S T A N F O R D U N I V E R S I T Y • Management Science & Engineering
CUTT
ING
YO
UR
TEET
H
1
Cutting Your Teeth: Learning from Entrepreneurial Experience
Chuck Eesley (Stanford), Edward B. Roberts (MIT)
Organization Science Winter ConferenceFeb. 3-7th, 2010
S T A N F O R D U N I V E R S I T Y • Management Science & Engineering
CUTT
ING
YO
UR
TEET
H Motivation
When do new ventures benefit from the prior entrepreneurial experience of their founders?
Under what conditions does organizational learning get transferred by individuals to new organizations (what type of learning in the case of entrepreneurial experience)?
2
S T A N F O R D U N I V E R S I T Y • Management Science & Engineering
CUTT
ING
YO
UR
TEET
H
Gruber, 1994; Rapping, 1965; Thornton & Thompson, 2001
Strategic Contexts:acquisitions (Finkelstein & Haleblian, 2002; Haleblian & Finkelstein, 1999; Hayward, 2002;
Vermeulen & Barkema, 2001), alliances (Anand, B. and T. Khanna, 2000; Hoang &
Rothaermel, 2005) and internationalization (Bingham, Eisenhardt, & Davis, 2009), innovation (Katila & Ahuja, 2002)
Type of ExperienceSuccesses/failures (Sitkin, 1992), variety (Schilling, Vidal,
Ployhart, & Marangoni, 2003), complexity, voluntary or not (Haunschild & Sullivan, 2002; Haunschild & Rhee, 2004)
Transfer of learning across organizations (Ingram & Baum,
1997; Kim & Miner, 2007; Miner & Haunschild, 1995) Vicarious (Haunschild & Miner, 1997; Huber, 1991)
Organizational Learning
S T A N F O R D U N I V E R S I T Y • Management Science & Engineering
CUTT
ING
YO
UR
TEET
H
Organization or industry-level phenomenon (Cyert & March, 1963; Baum & Ingram, 1998) Individual level?
Simon (1991):1) by ‘ingesting’ new members who have knowledge not previously in the organization, or 2) by its members learning
Huckman and Pisano (2006) - experience within particular organization
Hire employees / management to access internally (rather than externally) the accumulated learning – strategy/OT (Beckman & Burton, 2008; Ahuja & Katila, 2001)
Hypothesis 1: The benefit from learning transferred by an individual will be higher with greater levels of prior experience.
Hypothesis 2: The benefit from learning transferred by an individual will be higher with prior successful experience.
Motivation: Types of learning experiences
S T A N F O R D U N I V E R S I T Y • Management Science & Engineering
CUTT
ING
YO
UR
TEET
H
Transfer effects - loss of performance if skill is wrongly applied in a different context (Haleblian & Finkelstein, 1999; March, 1991)
Hypothesis 3a: The benefit from learning transferred by an individual will be higher with prior experience in the same industry.
Industry evolution - automobiles (Abernathy, 1978) typesetters (Tripsas, 1997)
Major disruptions - learning in the previous environment no longer relevantFind the right causal relationships and models to fit the changed environment (Kaplan & Tripsas, 2008)
Hypothesis 3b: The benefit from learning transferred by an individual with prior experience will be lower after a significant shift in the industry.
Hypothesis 4: The benefit from learning transferred by an individual will be higher with more recent prior experience.
5
S T A N F O R D U N I V E R S I T Y • Management Science & Engineering
CUTT
ING
YO
UR
TEET
H
Existing work mainly argues that processes and routines are learned from operating experience (Nelson & Winter, 1982; Winter, 2000)
Bingham, Eisenhardt, and Davis (2009) - rather than routines, increasingly sophisticated and refined portfolios of heuristics to guide actions
Content knowledge as important or more so than process knowledge
Hypothesis 5: The benefit from learning transferred by an individual will come from content learning about the industry gained from prior experience rather than about process learning.
