Eesley China Performance

S T A N F O R D U N I V E R S I T Y • Management Science & Engineering

Perf

orm

ance

in a

Dev

elop

ing

Econ

omy

1

Entrepreneurial Performance in a Developing Economy: Evidence from China

Chuck Eesley

March 8th, 2010


Perf

orm

ance

in a

Dev

elop

ing

Econ

omy

Institutional Environments and Entrepreneurship

Drivers of entrepreneurial entry and performance (different contexts)

Developed economy Entrepreneurs from Technology-Based Universities - with David Hsu

(Wharton), Ed Roberts (MIT) Bringing Entrepreneurial Ideas to Life Cutting Your Teeth - Prior entrepreneurial experience

Developing economy The Right Stuff

– Role of institutional environment in selection of high human capital entrepreneurs

Entrepreneurial Performance in a Developing Country: Evidence from China

What Drives an Innovation Strategy?– Role of Institutional Env./funding of S&T in search for ideas

2


Perf

orm

ance

in a

Dev

elop

ing

Econ

omy

Most Work on Entrepreneurship Done in Developed Economy

Individual Level Institutional Level

Overall Economic growth ++

Entrep. Entry ++ +

Entrep. Performance ++ ?

3

• Representative entrepreneurship • Self-employment (include lawyers and doctors)• Tech-based entrepreneurship


Perf

orm

ance

in a

Dev

elop

ing

Econ

omy

Who is an entrepreneur?

Entrepreneurs (self-employed) tend to have wealth (Blanchflower, Oswald, 1998;

Nanda, 2010), self employed parents (Sorensen, 2007; Dunn and Holtz-Eakin, 2000), low opportunity costs (Amit et al., 1995), more educated (Fairlie, Woodruff, 2007), in their 30s-40s (Levesque and Minniti, 2006), generalists (Lazear, 2004), tend to be brokers (Burt, Raider, 2002), low on uncertainty avoidance/higher on individualism (McGrath, MacMillan and Scheinberg, 1992), and achievement need (Johnson, 1990; Roberts, 1991)

Tech entrepreneurs are more highly educated (Roberts, 1991), come from VC-backed firms (Gompers et al., 2005) and often male, engineering/management background, and non-U.S. citizens (Saxenian, 1999; Hsu, Roberts & Eesley, 2007)

In developing world? Entrep. family, friends, value work, wealth (Djankov et al., 2006)

Human capital, overseas, academics less likely (Eesley, 2010) Interaction between individual and institutional environment characteristics

4


Perf

orm

ance

in a

Dev

elop

ing

Econ

omy

Who is a successful entrepreneur? (developed economy)

Self-efficacy (Baum, 2004) Education - Master’s degree (Roberts, 1991)

Prior entrepreneurial experience (Roberts and Eesley, 2010; Shane and Khurana, 2003)

Managerial achievements (Eisenhardt & Schoonhoven, 1990)

Work in an entrepreneurial prominent organization (Burton, Sorensen, & Beckman, 2002)

Cohesive social network (Shane, Cable, 2002)

Growth market (Eisenhardt and Schoonhoven, 1990)

Planning (Delmar, Shane, 2003)

Tech-entrepreneurs also: attract VC funding (Chemmanur, Krishnan, and

Nandy, 2008), larger founding teams, transfer tech. from parent organizations, product and marketing orientation (Roberts, 1991; 1992)

5


Perf

orm

ance

in a

Dev

elop

ing

Econ

omy

Factors that Drive Founding Do Not Drive Performance

Founding Performance

Parental entrepreneurship Managerial Experience, prior founding

Wealth Lower wealth

Low opportunity costs High opportunity costs

Male Either gender

VC-backed firm work experience Entrep. prominent organization

Broker network Cohesive network

Need for achievement Self-efficacy

Education Education

Low uncertainty avoidance, high individualism Product, marketing orientation

Generalists Growth market

Non-U.S. citizen Either

In 30s-40s Any age

Planning, VC6


Perf

orm

ance

in a

Dev

elop

ing

Econ

omy

7

Institutional Environment and Competitive Advantage

Successful entrepreneur, developing economy? Individual characteristics, context characteristics, and their

interaction drive performance – Team and industry (Eisenhardt and Schoonhoven, 1990)

