Innovation, disruptive technologies and SMEs

Innovation, Disruptive
Technologies, and SMEs:
Constraints and Policy
Alan Hughes
Centre for Business Research
University of Cambridge
Presentation at the Six Countries Programme
Conference on SMEs and Disruptive
Technologies
Vancouver June 5-6 2003
© Alan Hughes June 2003
Disruptive Technology, Innovation
and High-Tech SMEs
• Disruption based on technological innovation
measurable only with hindsight?
• Evidence base for policy focussing on support for
innovative experiments
–
–
–
–
Nature and Incidence of knowledge based SMEs
Distinctive business characteristics
Innovative behaviour
Constraints on meeting business objectives
• Problems of Public Policy Support and Evaluation for
Technology based Innovation
© Alan Hughes June 2003
Focus of Presentation
• CBR panel surveys of UK SMEs
• Evaluation of SMART scheme to
support SME technological innovation
© Alan Hughes June 2003
Sources of Data
• Cosh, A.D. and Hughes, A. (2000) (eds) British Enterprise in
Transition: Growth Innovation and Public Policy in the Small and
Medium Sized Enterprise Sector 1994-1999. ESRC Centre for
Business Research University of Cambridge, Cambridge.
• PACEC (2001) Evaluation of SMART 2001. by Boyns,N. Cox,M.
Hughes,A. and Spires,R. DTI Evaluation Report Series No 3
September
• Cosh,A.D. and Hughes,A.(2003) Enterprise Challenged: Small
and Medium Sized Enterprises in the UK 1999-2003 ESRC
Centre for Business Research University of Cambridge,
Cambridge.
© Alan Hughes June 2003
CBR Surveys
• Regular biennial survey of independent SMEs in the
UK
• Manufacturing and business services
• Size Stratified surveys over 2000 firms
• Latest results based on 5th survey in 2002
• Surveys incorporate questions on innovation input
and output constraints and over 200 company
specific variables on structure and performance
• Allows comparison of ‘high-tech’ and ‘conventional’
© Alan Hughes June 2003
Economic Indicators and the CBR survey periods
16
14
12
10
GDP grow th
(constant
prices) %pa
Short-term
interest rate %
%
8
6
RPI inflation %pa
4
Unemployment
rate %
2
0
-2
-4
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
© Alan Hughes June 2003
1999
2000
2001
2002
Macroeconomic Background
• Latest survey took place against relatively stable
macroeconomic conditions with low rates of inflation
and interest rates compared to earlier years.
• However
– stagnating demand especially for manufacturing output,
– falling capital markets especially for technology related
stocks.
– expect, therefore,
• some increase in the importance of demands constraints in
manufacturing, and
• some intensification of financial constraints for high-technology
sectors compared to earlier survey periods.
© Alan Hughes June 2003
Size Sector and Age Distribution
• 2002 sample of 2127 firms
– 35% <10 employees micro
– 50% 10<100 employees
– 15% 100<500 employees
– 61% drawn from manufacturing and
around 39% from business services
– around a half formed after 1980, and a
quarter after 1990, around 14% date from
the pre-war period
© Alan Hughes June 2003
•
•
•
•
•
•
SMART
Smart is the Small Business Service (SBS) initiative that provides grants to help
individuals and small and medium-sized businesses to make better use of
technology and to develop technologically innovative products and processes.
Technology Reviews
Grants of up to £2,500 for individuals and small and medium-sized firms (fewer
than 250 employees) towards the costs of expert reviews against best practice.
Technology Studies
Grants of up to £5,000 for individuals and small and medium-sized firms (fewer
than 250 employees) to help identify technological opportunities leading to
innovative products and processes.
Micro Projects
Grants of up to £10,000 to help individuals and micro-firms (fewer than 10
employees) with the development of low cost prototypes of products and
processes involving technical advances and/or novelty.
Feasibility Studies
Grants of up to £45,000 for individuals and small firms (fewer than 50
employees) undertaking feasibility studies into innovative technologies.
Development Projects
Grants of up to £150,000 for small and medium-sized firms (fewer than 250
employees) undertaking development projects. A small number of exceptional
high-cost projects may attract grants of up to £450,000.
© Alan Hughes June 2003
SMART CHANGES 2003
•
Smart research and development (R&D) project grants will be replaced
by a new R&D grant product on 1 June 2003.
•
The important differences are:
– Research projects (previously called Feasibility studies) - 60% of eligible
project costs up to a maximum grant of £75,000
– Development projects - 35% of eligible project costs up to a maximum grant
of £200,000
– Exceptional development projects - 35% of eligible project costs up to a
maximum negotiable grant of £500,000
– Micro projects - 50% of eligible project costs up to a maximum grant of
£20,000.