6
Shedding light on what the individual is learning
S T A N F O R D U N I V E R S I T Y • Management Science & Engineering
CUTT
ING
YO
UR
TEET
H MIT Data
Long time horizon in the cross section (1930s-2003)Note: not MIT-originated technology
Alumni: 105,000 surveyed; 42,930 records in 2001– Date of birth, country of citizenship, gender, major at
MIT, highest attained degree– 7,798 indicated founding at least one company
Survey of self-identified MIT alumni entrepreneurs in 2003
– 2,067 respondents (r.r. 27%) – More detailed info; new venture founding history
(multi-founder r.r. of 30.4%, 1.79 vs. 2.13 reported)
80% of the company names D&B database (obtain a credit rating) no bias towards larger firms (Aldrich, Kalleberg, Marsden and Cassell, 1989)
VEIC SIC codes (Dushnitsky & Lenox, 2005)
S T A N F O R D U N I V E R S I T Y • Management Science & Engineering
CUTT
ING
YO
UR
TEET
H Univariate t-tests of means
Panel A No prior founding experience At least 1 prior founding experience
Revenues 13.957 14.219*
Exits 0.205 0.240**
Panel BPrior founding exper. in a
different industryPrior founding exper. in the same
industry
Revenues 14.042 14.854**
Exits 0.060 0.173***
Panel C+Before dotcom boom, no prior
founding exper. Before dotcom boom, has prior
founding exper.
Revenues 14.614 15.227*
Exits 0.275 0.391*
Panel D+After dotcom boom, no prior
founding experience After dotcom boom, has prior
founding experience
Revenues 12.810 13.375
Exits 0.182 0.061**8
S T A N F O R D U N I V E R S I T Y • Management Science & Engineering
CUTT
ING
YO
UR
TEET
H
MethodsOLSY = F(α + 1’θi + 2’ *(exper.) + 3’ *(exper.)*(ind. disruption) + ’Xi
+ τt+ ηj + εi)
Cox Hazard Rate Model (robust to logit)
Prob (Y= 1) = F(α + 1’ *(experience) + ’Xi + τt+ ηj + εi)
Dependent variable: Revenues, exit (acquisition, IPO)
# prior experiences, # exits (acquired, IPO), same/different industry
Xi = Set of controls for firm age, external funding, num. cofounders, functional diversity, initial capital
Include (τ + η) year, industry sector fixed effectsBefore and after significant industry disruption
9Proportional Hazards Test
S T A N F O R D U N I V E R S I T Y • Management Science & Engineering
CUTT
ING
YO
UR
TEET
H Revenues
N=1106, 911. Controls for firm age, num. cofounders, functional diversity, industry, year, initial capital, VC funding are controlled. (more)
Pr(Exited)(4-1)
Ln(Rev)(4-2)
Pr(Exited)(4-3)
Ln(Rev)(4-4)
Pr(Exited)(4-5)
Ln(Rev)(4-6)
Pr(Exited)(4-7)
Ln(Rev)(4-8)
Exper. Founder 1.568*** 0.453**
(0.219) (0.181)Num. Prior Exper. 1.167** 0.272***
(0.083) (0.064)
Prior exits 1.382*** 0.401***
(0.116) (0.137)
Same SIC 1.166 1.167***
(0.356) (0.429)
S T A N F O R D U N I V E R S I T Y • Management Science & Engineering
CUTT
ING
YO
UR
TEET
H Large Industry Disruption
VARIABLESPr(Exited
)Ln(Revenue
s)Ln(Revenue
s)Ln(Revenue
s)Ln(Revenue
s)Ln(Revenue
s) Pr(Exited)
(5-1) (5-2)Before(5-3)
Before(5-4)
After(5-5)
After(5-6) (5-7)
Exper. founder 0.407 1.333** 2.139*** 0.440(0.287) (0.596) (0.697) (0.527)
Post-1997*Experienced founder -1.214** -0.302
(0.522) (0.840)Prior exits 1.173** 0.545
(0.473) (0.438)Lag 25-50th quartile 2.453**
(1.085)Lag 50-75th quartile 1.633
(1.407)Lag 75th+ quartile -6.608***
(2.549)Age founded -0.021* -0.029 -0.0537* -0.0663* 0.024 0.022 -0.015
Software firms only; N=205Controls: age founded, num. cofounders, Bachelor’s, Master’s, initial capital, firm age
Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
S T A N F O R D U N I V E R S I T Y • Management Science & Engineering
CUTT
ING
YO
UR
TEET
H
Content vs. process
1063 firms, 234 events and 16,068 years at risk. All models include controls for Master’s and Doctorate degrees, the number of cofounders, log(firm age), functional diversity and constant terms, but the coefficients are not shown to save space. Model (6-3) excludes firms that were VC funded. Models (6-4), (6-5) and (6-6) exclude firms that were funded by angel investors. The results are robust to leaving these firms in as well.
Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
VARIABLES Log(Rev) Pr(Exited) Log(Rev) Log(Rev) Log(Rev) Log(Rev)
Prior exper. 0.368*** 1.263** 0.388*** 0.356***
(0.072) (0.131) (0.087) (0.074)
Prior exper.*External funding -0.416*** 1.098
(0.149 (0.183)
Prior exper.*angel -0.553**
(0.273)
Prior exper.*VC -0.387**
(0.176)
Same SIC 1.124***
(0.392)
Same SIC*VC -2.399***
(0.904)
Different SIC 0.265***
(0.066)
Different SIC*VC -0.489
(0.592)
VC 1.016** 0.496 0.462
S T A N F O R D U N I V E R S I T Y • Management Science & Engineering
CUTT
ING
YO
UR
TEET
HAlternative stories/Robustness checks
1. Unique to particular measures of “performance”
2. More talented or persistent individuals select into serial entrepreneurship (individual fixed effects)
3. Learning about the start-up process (evidence on industries)
4. Increased social network size (evidence on location)- most communication is with those in closer physical proximity
5. Wealth
S T A N F O R D U N I V E R S I T Y • Management Science & Engineering
CUTT
ING
YO
UR
TEET
H Conclusion and Implications
14
Hypothesis:Supported
?
H1The benefit from learning transferred by an individual will be higher with greater levels of prior experience. Y
H2 “… higher with prior successful experience Y
H3a
“… higher with prior experience in the same industry Y
H3b
“… lower after a significant shift in the industry Y
H4 “… higher with more recent prior experience Y
H5“… will come from content learning about the industry gained from prior experience rather than about process learning. Y
Organizational Learning and Entrepreneurship Literatures:
Individual Level• Learning by ingesting new members
Strategy• Micro-foundations of competitive advantage – content vs.
process (Haliblian & Finkelstein, 2002) diversification• Active view on identification of valuable resources (routines)• Level playing field after disruptions• Challenge of sectors with more serial entrepreneurs
Institutions• Fostering this type of experience (exit events, non-competes)
S T A N F O R D U N I V E R S I T Y • Management Science & Engineering
CUTT
ING
YO
UR
TEET
H Relationship to Broader Research Stream
Drivers of entrepreneurial entry and performance (different contexts)
Developed economy Entrepreneurs from Technology-Based Universities - with David Hsu
(Wharton), Ed Roberts (MIT) Bringing Ideas to Life – Conditions when types of assets performance
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
15
S T A N F O R D U N I V E R S I T Y • Management Science & Engineering
CUTT
ING
YO
UR
TEET
H
Thank you!
Chuck EesleyStanford University
Management Science & Engineering (MS&E)[email protected]
S T A N F O R D U N I V E R S I T Y • Management Science & Engineering
CUTT
ING
YO
UR
TEET
H Index of Backup Slides
Panel DataIndividual Fixed Effects
Learning about start-up process
Descriptive statisticsResponse Bias
S T A N F O R D U N I V E R S I T Y • Management Science & Engineering
CUTT
ING
YO
UR
TEET
HAlternative stories/Robustness checks
1. Unique to particular measures of “performance”
2. More talented or persistent individuals select into cross-functional roles or attempt 2nd start-ups (individual fixed effects)
3. Learning about the start-up process (evidence on industries)
More difficult to rule out
4. Increased social network size (evidence on location, could be mechanism)
5. Wealth
S T A N F O R D U N I V E R S I T Y • Management Science & Engineering
CUTT
ING
YO
UR
TEET
H Probability of Acquisition
Dep. Variable = Acquisition year(subjects start being at risk at year of founding)Note: reported coefficients are hazard ratios
Independent variables
Model 7-1 Model 7-2 Model 7-3 Model 7-4
Age at founding
# of start-ups foundedNumber of Cofounders
Prior acquisitions
Prior IPOs
# Same State
# Different State
Same Industry
Different Industry
0.989(0.034)2.224**(1.444)1.551
(0.492)
--
--
--
--
--
--
0.955**(0.021)
--
1.563(0.527)
2.011***(0.370)1.777
(0.759)--
--
--
--
0.969(0.020)
--
1.489(0.470)
--
--
1.255**(0.171)1.333**(0.234)
--
--
0.965(0.029)
--
1.578(0.928)
--
--
--
--
37.621**(56.90)3.675
(3.015)
Firm age, indiv. degree, Industry, year, initial capital, VC funding are controlled.