As institutional environment changes, individual and context characteristics that drive performance change

Institutions – political, social, legal (formal and informal) constraints on indiv. and orgs. (Scott, 2001; North, 1990)

Transition – government planning and control free market institutions


Perf

orm

ance

in a

Dev

elop

ing

Econ

omy

Emerging Economies and Government Ties

Firms with government and managerial ties outperform (Peng and Luo, 2000) Picking winners, providing resources (nefarious) Privatization auctions that are not fully competitive (Schamis, 2002) Closer relationships with state-owned firms (Backman, 2001) Better access to credit (Khwaja and Mian, 2005; Leuz and Oberholzer-Gee,

2006; Li, et al, 2009) Government bailouts (Faccio, Masulis and McConnell, 2006)

Over time? Ties to sociopolitical elites increase the propensity to form cross-border

alliances yet when the regime changed, these ties became a liability (Siegel, 2007)

Elites in transitional economies have been able to translate their power into economic benefits? (Nee, 1996; Walder, 2002; Walder, 2003)

New firms? Technology firms?8


Perf

orm

ance

in a

Dev

elop

ing

Econ

omy

9

Elite Entrepreneurs in a Developing Economy

Push vs. pull entrepreneurshipEliminates subsistence entrepreneurshipFocus on form of entrepreneurship tends to result in economic

growth, not poverty alleviation

Desired homogeneity in skills, opportunity costs, more narrow, well-defined set, at risk for technology-based entrepreneurship

Comparing apples to apples


Perf

orm

ance

in a

Dev

elop

ing

Econ

omy

Model/Hypotheses

10

Relaxing constraints on entry/markets, allowing market institutions Providing information on opportunities Stepping away, other institutions play a role Incentives for entrepreneurial behavior

Central Planning/Control

Transition Free Markets

Phase I Phase II Phase III

Entry is difficult Other institution or gov. driven

Entrep. process/skill driven

Government ties Other institutional or govt. programs

Entrep. experience and innovation


Perf

orm

ance

in a

Dev

elop

ing

Econ

omy

11

Hypotheses:

H1: In the beginning, in institutional environments transitioning from an emphasis of government central planning and control, entrepreneurs in locations where government control was initially relaxed will create larger firms.

H2: In institutional environments characterized by an emphasis on government central planning and control, entrepreneurs with government ties will create larger firms compared to entrepreneurs without such ties.

H3: As institutional environments transition from government central planning and control to market-based incentives, entrepreneurs who can access programs by non-gov. institutions (such as science parks) to create entrepreneurial behavior will create larger firms.

H4: In institutional environments that have transitioned from government central planning and control to market-based institutions, greater competition will make entrepreneurs with exposure to the firm creation process and those who are innovating create larger firms.


Perf

orm

ance

in a

Dev

elop

ing

Econ

omy

Context

Alumni survey Tsinghua University26,700 mailed (correct addresses)3,000 surveys11% r.r.Growing tradition: Stanford GSB, Chicago, HBSDisadvantages: Biased towards tech., other response bias?

Advantages: Defined‘at risk’ set, first abroad, detailed work history and founding data, less biased by govt. concerns

Other surveys: business owners, self-employed and entrepreneurs merged together – Treiman and Walder, “Life Histories and Social Change in Contemporary China”

Self-employment (Chinese Health and Nutrition Survey, NBS HH Survey)

12


Perf

orm

ance

in a

Dev

elop

ing

Econ

omy

Tsinghua Univ.

