orange), and Philadelphia (lime) CSAs. This decrease is in the Telecom Sector (NAICS 517), and is most likely entirely due to a temporary work stoppage at Verizon. Most of the same workers returned to work in September 2011. Seemingly all of the August 2011 dip in technology employment in the U.S. as a whole is also due to this work stoppage.
In order to more easily see data trends, Chart 2b shows the same data as Chart 2a, but averages out the August 2011 dip in the effected CSAs and in the total U.S. technology sector employment. The same August 2011 adjustment is made in subsequent charts.
Click on chart to see full size.
Now, it is more apparent that the 12 CSAs fall into roughly 4 clusters in their rate of recovery from the bottom of he recession. The San Jose CSA (dark blue) stands out from the other CSAs, growing more than 12.8% during the 2 year period. Seven CSAs, Atlanta (plum), Denver (light yellow), New York (pink), Seattle (blue), Dallas (tan), Minneapolis (light orange), and Chicago (teal), ranging 4.4% to 7.4%, show better than or comparable growth to the U.S. technology sector as a whole. The Boston (turquoise) and Washington DC (orange) CSAs, ~2.4%, are lagging the U.S. as a whole. And the Philadelphia CSA (lime), slight above 0%, and the Los Angeles CSA (violet), slightly below 0%, lagging further still.
Chart 3 shows the same data as Chart 2b, but in addition to being indexed to January 2010, it is normalized to the U.S. technology sector as a whole. In this form, it might be easier to see the differences among the CSAs for the two year period.
Click on chart to see full size.
As noted earlier, the 12 CSAs had different relative losses from the pre-recession maximum in July 2008 to the recession minimum in January 2010. Thus, a portion of the San Jose-SF CSA gain from January 2010 to December 2011 is because it started from a relatively lower base. Charts 4 and 5 are similar to Charts 2b and 3, but start at the pre-recession high at July 2008.
Chart 4 shows technology sector employment for the 12 CSAs from the pre-recession peak at July 2008 to December 2011. The chart’s data is indexed to July 2008, and as in Chart 2b above, the data at August 2011 is averaged to remove the temporary work stopage.
Click on chart to see full size.
The San Jose-SF CSA (dark blue) is the lowest line on the left half of the timeline, most apparent from April to October 2009. After hitting its January 2010 low, San Jose-SF CSA employment increases relatively rapidly to end as the third highest data point at December 2011, 2.5% above its pre-recession high at July 2008.
The relatively rapid fall in technology sector employment in the San Jose-SF CSA compared to the other CSAs from July 2008 to January 2010 might be related to the general psychological mood at the time, which was perhaps most typified in a widely circulated presentation titled “R.I.P. Good Times” prepared by Sequoia Capital on October 6 2008 for its portfolio companies. Global financial markets had been gradually spiraling down since late 2007, and were scrambling during the summer of 2008; then, Lehman Brothers filed bankruptcy on September 15 2008; and global financial systems were getting slammed. Three weeks after Lehman’s collapse, the Sequoia presentation painted a grim financial picture for venture funds and venture backed companies. There was a lot of talk about panic, if not actual panic across the technology community in Silicon Valley and beyond.
Three CSAs, Atlanta (plum), Seattle (blue), Denver (light yellow), and Washington DC (orange), did consistently better overall than the U.S. as a whole. All four had smaller relative losses to the U.S. for the 3.5 year period. Atlanta, Seattle, and Denver finished the period higher than their July 2008 employment levels. Washington DC lost the fewest jobs relatively, but hasn’t yet reached its pre-recession peak. The Washington DC technology sector may have been protected from early losses by government spending, and then, may have slowed in recovery as more policy pressure has been placed on reducing government spending.
Chart 5 shows the same technology sector employment data as Chart 4 indexed to the pre-recession peak at July 2008, but in addition, it normalizes CSA data to the U.S. technology sector as a whole. Thus, Chart 5 shows CSA employment relative to the U.S.
Click on chart to see full size.
The Atlanta (plum), Seattle (blue), Denver (light yellow), and Washington DC (orange) CSAs consistently show relatively higher technology sector employment relative to the U.S. as a whole. Though they had job losses, their losses were relatively less than the U.S. as a whole.