S T A N F O R D U N I V E R S I T Y • Management Science & Engineering
CUTT
ING
YO
UR
TEET
H Panel DataPR(ACQUIRED) PR(IPO) LN(EMPL) LN(SURVIVAL)
Num. of start-ups founded 0.040 (0.051) 0.002 (0.069) 0.066 (0.057) -0.028* (0.016)Num. prior acquired
0.396*** (0.087) 0.084 (0.116) 0.160 (0.103) 0.058 (0.024)
Num. same 2 digit SIC -0.239* (0.125) -0.014 (0.161)
0.442*** (0.143) 0.014 (0.034)
Age at founding year
-0.012*
** (0.004) 0.001 (0.005)
-0.012**
* (0.004) 0.006 (0.001)Gender (1=male)
0.404** (0.202) 0.372 (0.289)
0.582*** (0.153) 0.059 (0.052)
Masters -0.016 (0.076) 0.170* (0.103)0.305**
* (0.086) 0.040 (0.028)Doctorate -0.192* (0.102) 0.117 (0.130) 0.181 (0.121) 0.111 (0.036)
ln(emp)0.055*
** (0.019)0.188**
* (0.025)
ln(firm age)0.173*
** (0.057)0.358**
* (0.097)0.532**
* (0.074)
MA0.330*
** (0.081)0.260**
* (0.104) 0.214** (0.098) -0.021 (0.030)
CA0.389*
** (0.092)0.440**
* (0.123) -0.030 (0.102) 0.010 (0.033)
Constant -1.422 (1.347)
-2.543**
* (0.994)
-3.290**
* (0.626)1.412**
* (0.198)Year F.E. YES YES YES YESSIC F.E. YES YES YES YESIndividual F.E. NO NO NO NOR-squared 0.160 0.228 0.150 0.622Num. of obs. 1997 1760 2092 2217
***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively. Robust standard errors in parentheses.
S T A N F O R D U N I V E R S I T Y • Management Science & Engineering
CUTT
ING
YO
UR
TEET
HAlternative stories/Robustness checks
1. Unique to particular measures of “performance”
2. More talented or persistent individuals select into cross-functional roles or attempt 2nd start-ups (individual fixed effects)
3. Learning about the start-up process (evidence on industries)
More difficult to rule out
4. Increased social network size (evidence on location , could be mechanism)
5. Wealth
S T A N F O R D U N I V E R S I T Y • Management Science & Engineering
CUTT
ING
YO
UR
TEET
HControls for individual characteristics
PR(ACQUIRED) PR(IPO)
LN(EMPLOYEES) LN(SURVIVAL)
Num. of start-ups founded
2.326*** (0.181) -0.099 (0.074) 0.029 (0.129) 0.161*** (0.043)
Num. prior acquired
-5.105**
* (0.221) 0.331*** (0.114) 0.078 (0.186) -0.119** (0.060)Num. same 2 digit SIC -0.298 (0.248) 0.090 (0.154) -0.034 (0.208) 0.010 (0.064)Age at founding year
-0.103**
* (0.010) 0.000 (0.005) -0.016 (0.011) 0.013 (0.013)
ln(emp)-
0.099** (0.045) 0.158*** (0.025)ln(firm age) 0.359** (0.157) 0.394*** (0.093) 0.322** (0.145) Year F.E. YES YES YES YESSIC F.E. YES YES YES YESIndividual F.E. YES YES YES YESR-squared 0.750 0.206 0.750 0.884Num. of obs. 463 1771 2135 2231
***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively. Robust standard errors in parentheses.
S T A N F O R D U N I V E R S I T Y • Management Science & Engineering
CUTT
ING
YO
UR
TEET
HAlternative stories/Robustness checks
1. Unique to particular measures of “performance”
2. More talented or persistent individuals select into cross-functional roles or attempt 2nd start-ups (individual fixed effects)
3. Learning about the start-up process (evidence on industries)
More difficult to rule out
4. Increased social network size (evidence on location , could be mechanism)
5. Wealth (no strong effects of prior IPO)
S T A N F O R D U N I V E R S I T Y • Management Science & Engineering
CUTT
ING
YO
UR
TEET
H
Industry Context / Network
Independent variables
Model 6-2 Model 6-3 Model 6-4Revenues
(N=964) (N=648) (N=964)
Founder char.Age at founding -0.013 -0.019 -0.012 (0.009) (0.012) (0.009)Bachelor’s deg. 0.298 0.586+ 0.346 (0.256 (0.335) (0.255)Master’s degree 0.402 0.508 0.434+ (0.255) (0.334) (0.254)# Same State 0.238* (0.096)# Different State 0.125 (0.104)Same 2 digit SIC 1.675** (0.614)Different 2 digit SIC 0.153 (0.623)Prior acquisitions 0.445** (0.189)Prior IPOs 0.408 (0.341)R-squared 0.291 0.362 0.346
Charles Eesley – Cutting Your Teeth
Firm age, functional diversity, Industry, year, initial capital, VC funding are controlled.