13

• Established in Beijing in 1911

• 1952 reorganized Soviet style

• 1966-1976 Battlefield during Cultural Revolution

• 1978 restored departments in sciences, economics and management, and humanities

• 1984 – First Graduate school in China created at Tsinghua

• 1998 – Tsinghua Science Park established


Perf

orm

ance

in a

Dev

elop

ing

Econ

omy

14


Perf

orm

ance

in a

Dev

elop

ing

Econ

omy

79 pages of notes from interviews with 42 individuals

– 26 Tsinghua alumni entrepreneurs– 2 Tsinghua staff (TLO, Science Park)– 5 Chinese venture capitalists (VCs)– 2 Government officials– 3 Other Chinese entrepreneurs (non-Tsinghua)– 2 MIT Alumni (non-entrepreneurs)– 2 Tsinghua alumni (non-entrepreneurs).

• Interviews were in Beijing, Shanghai and Xi’an

15

Context - interviews


Perf

orm

ance

in a

Dev

elop

ing

Econ

omy

Categories Tsinghua CHNS NBS HH survey

NBS HH survey

Sample Urban Rural and Urban

Urban – self-employed

Urban – non. Entrep.

Male 0.89 0.53 0.56 0.50

Age 50.13 41.45 36.2 37.2

Married 0.88 0.98 0.83 0.84

Years of Education

17.1 9.1 9.2 9.4

Household Size

3.40 3.9 -- --

Self-employed 0.26(0.8% in

1999)

0.14 (4% in 1999) --

Experienced a layoff

0.13 -- 0.26 0.19

Father’s Educ.

4.11 -- 5.4 5.2

Mother’s Educ.

4.89 -- 6.0 5.9

Parent Self-Empl.

0.08 -- 0.06 0.05

Comm. Party 0.62 -- 0.05 0.18

Benchmarking Tsinghua Data


Perf

orm

ance

in a

Dev

elop

ing

Econ

omy

Why China?

"The storm center of the world has shifted . . . to China, whoever understands that mighty Empire . . . has a key to world politics for the next five hundred years.” --U.S. Secretary of State John Hay, 1899

Tech-based entrepreneurship in developing countries rarely appears in academic literature (Lu 1997, 2000; Puga and Trefler, 2005)

Vernon’s (1966) product-cycle model

19892004 China 29% vs. US 1% (State Statistics Bureau)19782004 # employed in private business up 300X

Policies and institutions changing rapidly (Cull & Xu, 2006; Nee, 1998; 1992; 1996; Peng & Heath, 1996; Steinfeld, 2007)


Perf

orm

ance

in a

Dev

elop

ing

Econ

omy

China’s S&T Policy Reform

Experimental (PRO centered)

78-85

Structural Reform of S&T

85-95

Deepening S&T reform

95-2005

Firm-centered innovation system

2005+

Open door policy (1978)

SE Zones created (1980)

Univ. reform (1985)

Bankruptcy law for SOEs (1986)

Stock Exchange (1990)

Join WTO (2001)

Private ownership (1999)

Promotion of VC/PE (1998)

CAS Knowledge Innov. Prog. (1998)

Deng Xiao Ping reform report (1975)

2006 Adoption of medium and Long Term S&T Strategic Plan

National Key Tech. R&D Program (1984)

National Natural Science Fndtn. (1986)

Innov. Fund for Tech. SMEs (1999)

(Thanks OECD Review, 2007)

>7 emp. permittedTorch Program (1988)

Tsinghua Science Park(1998)

Gradual, local and sectoral experimentation, partial reforms dual-track approach (Gregory, Tenev, & Wagle, 2000; Nee, 1996) Economic, not political liberalization

Promotions and tax revenue were tied to local economic development (“eating from separate kitchens”, or fenzao chifan)

TVEs (getihu) FDI, 1988 (saying qiye)

Privatization (zhuanzhi) or “restructuring of ownership” (suoyouzhi gaizao) 1995 – 1996 “Grasping the large, letting go of the small.”


