slide1
Download
Skip this Video
Download Presentation
Determinants of New Technology-Based Firms Performance in Catch-Up Regions: Evidence from the U.S. Biopharmaceutical and IT Service Industries 1996-2005

Loading in 2 Seconds...

play fullscreen
1 / 41

Determinants of New Technology-Based - PowerPoint PPT Presentation


  • 229 Views
  • Uploaded on

Determinants of New Technology-Based Firms Performance in Catch-Up Regions: Evidence from the U.S. Biopharmaceutical and IT Service Industries 1996-2005. Wenbin Xiao December 16, 2009.

loader
I am the owner, or an agent authorized to act on behalf of the owner, of the copyrighted work described.
capcha
Download Presentation

PowerPoint Slideshow about 'Determinants of New Technology-Based' - emily


An Image/Link below is provided (as is) to download presentation

Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author.While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server.


- - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - -
Presentation Transcript
slide1

Determinants of New Technology-Based Firms Performance in Catch-Up Regions:Evidence from the U.S. Biopharmaceutical and IT Service Industries 1996-2005

Wenbin Xiao

December 16, 2009

1 objective
Investigate how region-level factors affect the early stage-performance of New Technology-based Firms (NTBF) in catch-up regions1. Objective

I. Introduction

2 research questions
First, why do some catch-up regions succeed while others continue to lag behind?

Second, besides the industrial cluster size, what other location-specific factors matter?

Finally, how do the causal patterns vary between the biopharmaceutical and IT service industry?

2. Research Questions
3 why important
Fostering “homegrown” NTBFs has become a popular strategy to reinvigorate a regional economy (Ellison & Glaeser, 1997; Feldman and Francis 2004)

The conventional wisdom posits that location enhances NTBF performance by industrial clustering (Marshall 1920, Krugman 1991, Porter 1990 1998)

Recent evidence suggests that some catch-up regions have better average NTBF performance than leading regions (Sorenson & Audia 2000, Florida 2002, Stuart & Sorenson 2003, Folta et al. 2006)

‘The best practice’ in leading regions may not be applicable to ‘catch-up’ regions, but there are few studies focusing on catch-up regions (Todtling & Trip, 2005).

3. Why Important?
slide5

+

Scientist job market

Attracting technological

entrepreneurs

+

Venture capital

+

Cultural diversity

Average early-stage NTBFs performance in a region

+

Academic research

Facilitating

Radical Innovations

+

Industrial structure

+

Entrepreneurial climate

II. Theoretical Framework

slide6
Local scientists job market conditions:

H1a: The size of local scientist job market positively affects the average NTBF performance in a region.

H1b: The average salary level of local scientists increases the average NTBF performance in a region.

Venture capital

H 2: The number of venture capital firms located in a region increases its average NTBF performance at a decreasing rate .

Cultural diversity

H 3: Cultural diversity increases the average NTBF performance in a region.

Academic research

H 4: Academic research positively affect the average NTBF performance in a region.

Industrial structure

H5a: The degree of industrial specialization in a region promotes its NTBFs’ performance.

H5b: Coagglomeration with buyer-industries promotes NTBFs’ performance.

H5c: Coagglomeration with supplier-industries promotes NTBFs’ performance.

Local entrepreneurial climate

H6: Entrepreneurial culture positively affects the average performance of NTBFs in a region.

III. Hypotheses

1 how to quantitatively define catch up regions
Define it at Metropolitan Statistical Area (MSA) level;

Create a 1995 high-tech index, with two equally weighted components:

Industrial size: share of national establishments in a specific high tech industry

Industrial density: number of industrial establishments per square mile

Choose the 95th percentile as the cutoff point to generate a stable list during the study period