As mentioned before, the San Jose/Sf CSA (dark blue) shows the widest swing in relative employment, falling more rapidly after July 2008 and then rising more rapidly after late 2009.
The Minneapolis (light orange) and Boston (turquoise) CSAs roughly track the U.S. as a whole.
The New York CSA (pink) is consistently below the U.S. during most of the period, but recovers to match the U.S. toward the end of 2011.
Dallas (tan), Chicago (teal), Philadelphia (lime), and Los Angeles (violet) CSAs consistently show relatively lower technology sector employment relative to the U.S. as a whole. The 4 CSAs lost jobs at a greater rate relative to the U.S., and have not yet recovered to match the U.S. The Philadelphia and Los Angeles CSAs show the lowest relative employment, and they continue to further lag the U.S. The Chicago CSA, after the relative losses at the beginning of the period, appears to stabilize. The Dallas CSA shows significant relative job losses into 2010, then gradual relative improvement, but not enough to match the U.S. as a whole.
Note again that Chart 5 shows the CSA technology sector employment data that is normalized to the U.S. as a whole. The U.S. has 3.2% less technology sector employment in December 2011 than in the pre-recession peak in July 2008. Only 4 CSAs have recovered to pre-recession employment levels.
The next 4 charts are similar to the previous charts, but for completeness, show longer periods. These charts further reinforce many of the observations above. Only a few notable observations will be repeated.
Charts 6 and 7 show the CSAs from the post dot com bubble minimum at February 2004 to December 2011. Chart 6 is indexed to February 2004. Chart 7 is indexed to February 2004 and normalized to the U.S. technology sector.
The Seattle CSA (blue) shows the most dramatic relative increase in technology sector employment since February 2004. The Washington DC CSA (orange) also shows a significant relative increase. The San Jose/SF CSA (dark blue) roughly tracks the U.S. through 2009, then breaks away from other CSAs. The other CSAs cluster around or below the U.S. as a whole.
Click on chart to see full size.
Click on chart to see full size.
Charts 8 and 9 show the CSAs for the entire 20 year period from January 1992 to December 2011. Chart 8 is indexed to January 1992. Chart 9 is indexed to January 1992 and normalized to the U.S. technology sector.
Again, the Seattle CSA (blue) shows the most dramatic relative increase (77K jobs added) across the 20 tear period, with the Washington DC CSA (orange) also showing a significant relative increase (114K jobs added).
Strikingly apparent in Chart 8, Seattle is the only CSA to have exceeded its peak technology employment levels during the dot com bubble, c.2000, and has done so by 10%. The Washington DC CSA is within 5% of its dot com bubble peak employment levels, and could exceed those as well. Other CSAs are typically 25-30% below their dot com bubble levels.
The relative gains in the Atlanta (plum) and Denver (light yellow) CSAs in the 1990s and early 2000s and the leveling off in the 2000s is more clearly apparent here, especially in Chart 9.
Click on chart to see full size.
Click on chart to see full size.
Location quotient (LQ) here is the ratio of the concentration of technology sector employment in a CSA compared to the concentration of technology sector employment across the U.S. as a whole:
lq = ( csa.tech_empl / csa.all_empl ) / ( US.tech_empl / US.all_empl )
Thus, if a CSA’s LQ is greater than 1.0, then the CSA’s ratio of technology sector employment to its employment across all sectors is greater than the ratio for the U.S. as a whole of technology sector employment to employment in all sectors. Conversely, iF a CSA’s LQ is less than 1.0, then the CSA’s ratio of technology sector employment to its employment across all sectors is less than the ratio for the U.S. of technology sector employment to employment in all sectors.
Location quotient is an indication of an industry sector’s relative importance to a region’s overall economy.
Chart 10 shows the technology sector location quotient for the 12 CSAs from January 1992 to December 2011.
Click on chart to see full size.
Immediately apparent is the San Jose/SF CSA’s significantly higher technology sector LQ, starting at 2.77 in 1992 and rising to 3.10 at the end of 2011. The San Jose/SF region has substantially higher LQ than other regions, and has continued to increase its LQ over the past 20 years.
Most of the 12 major CSAs maintained location quotients that were greater than 1.0.
