Download Assessing the impact on Entrepreneurial Outcomes

PUBLICLY FUNDED BUSINESS ADVISORY
SERVICES AND ENTREPRENEURIAL
OUTCOMES
DOUGLAS CUMMING
AND
EILEEN FISCHER
YORK UNIVERSITY
SCHULICH SCHOOL OF BUSINESS
MARCH 2011
Motivation
 Massive government spending to promote
entrepreneurship and innovation around the world

View from OECD and World Bank, among others:
entrepreneurship and innovation will be the key driver to
growth and prosperity in the 21st century
 Growing role of publicly funded business advisory
hubs (e.g., MARs in Toronto, ISCM in Markham)


What is the value added?
How to most effectively deliver services?
Additional Motivation
Prior Research on Public Funded Advisory Hubs
 No direct evidence of whether advisory services can
effectively enhance entrepreneurial outcomes when
they are deliberately targeted to innovation and
growth oriented firms.
 Scant evidence on impact of the intensity or volume
of advisory services
 Scant evidence considering endogenity and selection
econometric issues
In this paper....
 Unique dataset from Investment Network


Affiliated with Innovation Synergy Center, Markham (ISCM)
Thanks to Catarina von Maydell (detailed records)
 Examine the impact of mentor hours on entrepreneurial
outcomes




Sales
Patents
Financing
Alliances
 Control for endogeneity and selection effects
Investment Network
 ISCM started in 2003
 Investment Network in 2006-Q4, subprogram of ISCM
 Focus on companies that:





Generating revenue or will generate revenue within 12 months
Have the capacity to generate a minimum of $2M in revenue within
3 to 4 years
Have a sustainable competitive/technical advantage
Have a current company valuation of less than $2M
Will be looking for up to $500,000 in financing within 24 months
 Appropriate program to assess impact of advisory
services on entrepreneurial outcomes among firms with
growth and innovation intentions and potential
Hypotheses
 The amount of advising a growth-oriented firm
receives is positively and significantly related to
entrepreneurial outcomes subsequently attained by
the firm




Sales
Patents
Financing
Alliances
 Controlling for issues of causality and sample
selection
Data
 228 firms in contact with Investment Network
 101 entered the Investment Network Program, 2006-Q4 to 2009-Q2
 Types of Variables (over 100 variables in dataset)
 (1) dependent variables


(2) factors that influence whether or not the firm is part of the Investment
Network


Industry, incorporation date, business acumen, coachability, etc
(5) top management team characteristics


Hours spent, number of mentors, number of companies advised / mentor
(4) entrepreneurial firm characteristics


Referral sources, market conditions
(3) value added provided by advisors


sales, patents, financing, alliances
Age, race, experience, etc.
(6) market conditions

Public market returns over investment horizon, year effects, etc.
Some Costs and Benefits
 As at June 2009 early-stage entrepreneurial firms
had raised $6,545,000 in financing
 The program costs were totaled at $662,360
 Ratio of financing raised per dollar of cost is $0.10
 From a public policy perspective, therefore, the
program is highly efficient and cost effective
Table 2. Comparison Tests
This table presents comparison of means and medians tests for high (>20) versus (<20) number of hours that the advisor(s) spent with the
firm. Percentage changes for sales are for the subset of firms (48 in total) that had sales in 2008. Variables are defined in Table 1. P-value
presented for the median test. *, **, *** Significant at the 10%, 5% and 1% levels, respectively.
Advisors >20 hours
Advisors <20 hours
Variable
Number of
Observations
Mean
Median
Number of
Observations
Mean
Median
Comparis
on of
Proportio
ns
Comparis Compar
on of
ison of
Means
Medians
104.282*
**
Sales 08/09
28
106.674
72.321
20
360.915
35.417
Financing
53
0.755
1
48
0.583
1
1.834*
Patents
53
0.887
1
48
0.542
1
3.870***
Strategic Alliances
53
0.472
0
48
0.375
0
0.981
P<=0.00
0***
Econometric Tests: Tables 4 and 5
 Step 1: Is the firm part of the Investment Network
Program? (Logit)
 Step 2: How many mentor hours does the firm
receive accounting for step 1? (Heckman Selection)
 Step 3: What is the impact of advisor hours on
outcomes, controlling for steps 1 and/or 2?
(Instrumental Variables)