1. How to quantitatively define catch-up regions?

IV. Methods

slide8

Fig.1: Leading and catch-up regions in the biopharmaceutical industry,1995

Total MSAs: 168

Leading: 8

Catch-up: 160

slide9

Fig.2: Leading and catch-up regions in the IT service industry, 1995

Total MSAs: 316

Leading: 15

Catch-up: 301

2 measures
Dependent variable: average NTBF performance in a MSA

Number of Initial Public Offering (IPO) and Merger &Acquisition (M&A) event

Independent variables

Local scientist job market conditions

Job market size, absolute salary ratio, relative salary ratio

Venture capital

Number of industry-related venture capital firms

Cultural diversity

Share of population that were foreign born

Academic research

Share of industry-related university R&D expenditure

Industrial structure

Location quotients of the industry, buyer-industries, and supplier-industries

Local entrepreneurial climate

New small firm birth rate

Control variables

industrial cluster size and density, time fixed effect

2. Measures
3 data sources
List of IPO and acquisition companies

Thomas Financial Co.’s SDC New Issue Database

Hoover’s IPO Center

Jay Ritter\'s 1975-2003 IPO dataset

Company information

Prospectus and other legal reports in the SEC filing database

Region-specific data

County Business Patterns, US Census Bureau

Regional Economic Account, US Bureau of Economic Analysis

Occupational Employment Statistics, US Bureau of Labor Statistics

VC data

Moneytree Venture Capital Profiles

3. Data sources
4 models
4. Models

Cross-sectional model

Dependent variable: total number of NTBF IPO and M&A events that occurred in industry i within a MSA m between 1996 and 2005.

Independent variables: values in 1995

Control variables: industrial cluster size and density in 1995

Two-period panel model

Dependent variable: number of NTBF IPO and M&A events that occurred in industry i within a MSA m between 1996-2000 or 2001-2005

Independent variables: values in 1995, 2000

Control variables: industrial cluster size and density, time fixed effect

5 zero inflated negative binomial zinb model specification
5. Zero-Inflated Negative Binomial (ZINB) Model Specification

Non-negative integer values, count data

Overdispersion

Sample variance bigger than mean

Significant alpha test statistics

Excess zeros

Biopharmaceutical : 59% zero count values

IT service industry: 34%

Significant Vuong test statistics

Heteroskedasticity

Significant Breusch-Pagan/Cook-Weisberg test statistics

Use robust standard errors

6 distance weighted measures
6. Distance-Weighted Measures

Use distance-weighted measures to capture the spill over effects from the adjacent regions

Step 1: calculate the physical distance between two MSAs

Step 2: weight the contribution of each MSA to the focal MSA by the inverse of their distance

Step 3: sum these weighted contributions across all MSAs to yield a distance-weighted value for the focal MSA.

Construct five distance-weighted measures

scientist job market size, venture capital firms, immigrants, academic R&D expenditure, and industrial establishments.

1 1 local scientist job market size
1.1: Local Scientist Job Market Size

V. Findings

  • Life scientist job market size has positive and significant impacts on biopharmaceutical NTBF performance in catch-up regions
    • One additional percentage point of life scientist job market share increases the expected count of IPO and M&A events by a factor of exp(0.493), or 1.64, holding other variables constant
  • Computer scientist job market size has negative but insignificant impact on IT service NTBF performance in catch-up regions
1 2 local scientist absolute salary ratio
1.2: Local Scientist Absolute Salary Ratio
  • Life scientist absolute salary ratio has positive and significant impact only in the full sample data.
  • Computer scientist absolute salary ratio has positive and significant impact in catch-up regions, but its impact is stronger and more significant in the full sample size.
1 3 local scientist relative salary ratio
1.3: Local Scientist Relative Salary Ratio
  • The impact of scientist relative salary ratio is negative but insignificant in catch-up regions for both industries;
  • The impact is negative and highly significant impact in IT leading regions.
2 venture capital
2. Venture Capital
  • The impact of biotech VC firms is not significant
  • The number of IT service VC firms increases NTBF performance at a decreasing rate in catch-up regions. The optimal number is around 7
  • The impact of IT service VC firms is negative and significant in the full sample regions
3 cultural diversity
3. Cultural Diversity
  • Cultural diversity has positive and significant impact on IT service NTBFs in catch-up regions
  • One explanation is that the measure of cultural diversity, which is the share of foreign born population, favors the IT service industry over the biopharmaceutical industry
4 academic research
4. Academic Research
  • Academic research is not significant in catch-up regions for both industries.
  • The impact of life science academic research became significant in the leading regions, only after capturing the spillover effect.
  • The impact of computer science academic research is positive and significant only in the full sample regions.
5 1 industrial specialization
5.1: Industrial Specialization
  • Industrial specialization has positive and highly significant impact only in the IT service industry.
5 2 co aggolomation with buyer industries
5.2: Co-aggolomation with Buyer-Industries
  • Coagglomeration with supplier-industries has little impact on NTBF performance in both industries in catch-up regions.
5 3 co aggolomation with supplier industries
5.3: Co-aggolomation with Supplier-Industries
  • Coagglomeration with Supplier-industries has little impact on NTBF performance in both industries in catch-up regions.
6 local entrepreneurial climate
6. Local Entrepreneurial Climate
  • There is strong evidence that the local entrepreneurial climate has positive and significant impact on NTBF performance in catch-up region for both industries
  • The impact of entrepreneurial climate is stronger in the leading regions than in catch-up regions.
7 industrial cluster size
7. Industrial Cluster Size
  • Industrial cluster size increases NTBF performance at a decreasing rate
    • The optimal biopharmaceutical cluster size is about 68, which is very close to the result (65 firms) by Folter (2006).
    • The optimal IT service cluster size is about 742.
8 industrial cluster density
8. Industrial Cluster Density
  • Industrial cluster density increases NTBF performance in general.
1 main findings
1. Main findings

