Factors Influencing the Long Term Sustainability of ...

Factors Influencing the Long Term Sustainability of Entrepreneurial

Technology Centers

Ross Gittell, Jeffrey Sohl and Edinaldo Tebaldi

University of New Hampshire

Objective of inquiry

To provide insights into the contributing factors to economic success in technology centers over the last business cycle (approximately 1990-2003)Explain and understand differences among USA tech centers in changes in employment and per capital income This inquiry is a follow-up to previous study (Gittell and Sohl, forthcoming) of the experience in USA technology centers in the early 2000s economic downturn

USA Entrepreneurial Technology Centers

Growth in high technology goods and services explained 65 percent of the difference between USA metropolitan areas with the fastest-growing economies and the average (Milken, 1999)How the Milken top ranked “tech poles” fared over the full course of the last business cycle and why some performed better than others is the focus of inquiry

Factors effecting tech pole economies during the tech downturn (Gittell & Sohl,

forthcoming)

Lack of diversification in overall economic baseLimited diversity within high technology industriesHigh average wagesHigh levels of venture capital funding during the end of the boom period of the late 1990s“Contradictions” in Tech Center Development. Some of the same factors that were negative influence in downturn over the full business cycle contributed to long-term economic growth

The Tech Poles ranked ordered in aggregate performance, LT and ST employment changes and per capita income change rankings. From strongest

performing to weakestGrowth Rankings 90-03 Employ. 2001-02 Employ. Per Capita Inc. AggregateTECH POLES Rank Rank Rank RankAustin 1 16 4 21Rochester 9 4 9 22Houston 10 8 5 23Denver 6 17 3 26San Diego 11 1 15 27Boise 2 14 12 28Atlanta 5 9 14 28Albuquerque 7 3 19 29Raleigh 4 11 18 33New Haven 21 7 8 36PhoeniX 3 10 23 36Newark 20 6 11 37Washington, DC 13 2 22 37Oakland 16 13 10 39Seattle 14 23 6 43Middlesex 12 19 13 44Dallas 8 20 16 44San Jose 19 25 1 45Boston 18 21 7 46San Francisco 22 24 2 48Philadelphia 25 5 21 51Orange County 15 12 24 51Chicago 17 18 17 52Los Angeles 24 15 25 64New York City 23 22 20 65

Creative Destruction Process in the Technology Centers

The tech poles with the greatest variance (ups and downs) in employment growth had the strongest overall growth performance over the full business cycle Among the tech poles with the greatest differential between long and short term rank all but one ranked in the top tier in employment growth 1990 to 2003 and three (of the eight among the top third) ranked among the top tier in per capita income growth. This suggests the force of the creative destruction process in the USA tech poles

Groupings of the USA tech centers --from most robust to least -- over the full course of the business cycle

(1) the high tech growth centers with employment concentrations in growth sectors within high technology, such as bio-technology and health care-related industries (e.g., San Diego, Rochester);(2) relatively recession resilient USA technology centers based in metropolitan areas with high concentrations of non-profit and technology related institutions, such as universities and governmental agencies (e.g., Raleigh,New Haven); 3) mature tech centers vulnerable to decline emanating from lack of diversification and high costs (e.g., Silicon Valley and Boston); (4) large metropolitan area centers lagging significantly behind in overall economic performance the other tech centers (e.g., NYC, Los Angeles, Chicago)  

Econometric Modeling

Dependent Variables Long-term (full 1990-2003/01 business cycle) growth in:

employment per capital income

Tested a range of explanatory variables suggested by theory and literature as effecting technology center economic growth

Econometric Model ResultsModel 1: Dependent Variable: Change in Employment, 90-03 Model 2 : Change in Personal Income, 90-01

 Variable

Employment,% change 1990-2003

 Variable

Personal Income,% change 1990-2001

Coefficient t-ratio Coefficient t-ratio

Constant 0.081 0.73 Constant 0.4735 4.25

High Tech Gini 1990 0.5355 3.32 Gini Supersector 1990 0.9761 2.37

Chg in Tax burden, 90-03

-10.139 -2.15 Tax Burden 03 -3.5260 -2.82

Chg home price 83-03

-0.0014 -2.13 Venture Capital Per Worker 93

0.00014 7.01

Adjusted R-squared 0.47   Adjusted R-squared 0.43  

Included observations 25   Included observations 25  

Method OLS   Method OLS  

Employment growth model: Significant explanatory variables

Specialization within high tech (as measured by gini coefficient) contributed to growth of employment over the last business cycle This is in contrast to earlier findings (Gittell and Sohl, forthcoming) that concentration within high technology contributed to pronounced employment decline during the economic downturnTen percent higher concentration within high tech activities in the early 1990s added approximately five percent to employment growth over the business cycleExamples of this are tech poles with the highest employment growth 90-2003 -- Austin and Boise.. had the highest concentration of employment within high technology