The location quotient curves do not show a marked effect from the dot com bubble. This indicates that the dot com bubble was not localized to particular regions but was more pervasive across all regions and across the U.S. as a whole.
The Seattle (blue) and the Washington DC (orange) CSAs show significant increases. Seattle roughly doubles its location quotient, rising from 1.02 to 2.07, and rises from 9th ranked to 2nd ranked. Washington DC increases its location quotient from 1.47 to 2.03, and rises from 5th ranked to 3rd ranked. Washington DC was 2nd ranked for much of the 2000s, but was gradually superseded by Seattle in 2010.
The Boston CSA (turquoise) started as the 2nd ranked location quotient at 2.04, but gradually decreased to 5th rank at 1.70.
The Los Angeles CSA (violet) also shows a notable decrease in LQ ranking from 7th at 1.29 to 9th at 0.95.
Table 5 shows the location quotient for the 5 transition dates used above for each of the 12 CSAs and for the aggregate of the 12 CSAs. Also, shown is the LQ for the U.S. as a whole, which is of course always 1.00. A location quotient that is less than 1.0 is shown in a red font color.
|CS488 San Jose-SF CSA||2.771||2.881||2.913||2.903||3.106|
|CS408 New York CSA||1.011||0.943||0.948||0.908||0.919|
|CS548 Washington DC CSA||1.474||1.927||2.018||2.076||2.032|
|CS348 Los Angeles CSA||1.287||1.036||1.001||0.995||0.951|
|CS148 Boston CSA||2.039||1.714||1.738||1.727||1.702|
|CS206 Dallas CSA||1.654||1.689||1.559||1.502||1.487|
|CS500 Seattle CSA||1.017||1.709||1.929||2.052||2.068|
|CS176 Chicago CSA||0.879||0.844||0.840||0.838||0.831|
|CS122 Atlanta CSA||1.119||1.322||1.280||1.333||1.358|
|CS428 Philadelphia CSA||0.959||0.931||0.972||0.948||0.925|
|CS378 Minneapolis CSA||1.303||1.365||1.347||1.356||1.340|
|CS216 Denver CSA||1.544||1.805||1.685||1.739||1.762|
Table 6 shows the percentage change in location quotient for the 6 time periods used above. This is the total change for the time period; the annualized change is shown in Table 7. A decrease in location quotient across a time period is shown in red.
|Percent Change LQ||1992.01||1992.01||2004.02||2008.07||2010.01||2008.07|
|CS488 San Jose-SF CSA||12.09%||5.14%||7.82%||-0.35%||6.99%||6.61%|
|CS408 New York CSA||-9.10%||-6.26%||-2.55%||-4.22%||1.25%||-3.03%|
|CS548 Washington DC CSA||37.84%||36.92%||5.44%||2.84%||-2.11%||0.67%|
|CS348 Los Angeles CSA||-26.07%||-22.22%||-8.13%||-0.63%||-4.35%||-4.95%|
|CS148 Boston CSA||-16.53%||-14.75%||-0.73%||-0.64%||-1.45%||-2.09%|
|CS206 Dallas CSA||-10.05%||-5.75%||-11.93%||-3.63%||-0.98%||-4.57%|
|CS500 Seattle CSA||103.31%||89.60%||21.01%||6.41%||0.77%||7.23%|
|CS176 Chicago CSA||-5.47%||-4.41%||-1.47%||-0.25%||-0.87%||-1.11%|
|CS122 Atlanta CSA||21.31%||14.31%||2.74%||4.15%||1.89%||6.12%|
|CS428 Philadelphia CSA||-3.50%||1.35%||-0.55%||-2.43%||-2.41%||-4.78%|
|CS378 Minneapolis CSA||2.80%||3.35%||-1.81%||0.70%||-1.23%||-0.54%|
|CS216 Denver CSA||14.10%||9.07%||-2.39%||3.21%||1.36%||4.61%|
Table 7 shows the annualized percentage change in location quotient for the 6 time periods used above. A negative change is shown in red.