© Alan Hughes June 2003
What is a small Knowledge-Based/High-Tech Small firm?
Variety of definitions proposed but consistent elements are:
Dependence upon application of scientific or technological skills, or knowledge,
in production of new goods or services
Emphasis upon technology component of activities as a source of competitive
advantage
Independent owner/controlled status with employment less than some varying
specified cut off point (e.g. 200 or 500 workers)
In practice in most studies hi tech firms are identified as “firms in hightech industries” and high-tech industries are then identified as:
Industries with high or above average R & D/sales ratios (e.g 20% above mean
for all industries)
Industries with above average shares of technologists, scientists and higher
professionals in their labour force (e.g 20% above mean for all industries)
But this is misleading high-tech firms can occur in any sector
© Alan Hughes June 2003
High-Tech Firms Defined
 CBR SME dataset allows classification of firms irrespective of
industry in to High-Tech and conventional categories by using
firm level data
 Classification is possible in terms of various factors eg:
-
R & D/Sales Ratio
-
Proportion of Technologists Scientists and Higher Professionals
-
Self perception of sources of competitive advantages
For simplicity High-Tech defined here as firm with R&D/Sales  2%
-
© Alan Hughes June 2003
High Tech Industries v. High Tech
Firms
• Many firms in high tech inds have low R&D/Sales
(45% have zero R&D)
• Many firms in conventional inds have high R&D/Sales
(10% > 3%R&D/sales)
• Conventional Sectors c.75% of UK firms therefore
more R&D intensive firms in conventional sectors
• High Tech Firms R&D intensity 1997-2002
– Mean R&D/Sales
1997 12.6%
– % with R&D/Sales >3%
1997 57.4% 2002 56.1%
© Alan Hughes June 2003
2002 16.5%
The Distribution of Firms by R&D to Sales Ratio in High-tech and Conventional Sectors 2002 – Firm-based
definition of high-tech.
Number of
Firms
Mean
R&D/Sales
Ratio
%
0
>0<1
1<2
2<3
>3
R&D/Sales Ratio %
High-tech Manufacturing
Conventional Manufacturing
All Manufacturing
183
918
1101
14.7
0.2
2.6
2.7
71.7
60.2
0.5
17.0
14.3
1.6
11.3
9.7
20.2
0.0
3.4
74.9
0.0
12.4
High-tech Services
Conventional Services
All Services
238
435
673
18.0
0.1
6.4
39.5
83.7
68.1
3.4
12.4
9.2
3.8
3.9
3.9
11.8
0.0
4.2
41.6
0.0
14.7
All High-tech
All Conventional
All Firms
421
1353
1774
16.5
0.2
4.1
23.5
75.5
63.2
2.1
15.5
12.3
2.9
8.9
7.5
15.4
0.0
3.7
56.1
0.0
13.3
The differences across sectors in the proportion of firms with no R&D, in the median R&D/Sales ratios and in the
distribution of firms across size categories of R&D/Sales, are all significant at the 5% level or better.
© Alan Hughes June 2003
Why are Knowledge-Based
/High-Tech Firms Important?
“Smaller technology firms are entrepreneurially driven and form an
important seedbed out of which innovation experiments flow to be
tested in the wider economic context. Smaller firms with
competitive advantages can grow very rapidly and diffuse their
products into the economic structure, so raising the level of
average practice economic performance in their sector. It is
because of this connection between diffusion and average
efficiency that barriers to growth for the individual firm become
barriers to raising the national level of economic performance”
ACOST (1990)
© Alan Hughes June 2003
High Tech Producers and High
Tech Users
• Focus here is on high tech producers
• Impact on productivity growth at macro
level depends upon both output of hightech producers and high tech users
– US productivity acceleration post 1995
mainly accounted for by wholesaling,
retailing, financial services
© Alan Hughes June 2003
Golden Oldies v New Kids on
the Block
• Great current emphasis on spin-outs from universities and startup… New kids on the block
– Seed bed role
– But tiny proportion of all knowledge based start-ups
– Very small proportion grow substantially (e.g. in 2002 only
125 of 21000 US university licensed firms yielded >
$1.million)
• Emphasis also needed on existing firms…Golden Oldies
– Much more important for productivity growth
– Key difference UK/EU v, USA not start-up but rapid growth
after start-up
– Sustained innovation to disrupt leader
– Examine constraints on innovation and growth
© Alan Hughes June 2003
Competitive Position
• All firms have low numbers of perceived
competitors (4-6)
• High Tech perceive greater proportion of
these overseas
– Mfg
– Services
31%
15%
• High Tech esp. services more reliant on
largest customers (40%-50% sales from top 5
customers)
© Alan Hughes June 2003
Why Collaborate?