Perf

orm

ance

in a

Dev

elop

ing

Econ

omy


Perf

orm

ance

in a

Dev

elop

ing

Econ

omy

Variable Respondents(N=2,667)

Non-respondents(N=299)

t-stat for equal means

Age 49.3 54.1 -4.216**Age (founders only) 38.4 37.4 0.602Bachelor’s Graduation Yr 1980.9 1977.4 3.777**Bach. Grad yr (founders only) 1991.6 1993.2 0.941Years of Education 17.2 17.0 2.381**Entrepreneur parents 0.09 0.12 -0.713EntrepreneurPrivatizedFirst start-up founded

0.290.10

2000.3

0.400.05

2001.1

-2.168**1.392-0.661

Tech only 0.28 0.29 0.757Business only 0.10 0.09 0.235Gender 0.88 0.90 0.901Family economic status 3.75 3.85 -1.871*High Salary 3.21 2.93 3.351**Avg. Tenure 6.94 8.01 -2.045*Overseas work exp. 0.26 0.26 -0.126Number of positions 2.39 2.26 -2.012*High governmentLow government

0.030.18

0.030.17

-0.2390.617

Last job academia 0.19 0.19 -0.051Ever job academia 0.32 0.27 2.323**Last job business 0.62 0.61 0.348Student Leader 0.61 0.57 0.874GPA Rank 2.28 2.58 -2.661**Bach. Grad Yr. 10th percentile 1954 1953 --Bach. Grad Yr. 25th percentile 1965 1961 --Bach. Grad Yr. 50th percentile 1986 1979 --Bach. Grad Yr. 75th percentile 1996 1993 --Bach. Grad Yr. 90th percentile 2001 2001 --

20

Comparison of Key Demographic Characteristics by Survey Wave

***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively.

Back


Perf

orm

ance

in a

Dev

elop

ing

Econ

omy

1982

1984

1986

1987

1988

1989

1990

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

0

5

10

15

20

25

30

35

First Foundings by Year

First Founding Year

Nu

mb

er o

f S

tart

-up

s


Perf

orm

ance

in a

Dev

elop

ing

Econ

omy

INDUSTRY NUMBER OF FIRMS %AEROSPACE 3 0.90ARCHITECTURE 13 3.88BIOTECH AND DRUGS 7 1.09CHEMICALS 8 2.39CONSUMER PRODUCTS 17 5.07ELECTRIC 12 3.58ELECTRONICS 69 20.60ENERGY 14 4.18FINANCE 10 2.99INTERNET 33 9.85LAW, ACCOUNTING 22 6.57MACHINERY 19 5.67MANAGEMENT 21 6.27MATERIALS 13 3.88MED DEVICES 4 1.19OTHER MFG 16 4.78PUBLISHING 11 3.28SOFTWARE 34 10.15TELECOM 9 2.69TOTAL 335 100

Industries


Perf

orm

ance

in a

Dev

elop

ing

Econ

omy

Descriptive results

Could create a figure or t-tests of means across time periods here

23


Perf

orm

ance

in a

Dev

elop

ing

Econ

omy

China is more state-led than a market economy still. The government controls many resources. For firms the quickest way to make a lot of money is through the government. The government has incentives for firms to put up good numbers (it’s good for the politicians’ careers).

Interview Quotes

In the 1990s … government was giving them less support. Many universities created “university –run” enterprises and were basically selling off the periphery of campus. Even high schools had school—run enterprises that were considered acceptable. There is a Tsinghua name to many of them. - GC – 2nd generation investor


Perf

orm

ance

in a

Dev

elop

ing

Econ

omy

25

In the US you don’t see the government, but in China the role of the government is seen from the very beginning. They could be a major funding source, help in penetrating the markets, and in protecting the competitive advantage. There is some over-emphasis of the role of the government however. - GY– VC investor

There is a big government legacy. In the early days the government connections and system were the currency. Power was the currency, but that is changing to a system whose currency is monetary. There was an opportunity to monetize that power. - GC - investor

Interview Quotes, cont.