|Annualized % Change LQ||1992.01||1992.01||2004.02||2008.07||2010.01||2008.07|
|CS488 San Jose-SF CSA||0.57%||0.30%||0.97%||-0.23%||3.59%||1.89%|
|CS408 New York CSA||-0.48%||-0.39%||-0.33%||-2.83%||0.65%||-0.90%|
|CS548 Washington DC CSA||1.62%||1.92%||0.68%||1.88%||-1.10%||0.20%|
|CS348 Los Angeles CSA||-1.50%||-1.51%||-1.08%||-0.42%||-2.29%||-1.47%|
|CS148 Boston CSA||-0.90%||-0.96%||-0.09%||-0.43%||-0.76%||-0.62%|
|CS206 Dallas CSA||-0.53%||-0.36%||-1.61%||-2.43%||-0.51%||-1.36%|
|CS500 Seattle CSA||3.63%||3.95%||2.46%||4.23%||0.40%||2.06%|
|CS176 Chicago CSA||-0.28%||-0.27%||-0.19%||-0.16%||-0.45%||-0.33%|
|CS122 Atlanta CSA||0.97%||0.81%||0.35%||2.75%||0.98%||1.75%|
|CS428 Philadelphia CSA||-0.18%||0.08%||-0.07%||-1.63%||-1.26%||-1.42%|
|CS378 Minneapolis CSA||0.14%||0.20%||-0.23%||0.47%||-0.64%||-0.16%|
|CS216 Denver CSA||0.66%||0.53%||-0.31%||2.13%||0.71%||1.33%|
References and Notes
Quarterly Census of Employment and Wages.
The BLS/QCEW data set provides detailed monthly employment data for every county across the U.S. classified by industry segments using NAICS industry codes. The BLS/QCEW is reported quarterly, and is published approximately 6 months after the close of the reported quarter. This 6 month lag is the tradeoff for gaining access to this granular level of detail. The BLS provides various interactive tools for perusing the data.
BLS QCEW Databases.
BLS Location Quotient Calculator.
Update of Statistical Area Definitions and Guidance on Their Uses
U.S. Office of Management and Budget maintains the definitions of regional statistical areas. Roughly speaking, a Combined Statistical Area (CSA) contains 2 or more Core Based Statistical Areas (CBSA). A CBSA is designated as either a Metropolitan Statistical Area (MSA) or Mircopolitan Statistical Area (μSA). A CBSA is made up of a collection of counties. The most recent area definitions are listed in the above Bulletin.
“Combined Statistical Areas can be characterized as representing larger regions that reflect broader social and economic interactions, such as wholesaling, commodity distribution, and weekend recreation activities, and are likely to be of considerable interest to regional authorities and the private sector.”
There are 128 CSAs, which represent ~2/3 of the U.S. population. Not all counties in the U.S. are included in a CSA. For example, San Diego is not part of a CSA. Austin County TX was not part of a CSA until a new CSA was designated a couple of years ago in OMB 10-2 for Austin and the greater Austin area.
A CSA can extend beyond state boundaries. For example, the Boston CSA includes most of eastern Massachusetts plus southern New Hampshire and Rhode Island. The New York CSA includes New York City, Long Island, Southern Connecticut, Northern New Jersey, and nearby portions of Pennsylvania.
The San Jose/San Francisco CSA includes the counties north and south of San Francisco including Santa Clara County and Silicon Valley.
Wikipedia provides an easily accessible list of U.S. CSAs, though the OMB Bulletin is the definitive reference.
NAICS classifies all business establishments in a hierarchical 2 to 6 digit code. NAICS has replaced Standard Industrial Classification (SIC) codes in most uses; though, SIC codes continue to be used by the U.S. Securities and Exchange Commission (SEC). There are mappings between NAICS and SIC.
The U.S. Census Bureau’s American Fact Finder and QuickFacts provide access to all of the underlying demographic data presented above and more.
Approximately 45,000 Verizon employees were on strike during the period when the Department of Labor collected August jobs data. The employees returned to work at the end of August.
From the Valley Comes a Warning. The New York Times. 2008.11.16.
“Maybe it was the tombstone that did it. Even as Wall Street burned, Silicon Valley seemed strangely sanguine. Then a PowerPoint presentation from Sequoia Capital prophesying Armageddon — featuring an “R.I.P. Good Times” headstone — made the rounds. A month later, venture capital firms are slashing investments and counseling portfolio companies to cut jobs.” – NYT