Local entrepreneurial climate plays a significant and positive role on NTBF performance in both industries.

Scientist job market size and absolute salary ratio have positive impacts, but the former matters more in the biopharmaceutical industry, and the latter matters more in the IT service industry.

The effect of relative salary ratio is negative and insignificant.

Venture capital increases NTBF performance at a decreasing rate.

Cultural diversity has stronger impact in the IT service industry than in the biopharmaceutical industry.

Academic research has little impact in catch-up regions for both industries.

Industrial specialization is significant and positive only in the IT service industry.

Industrial cluster size increases NTBF performance at a decreasing rate.

Industrial cluster density generally increases NTBF performance.

VI. Conclusions

2 policy implications
Promote local entrepreneurial culture

Lower the barriers to entry for human capital and increase cultural diversity and regional tolerance

Increase the availability of venture capitals

Invest in academic research and strengthen the collaboration between academia and NTBFs.

2. Policy implications
3 limitations
First, the event of IPO and M&A is only a rude measure of the early stage success of NTBF.

Second, this study doesn’t decompose the impacts of industry-related entrepreneurial activities and that of general entrepreneurship.

Third, this study doesn’t explicitly examine the effect of existing economic development policies on local NTBF performance.

Finally, the temporal stability analysis in this study is based upon a two-period, ten-year-long time frame.

3. Limitations
4 future research directions
Develop alternative measures of early-stage firm performance to obtain more robust results.

Explore whether industry-specific entrepreneurial activities or just the overall entrepreneurial activities are the true determinant of NTBF performance.

Conduct similar analysis at the state level which would allow for the addition of policy instruments to the model.

Examine the temporal stability based upon a longer time span

4. Future Research Directions
slide32

Fig.3: IPO and M&A Events in the Biopharmaceutical Industry, 1995-2005

Statistics for catch-up regions

Sum: 221 (575)

Mean: 1.38

Max: 26

Min: 0

Std: 3.00

slide33

Fig.4: IPO and M&A Events in the IT Service Industry, 1996-2005

Statistics for catch-up regions

Sum: 2399 (6982)

Mean: 7.97

Max: 189

Min: 0

Std: 19.07

table 1 results of cross sectional and two period panel zinb models biopharmaceutical
Table 1: Results of Cross-Sectional and Two-Period Panel ZINB Models ( biopharmaceutical)

Note: Robust standard errors in parentheses; *** p<0.01, ** p<0.05, * p<0.1

table 1 continued
Table 1: Continued

Note: Robust standard errors in parentheses; *** p<0.01, ** p<0.05, * p<0.1

table 2 results of distance weighted zinb models biopharmaceutical
Table 2: Results of Distance-Weighted ZINB Models ( biopharmaceutical)

Note: Robust standard errors in parentheses; *** p<0.01, ** p<0.05, * p<0.1

table 2 continued
Table 2 (continued)

Note: Robust standard errors in parentheses; *** p<0.01, ** p<0.05, * p<0.1

table 3 results of cross sectional and two period panel zinb models it service industry
Table 3: Results of Cross-Sectional and Two-Period Panel ZINB Models ( IT service industry)

Note: Robust standard errors in parentheses; *** p<0.01, ** p<0.05, * p<0.1

table 3 continued
Table 3 (continued)

Note: Robust standard errors in parentheses; *** p<0.01, ** p<0.05, * p<0.1

table 4 results of distance weighted zinb models it service
Table 4: Results of Distance-Weighted ZINB Models ( IT service)

Note: Robust standard errors in parentheses; *** p<0.01, ** p<0.05, * p<0.1

table 4 continued
Table 4 (continued)

Note: Robust standard errors in parentheses; *** p<0.01, ** p<0.05, * p<0.1

ad