Employment growth model

Tech poles with lower growth in state and local tax burdens also had significantly higher growth rates in employment As the local and state tax burden increased 1 percent, the growth rate of employment decreased by approximately 10 percentPhiladelphia and Los Angeles had the lowest employment growth and were among the tech poles with the highest increase in state and local taxes

Employment growth…

The third significant variable was long term housing price increase, 1983 to 2002As the growth rate of housing price doubled the growth rate of employment decreased by approximately .15 percentRapid housing price increases appeared to dampen employment growth most in San Jose, San Francisco and Boston. These were the top three tech poles in housing price increase 1983 to 2003 and were among the slowest growth tech poles in employment

Per capita income model

The per capital income growth model identified three explanatory factors

local and state tax burden level (2003)general (super-sector) industry specialization (1990) venture capital flow at the beginning of the economic boom (high growth) period or 1993

Per capital income (pci)…Local and state tax burden level was significantLevel was more significant than rate of increase (which was significant variable in employment model)The pci model suggests that for each 1 percent increase in tax burden, there was approximately a 3.5 percent decline in growth rate of personal income per capital in the tech poles 1990 to 2001Raleigh and New Haven are examples of tech poles with high per capita income growth and relatively low state and local tax burden level

Per capital income..

Super-sector employment specialization. (This was not a significant factor in the employment growth model but within high technology industries concentration was) This finding is in contrast to earlier findings (Gittell and Sohl, forthcoming) that concentration of employment within super-sectors contributed to pronounced employment decline during the economic downturnWith every 1 percent increase in super-sector specialization (as measured by super-sector ginis) in 1990 personal income per capital increased approximately 1 percent Silicon Valley and San Francisco both had high super-sector concentration of employment (the 4th and 2nd highest) and had the highest growth in per capital income among tech poles

Per Capital Income…

Venture capital $’s per worker, 1993 This variable had no significant effect on employment growth over the full business cycle and had a negative effect on employment growth during the economic downturn (Gittell and Sohl, forthcoming)… as too much VC$ was “chasing” too few good start-up venturesAn increase of $1,000 in 1993 venture capital per worker increased personal income per capital by approximately .14 percent over the full business cycle in the tech polesBoston, San Jose and San Francisco all ranked among the highest tech poles venture capital per worker in 1993 and had significant growth in per capital income over the business cycle

Insignificant factors

There were several variables suggested by the literature as affecting growth in technology centers that did not have statistically significant effectQuality of life (as measured by Morgan and Morgan 2003) Diversity (as measured by Florida 2002 and including foreign-born and gay percentages) Housing prices was not significant in the per capita income model (it was in the employment model)

Main FindingThe Schumpterian (1934) process of creative destruction worked with strong force in USA technology centers in the 1990s and early 2000sThe core contributing factors to growth long run --such as concentration of employment in particular high technology employment sectors and high venture capital flow -- also had significant but negative effect during downturns

Leading examples of the creative destruction process: Silicon Valley and Boston compared to USA 1970 to 2003. Last “cycle” peak (2000) to

trough (2001) was shortest

Percent Change in Total Employment: 1970 to 2003 (33 years)

-15.0%

-10.0%

-5.0%

0.0%

5.0%

10.0%

15.0%

1 3 5 7 9 11 13

15

17

19

21

23

25

27

29

31

33

San Jose, CA Boston US

Factors influencing the creative destruction process and contributing to its shortening

Globalization of the economy and increased competition among cities More rapid product life cycles.. fostered by institutional process that facilitate accelerated innovation and commercialization (e.g., University R&D, venture capital)Accelerated process life cycles.. In the late 1990s..fast, broad and deep application of IT, Internet and now wireless technologies. USA tech centers were among “first adopters” of IT in broad range of industries but residents in other areas caught up fast… This was affected by rising education and tech know-how across the population“Sticky” factor prices (e.g., housing and taxes) in major tech centers

Future Inquiry More detailed “case” analysis of USA technology centers

This could involve inquiry over the last quarter century and even further back in time. More in-depth historical analysis should consider in detail changes in the character and length of business cycles in tech centers

Analysis of how global competition, technology change, product and process life cycles and local factor prices effect local development cycles Consideration of whether the dynamics in technology centers outside the USA were similar to what was observed in USA tech poles in the 1990s and early 2000s

Future Inquiry…

The character and pace of economic change in technology centers is an important area of inquiryMany useful insights can be gained from on-going detailed empirical study of technology centers in the USA and other nationsA challenge will be for the inquiry, and insights provided thereof, to keep pace with changes in the technology centers