Perf

orm

ance

in a

Dev

elop

ing

Econ

omy

Interviews

Historically, investors in China see fewer experienced, serial entrepreneurs

“Forced to rely more on work experience outside of an entrepreneurial context to judge the quality of entrepreneurs – GY – VC investor

26

Tsinghua alumni, entrepreneurship among alumni has a meta-effect where some learn the process and mentor each other so in the future there will be more and more entrepreneurs from Tsinghua as has already happened with MIT alumni. – RC – Wave Communications (his 6th start-up firm)


Perf

orm

ance

in a

Dev

elop

ing

Econ

omy

27

Analysis:

E[yi | x] = α + ρ’zi + ’xi + ’yi + ’pre xi + ’Xi + pre, mid, post + η + φ + i

Dep. Variables: yi represents our measures of firm

performance

xi is government ties (father in government, privatized, coastal province)

yi is government programs (specifically science parks)

zi is separate measures for exposure to entrepreneurship, prior founding experience, or innovation

Xi vector of control variables - Everjob_govt, Comm. Party, Govt. customer, Privatized, Entrep. parents, Entrep. index, rural, EECS, Overseas, SEZ, software, electronics, wealthy family, Master’s, PhD, number of cofounders

η and φ represent year and industry sector dummies >0


Perf

orm

ance

in a

Dev

elop

ing

Econ

omy

28

Results: firm performance

N=230; (R2 0.627) ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively. Everjob_govt, Comm. Party, Govt. customer, Privatized, Entrep. parents, Entrep. index, rural, EECS, Overseas, SEZ, software, electronics, wealthy family, Master’s, PhD, number of cofounders, industry and year fixed effects were included as controls but coefficients are not shown to save space.

VARIABLES Log(rev) Log(rev) Log(rev)Log(rev

) Log(rev)Coastal 1989-2000 2.637**

(1.049)Privatize 1989-2000 2.406**

(1.058)Gov. dad 1989-2002 0.868**

(0.403)Park x 2000-02 1.391*

(0.712)Park x 2004-07 -1.545**

(0.646)E-index x 2004-07 0.214**

(0.100) Serial x 2003-07 1.409**

(0.689) Innovation x 03-07 1.939**

-0.912 1989-2002 5.021**

(1.942)1989-2002 0.834

(0.925)2000-02 -3.396**

(1.388)Log(firm age) 0.360 0.448** 0.417** 0.450** 0.007 0.085

(0.243) (0.190) (0.183) (0.180) (0.458) -0.431


Perf

orm

ance

in a

Dev

elop

ing

Econ

omy

Robustness Checks / Limitations

Deeper analysis of how gov. father and other factors helped (doesn’t appear to be through gov. purchases or more government funding) Information on opportunities?

Broad set of controlsNext Steps

– Additional analysis – government role in certain industries remains?

LimitationsRepresentativeness, response rates, self-reportingReverse causality, lobbying for reformsEffect of higher competitionMay be some shift back towards government connections with more recent govt. focus on standards

29


Perf

orm

ance

in a

Dev

elop

ing

Econ

omy

Factors driving which individuals start larger, more successful firms is not constant over time.

Shifts in the directions consistent with the model that the drivers of performance follow the institutional changes: government, science parks, market-based free competition

Conclusions

H1: In the beginning, … entrepreneurs in locations where government control was initially relaxed will create larger firms. (coastal regions, privatized)

H2: …, entrepreneurs with government ties will create larger firms compared to entrepreneurs without such ties. (father in government)

H3: … entrepreneurs who can access programs by non-gov. institutions (such as science parks) to create entrepreneurial behavior will create larger firms. (science parks)

H4: In institutional environments that have transitioned … to market-based institutions, greater competition will make entrepreneurs with exposure to the firm creation process and those who are innovating create larger firms. (entrep. index, serial, innovation)


Perf

orm

ance

in a

Dev

elop

ing

Econ

omy

Institutions guide macro-level economic performanceInstitutions – may guide (or constrain) the individual characteristics and firm creation strategies leading to performanceIndividual and institutional characteristics interact to generate performance, no single blueprint

Comparable process at University LevelMIT Case Study – cite Entrepreneurial ImpactFrom Administration/TLO central control to relaxing constraintsNext, University programs – b-plan competitions, etc.Finally, faculty/student/alumni entrepreneurial behavior grows outside of programsMay explain mixed-results in university entrepreneurship literature - Rothaermel review

Entrepreneurship Literature Drive to find “universal” drivers of performance Context-specific Opportunity recognition vs. resource mobilization

Implications


Perf

orm

ance

in a

Dev

elop

ing

Econ

omy

Boundary Conditions

China had gradual economic reforms while political system remained largely unchanged.