Model 7-3
Acq.0.969
(0.020)--
1.489(0.470)
1.255**(0.171)1.333**(0.234)
--
--
S T A N F O R D U N I V E R S I T Y • Management Science & Engineering
CUTT
ING
YO
UR
TEET
H Alternative stories/Robustness checks
1. Unique to particular measures of “performance”
2. More talented or persistent individuals select into cross-functional roles or attempt 2nd start-ups (individual fixed effects)
3. Learning about the start-up process (evidence on industries)
More difficult to rule out
4. Increased social network size (evidence on location , could be mechanism)
5. Wealth
S T A N F O R D U N I V E R S I T Y • Management Science & Engineering
CUTT
ING
YO
UR
TEET
H Descriptive StatisticsVariable Obs. Mean Std. Dev. Min Max
Log revenues 1264 14.05 3.08 0.03 21.66Acquired 1840 0.19 0.39 0 1IPO 1790 0.11 0.32 0 1Lag between 1502 12.11 9.41 0 50Number of Firms 2058 1.61 1.30 1 11Prior Acquisitions 2067 0.13 0.42 0 3Prior IPOs 2067 0.04 0.23 0 3Prior Same SIC 1473 0.02 0.14 0 2Prior Different SIC 1473 0.02 0.18 0 3Prior Foundings in the Same State
2067 0.38 0.90 0 8
Prior Foundings in a Different State
2067 0.23 0.79 0 7
Age Founded 1807 39.65 10.59 18 83Bachelor's degree 2000 0.43 0.49 0 1Master's Degree 2000 0.41 0.49 0 1Operating Years 1837 14.34 11.30 0 74Industry 1600 9.77 4.34 1 16Number of Cofounders
2056 1.05 1.22 0 4
VC funded 1691 0.13 0.34 0 1Log initial capital 1264 11.91 2.72 0.28 21.02Functional Diversity of Team
1964 1.23 0.48 1 3
Charles Eesley
S T A N F O R D U N I V E R S I T Y • Management Science & Engineering
CUTT
ING
YO
UR
TEET
H
Initial Evidence
Panel A – Likelihood of Exit Events and Revenues (in 2001 dollars)
Firm Rank1st firms(N=556)
2nd firms(N=182)
3rd firms(N=84)
4th firms(N=21)
5th firms and higher
(N=36)Median Revenues (‘000s) 836 1,784 924 1,181 7,274Standard Dev. (‘000s) 153,000 117,000 130,000 10,800 21,200
Charles Eesley – Cutting Your Teeth
Revenues adjusted to constant 2006 dollars
S T A N F O R D U N I V E R S I T Y • Management Science & Engineering
CUTT
ING
YO
UR
TEET
H
Characteristics of Non-respondents
Variable Responded to 2001 survey(N=43,668)
Did not respond to 2001 survey (N=62,260)
t-stat for equal means
Male 0.83 0.86 10.11Engineering major 0.48 0.47 -4.49Management major 0.16 0.15 -5.75Science major 0.23 0.23 0.37Social sciences major 0.05 0.06 4.07Architecture major 0.06 0.08 11.82Non-US citizen 0.81 0.82 3.77North American (not US) citizen 0.13 0.11 -4.14Latin American citizen 0.13 0.12 -1.44Asian citizen 0.33 0.34 1.45European citizen 0.30 0.26 -5.08Middle Eastern citizen 0.05 0.08 6.32African citizen 0.03 0.05 6.25
Variable Responded to 2003 survey(N=2,111)
Did not respond to 2003 survey(N=6,131)
t-stat for equal means
Male 0.92 0.92 0.12Engineering major 0.52 0.47 -3.63Management major 0.17 0.21 4.17Science major 0.17 0.18 1.09Social sciences major 0.06 0.05 1.18Architecture major 0.09 0.09 1.06Non-US citizen 0.82 0.81 -1.36North American (not US) citizen 0.17 0.14 -1.34Latin American citizen 0.19 0.19 0.13Asian citizen 0.22 0.24 0.73European citizen 0.31 0.32 0.38Middle Eastern citizen 0.08 0.07 -0.59African citizen 0.04 0.04 0.17
Charles Eesley – Cutting Your Teeth
S T A N F O R D U N I V E R S I T Y • Management Science & Engineering
CUTT
ING
YO
UR
TEET
H
Thank you!
Chuck EesleyStanford University
Management Science & Engineering (MS&E)[email protected]