(Sales, Patents, Financing, Alliances)
Use log (hours) to account for diminishing effect
Steps 1 and 2: Findings
 More help needed in bad markets
 A 1-standard deviation increase in quarterly stock returns
lowers the probability that a firm will join the Investment
Network by 10%.
 An decrease in market conditions by one standard deviation
increases the total number of hours by 10
 Referral source matters
 If the firm is referred to the network by the Network’s
coordinators or a governmental organization then the
probability that the firm becomes part of the Investment
Network by 56% and 35%, respectively.
Steps 1 and 2: Findings (Continued)
 Entrepreneur characteristics matter
 Entrepreneurs of Middle East origin receive on average 15.5
more hours of advice
 Entrepreneurs with a Masters degree receive on average 20
more hours of advice
 Females receive on average 16 fewer hours advice
 Harder to measure personal traits also matter
 More advice is provided to firms with higher rankings in terms
of coachability and business acumen
 Less advice is provided to firms with more people that are part
of the top management team
Key Findings from Regressions: Sales 08/09
 Additional mentor hours  greater sales
 Regardless of controls for selection and endogeneity
Move 10-11 hours increases sales by 13.3%
 Move 20-21 hours increases sales by 6.8%

 Interaction terms with advisor hours and other
variables that reflect the potential learning capacity
of a young firm


E.g., age, size of top management team and business acumen
All statistically insignificant.
Key Findings for Patents
 Accounting for the possible endogeneity of hours
spent:

Hours do not statistically increase the probability of patents.
 Without controlling for endogeneity:
 There is a statistical association between hours and patents
which is significant at the 5% level of significance.
Move 10-11 hours increases the probability of patents by 0.6%
 Move 20-21 hours increases the probability of patents by 0.3%.

Key Findings for Financing
 Additional mentor hours  higher probability of financing
 Regardless of controls for endogeneity
 Accounting for endogeneity:



A move from 10-11 hours increases the probability of angel financing by 0.7%,
A move from 20-21 hours increases the probability of angel financing by 0.4%.
Not controlling for endogeneity:


A move from 10-11 hours increases the probability of angel financing by 0.6%
A move from 20-21 hours increases the probability of angel financing by 0.3%
 Interaction term between hours and business acumen
 Reduces effect of hours on financing by 10%
 Due to learning capacity of the firm
Key Findings for Alliances
 Positive association between hours and obtaining a
strategic alliance


Statistically insignificant with controls for endogeneity
Statistically significant without controls for endogeneity
Move 10-11 hours increases the probability of an alliance by 0.6%
 Move 20-21 hours increases the probability of an alliance by 0.3%

Conclusions and Takeaways
Primary Conclusions
 Advising hours significantly and positively impact sales
and financing, regardless of econometric controls for
sample selection and endogeneity.
 There is a positive association between hours and patents
and alliances, but the causality is more ambiguous.
 Cost effective:



Entrepreneurial firms had raised $6,545,000 in financing
The program costs were totaled at $662,360
Ratio of financing raised per dollar of cost is $0.10
Takeaways
 What's working here?


Selecting firms with high potential
Intensive advising rather than minimal advising
 Should more advising programs be sponsored with
public dollars?




Ideally, we need more research with larger panels and equal
attention to data collection
Record keeping (and likely advice) differs across advisors / programs
Need to be willing to target selective firms; can't expect advising to
pay off equally for all types of firms
Need to recognize that minimal advising is likely to have minimal
payoffs; UP TO A POINT, more is better