Contrast with Russia where economic liberalization was sudden and accompanied political changes as well

Careful study of Western models and intention to move from government-centered innovation to firm-centered innovation

Regional experimentation, incentives for local economic development

Focus on tertiary education reform

Strong historical culture of entrepreneurship (esp. in certain locations)

Focus on firm performance, not social welfare or aggregate economic impact

32


Perf

orm

ance

in a

Dev

elop

ing

Econ

omy

Thank you!

Chuck EesleyStanford University

Management Science & Engineering (MS&E)[email protected]

mailto:[email protected]


Perf

orm

ance

in a

Dev

elop

ing

Econ

omy

34

Results: Conditioning on “inefficient matching”

Use a subsample where we are more confident of seeing inefficient matching. Removing firms where the team and idea both came from the same source (less evidence of search for optimal pairings)

Coefficients remain significant and of similar magnitude reassuring us that endogeneity is not driving the results.

To further test whether our contracting variables might be serving as a proxy for higher human capital founders who would only become involved if there is a chance for a very high outcome …

Probit - in the top 5% of the revenue distribution or the valuation at exit distribution (in the case of IPO or acquisition). The results held, and likelihood of being in the extreme right tails of the distribution increased only when both capabilities and contracting expertise were present.

Unobservable heterogeneity, mainly in the form of individual ability is a concern, but mitigated by the relative homogeneity.


Perf

orm

ance

in a

Dev

elop

ing

Econ

omy

35

Results

Ln(Alliances) Ln(Employees) Ln(Revenues) Pr(Public) Pr(Acquired)Capabilities 0.023 0.328 0.256 -0.266 -0.188

(0.094) (0.322) (0.284) (0.346) (0.242)Capab.*contract 0.203** 0.864*** 0.475* 0.861*** 0.433*

(0.101) (0.355) (0.323) (0.328) (0.246)Contracting -0.068 0.061 0.029 -0.004 0.273**

(0.049) (0.164) (0.153) (0.146) (0.119)ControlsWork Idea 0.242*** 0.452 0.801*** 0.607** 0.192

(0.087) (0.302) (0.274) (0.283) (0.212)

Social Idea 0.119 0.142 0.595* 0.418 0.444*

(0.104) (0.362) (0.324) (0.316) (0.252)

Military/Gov. Idea 0.325* -0.044 0.272 1.534*** 0.210

(0.168) (0.623) (0.511) (0.538) (0.416)

Work Team 0.048 0.183 0.286 -0.572** -0.114

(0.087) (0.297) (0.266) (0.292) (0.207)

Research Team 0.086 -0.333 -0.013 -0.630** 0.019

(0.093) (0.322) (0.289) (0.307) (0.222)

Social Team 0.203*** -0.089 -0.101 -0.417* -0.007

(0.075) (0.262) (0.236) (0.252) (0.182)

Family Team 0.022 -0.185 -0.147 -1.397** -0.531*

(0.115) (0.394) (0.348) (0.619) (0.313)

N=500; Controls: idea/team source, education, external funding, age, firm age, num. cofounders, industry, year. ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively. P-values represent one-tailed tests. All regressions include industry sector dummies, though the coefficients are not shown.


Perf

orm

ance

in a

Dev

elop

ing

Econ

omy

Boundary Conditions

New Business Line & New Entrepreneurial Firms

Stable Institutional and Industry Environment

Frictions in Markets for Technology

Industry Life Cycle– Mature contracting– Fluid/Early-stage capabilities– High velocity – firms selected out more quickly

36