• For all firms expanding product range
expertise comes top
• For High Tech relatively more important
– To share R&D
– Develop specialist products
– Access overseas markets
• For High Tech relatively less important
– Keep current customers
© Alan Hughes June 2003
Competitive Advantage
• High Tech Firms emphasise absolutely
and relatively
– Product Quality and Specialisation
– Creativity and Flair
• High Tech Firms place low emphasis
absolutely and relatively
– Price
– Speed of service
© Alan Hughes June 2003
Innovation Activity
• High tech firms are more innovative
– Product
– Process
– New to firm
– New to industry
© Alan Hughes June 2003
Innovative activity in high-tech and conventional firms 1999-2002
Percentage of Firms Introducing:% of firms
Product Process Product or
Novel
Novel
Planning to
No of Innovation Innovation Process
Product Process Introduce New
+
Firms
Innovation Innovation Innovation
Products or
Processes in
Next 3 years
High-tech Manufacturing
204
Conventional Manufacturing 1072
84.3
47.3
60.8
44.3
91.2
58.2
58.3
20.4
28.9
14.3
94.6
58.8
High-tech Services
Conventional Services
272
542
71.3
45.0
55.7
36.0
76.8
51.8
44.1
22.0
28.2
14.0
78.9
53.4
All Firms
2130
53.5
45.3
62.2
27.7
17.5
63.6
+
Including innovations in supply, storage or distribution systems. All measure of innovative performance across
firms are significantly different at the 5% level or better.
© Alan Hughes June 2003
Significant Constraints
• High Tech mfg
–
–
–
–
Demand
Finance for expansion
Marketing skills
Overseas market access (relatively)
• High Tech Services
– Overdraft finance
– Marketing skills
• Acquisition of Technology not a common
constraint
© Alan Hughes June 2003
Key Characteristics and
Evidence Based Policy Issues
• High-tech firms not high-tech sectors
• Product Development Focus with few Competitors
• Overseas Orientation
– Exports and Collaborative agreements
– Collaboration
• High Collaboration
– Customers and Suppliers key collaborators
• Constraints
– Finance still an issue
– Marketing skills (link to focus on new prod. devpt)
© Alan Hughes June 2003
CAPITAL MARKET FAILURE AND THE
RATIONALE FOR SMART
 SMEs invest too little in innovative technology for new
products and / or services
 SMEs have difficulty in obtaining debt or equity finance for
research and innovation in packages of less than around
£250,000
o Cost of due diligence, including technological
appraisal, makes deals below this size unprofitable for
lenders and investors
 Scheme has disbursed over £250 million to over 5000
firms since 1988
© Alan Hughes June 2003
SMART WINNERS (1)
• At time of award
–
–
–
–
–
93% independent single site businesses
50% less than 10 staff, 30% 10<50 staff
86% were formed as new start ups
66% less than 10 years old
87% already had R&D expenditure and staff
• Median £70K
• Median 2 full time staff
© Alan Hughes June 2003
SMART WINNERS (2)
1988-1998
• Success Rate
– 14700+ applications, 4000+ awards
– Success rate for applications
• Lower for smaller firms
• Rising as scheme matures
• Sector concentration
– IT, computing
– Scientific instruments
– Electrical engineering
© Alan Hughes June 2003
Numbers of applications and success rates, 1988-1998
a) No. of b) No.
applications
of
awards
By period:
88-92
93-95
96-98
By size of
firm:
Micro
Small
Medium
Total
c) No. of
unsuccessful
applications
d) Application
success rate, (b)
as % of a))
6,563
5,119
3,088
1,291
1,384
1,359
5,272
3,735
1,729
20
27
44
11,049
3,021
700
14,770
2,175
1,405
454
4,034
8,874
1,616
246
10,736
20
47
65
27
© Alan Hughes June 2003
Sectoral breakdown of award winners and market penetration
Selected
SIC
Divisions Description
23-24
Fuels / Chemicals
a) No. of
awards winners
'88-'98
243
b) No. of UK SME
Establishments, 1998
(000s)
6
c) Penetration
rate (a) as
% of b))
4.34
29
Machinery and
equipment nec
327
18
1.82
30
Office machinery and
computers
161
2
7.12
31
Electrical
machinery/apparatus
158
7
2.26
32
Radio, tv/comms
equipment
248
4
6.70
33
Medical, precision
instruments, etc
486
7
6.67
72
Computing and related
activities
245
119
0.21
73
Research and
development
162
3
4.86
All SIC
Divisions
3,144
© Alan Hughes June 2003
2,307
0.14
Evaluating ‘SMART like’
programmes
• Counterfactuals, Selection Bias and Information failure
– Randomization
– Matched Control groups..multiple characteristics
– Selection modelling
– Instrumental variables
– Subjective Counterfactuals
– Scheme Based Information
• Needle in a haystack
• Skewness of Outcomes
• Additionality
© Alan Hughes June 2003
Smart Evaluation Methodology
• Comparison of Successful v. Unsuccessful
applicants
• Survey data plus program information
• Econometric analysis with correction for
selection restricted to 1995-8 data gaps for
unsuccessful firms pre 1995
• Subjective counterfactuals
• Case Studies
• Focus on award winners to 1998..post effects
© Alan Hughes June 2003
Award Winners v Losers 1995-8
• Winners were on average faster
growing post award
– Robust to sample selection bias
– But rarely statistically significant
© Alan Hughes June 2003
Annual rates of change in award winners' and
unsuccessful applicants' turnover, exports and
employment.
Award Winners
Unsuccessful Applicants
Mean values of % change in:
Turnover Exports Employment
N % N
%
N
%
147 22 86 28 165
17
152 13 78 21 160
14
© Alan Hughes June 2003
ble 5.2
Changes in turnover, employment, exports and productivity, by year of
application.
Median values of % change in:
Turnover
Year:
1995
Award Winners
Employment
Exports
Productivity
N
%
N
%
N
%
N
%
34
100.0
38
33.3
22
122.2
34
27.3
Unsuccessful Applics.
1996
Award Winner
20
61.8
23
37.5
23
33.3
17
0.0
29
52.9*
33
42.9*
14
137.6
29
25.0
Unsuccessful Applics.
1997
Award Winners
13
10.0
17
0.0
5
100.0
12
17.5
37
0.0
41
0.0
19
0.0
37
0.0
Unsuccessful Applics
1998
Award Winners
16
27.7
15
25.0
6
85.3
15
0.0
58
22.0
64
0.0*
35
42.9
57
0.0
Unsuccessful Applics
14
29.2
17
50.0
9
33.3
14
0.0
Source: PACEC - Surveys of Award Recipients and Unsuccessful Applicants
Note * Significant difference at the 10% level between Award Winners and
Unsuccessful Firms (Mann-Whitney Test)
© Alan Hughes June 2003
Table 5
The Impact of Awards on Performance: Estimated coefficients on
"success" Dummy variable in Regression Analysis of Performance
Effects (corrected for selection bias)
Exports
Turnover Employment Productivity
Year
N
N
N
coeff.
coeff.
coeff. N
coeff.
1995
53
0.252
(0.08)
51
0.180
(0.18)
51
0.903
(1.01)
28
1.706
(0.49)
1996
41
1.331
(1.62)
40
1.122
(1.72)*
40
0.674
(1.86)**
17
1.352
(1.74)
1997
53
-0.182
(-0.71)
56
-0.208
(-1.20)
52
0.151
(0.62)
24
-0.274
(-0.62)
1998
71
-0.401
(-0.17)
72
0.175
(0.27)
70
0.840
(0.63)
44
3.361
(2.09)
Note:
* Significantly different from zero at 10% level
** Significantly different from zero at 5% level
't' statistics in brackets
© Alan Hughes June 2003
Subjective Estimates of SMART
Impact
• Turnover
– C.57% no change, 41% some increase
• Exports
– C 70% no change, 28% some increase
• Employment
– C 64% no change, 32% some increase
• Profitability
– C 46% no change, 53% some increase
© Alan Hughes June 2003
Skewness of Outcomes
• Typical result is no change
• But some firms outstandingly successful
– Top 5% growers 50% all SMART
associated sales
– Top 20% growers 80% all SMART
associated sales
• Pareto Distribution
© Alan Hughes June 2003
Total value of sales resulting from projects
(cumulative % sales against cumulative % companies)
Cumulative % of total sales
100
90
80
70
60
50
40
30
20
10
0
0
10
20
30
40
50
60
70
Cumulative % of companies
:
© Alan Hughes June 2003
80
90
100
Conclusions on SMART
• Successful scheme with positive outcomes
dominated by small proportion of high fliers
– Skewness to be expected
– Focus on average effects misleading
• Scheme has evolved in response to monitoring of
use and outcomes
• Evaluation requires that information requirements on
baseline data and outcomes should be built into
scheme
• Further work should focus on characteristics of super
growers and management constraints/strategy
© Alan Hughes June 2003