Transcript Kwazulu-Natal ICTE Cluster: Strategy and Action Plan

Institutions, innovation and growth
Andrés Rodríguez-Pose
London School of Economics
IRIS
Stavanger, 5 May 2009
Introduction
 Do institutions matter for innovation and growth?
 The link between institutions and growth has traditionally
been overlooked
 Traditional approaches to spurring economic development
were based on:
1. Greater investment in the stock of physical capital (neoclassical)
2. Endogenising innovation, technology, and physical capital (endogenous
growth)
3. Agglomeration, externalities, and distance (new economic geography)
 This system tended to work in the past
1. National intervention in the postwar period had coincided with growth and
a reduction in disparities
2. The first two approaches informed the European regional development
effort during the reform of the Structural Funds
Do institutions matter for regional development?
Andrés Rodríguez-Pose
Page 1
Do traditional approaches still work?
 These approaches, however, seem today less capable of
explaining economic growth and development
1. Regional convergence shifted to stability and divergence
- Cross-country convergence
- Intranational stability or even strong divergence
2. Growth in the residual factor
 Institutions being rediscovered
1. Work by sociologists, geographers, and some economists since the mid1980s
2. Becoming mainstream in economics
- ‘Institutions matter’ (Hall and Jones, 1999; Rodrik et al, 2004)
- Research now trying to understand which institutions matter
- Some institutional arrangements are more appropriate than others, depending
on the circumstances (Aghion and Howitt, 2006)
Do institutions matter for regional development?
Andrés Rodríguez-Pose
Page 2
Aim of the presentation
 If
1. We can explain less and less how economic development is being generated
2. The role of institutions as a shaper of economic development is being
regarded as more prominent
3. Economic development efforts have, by and large, overlooked local
institutions
 Ergo
1. Institutions should become an essential element of development effort.
 But, is that the case?
 If so, how can institutions be included in the local innovative
and development effort?
Do institutions matter for regional development?
Andrés Rodríguez-Pose
Page 3
The role of institutions in growth
 Across the world development strategies seem to be becoming
less effective
1. Neoclassical orthodoxies regarded as inadequate and perhaps providing
imperfect interpretations of regional development (Yeung, 2000)
 Growing attention has been paid to other factors and,
especially institutions
 Belief that different local institutional arrangements are key to
our understanding of development
1. Emphasis on social capital (Putnam, 1993, 2000)
2. On institutional thickness (Hudson, 1994; Amin and Thrift, 1995)
3. Learning regions (Morgan, 1997)
Do institutions matter for regional development?
Andrés Rodríguez-Pose
Page 4
But, what are institutions?
 Concept of institutions is:
1. Subjective
2. Controversial
3. Difficult to operationalise
 Numerous nuances and distinctions in concepts
1. Formal vs. informal institutions
2. Informal institutions of community (norms, trust, face-to-face) vs. social
capital
3. Institutions vs. organisations (rules vs. players) (North, 1990)
Do institutions matter for regional development?
Andrés Rodríguez-Pose
Page 5
How do institutions promote development?
 Markets as socially constructed (Bagnasco, 1988)
1. This makes the role of institutions greater than simple regulators of
economic activity
2. They determine the level of activity and its efficiency
3. They facilitate knowledge and innovation transfer
4. They shape incentives and disincentives
 Different forms of institutions are in constant interaction
1. The balance between formal and informal institutions (society and
community)
 ‘Institutional thickness’ determines the development capacity
of every territory
Do institutions matter for regional development?
Andrés Rodríguez-Pose
Page 6
Society, community and development
 Communities (Gemeinschaft or social capital) are
complemented by Society
 Communities refer to:
1. features of group life (i.e.: informal rules and norms, tradition and
social expectations, contacts and connections, and relationships).
 Society refers to:
1. universal and transparent rules (i.e.: property rights, rule of law,
promotion of individual choice, and factor mobility).
Do institutions matter for regional development?
Andrés Rodríguez-Pose
Page 7
The economic impact of communities
 The optimistic view: Communities are good for
development
1. Community improves:
- provision of public goods (Coleman, 1990; North, 1990)
- market organisation (Granovetter, 1985)
- promotes the embedding of firms in efficiency-enhancing networks of
relationships (Grabher, 1993)
- generates institutions such as trust (Fukuyama, 1999; Putnam, 2000;
Bowles and Gintis, 2002)
- reduces transaction costs (Storper, 1997)
- reduces moral hazards and free-riding (Streeck, 1992)
- mitigates information asymmetries (Granovetter, 1985; Wade, 1987)
- matches individual and aggregate interests (Rodríguez-Pose, 1999)
Do institutions matter for regional development?
Andrés Rodríguez-Pose
Page 8
The economic impact of communities (II)
 The pessimistic view: Communities are bad for
development
1. Community leads to:
- pervasiveness of rent-seeking (Trigilia, 1992)
- insider-outsider problems
2. unsatisfactory distributional effects
- clientelism, and nepotistic practices (Trigilia, 1992)
- it may be a second best solution in the absence of developed societal
institutions
3. Communities may
- generate greater social polarization
- hamper equal opportunity
- exacerbate problems of imperfect competition, impacted information,
and principal-agent problems (Durlauf, 1999; Durlauf and Fafchamps,
2004).
Do institutions matter for regional development?
Andrés Rodríguez-Pose
Page 9
The economic impact of society
 The development of societal institutions is generally
perceived as a positive sign
 But, under certain circumstances it may be detrimental
for development:
1. In contexts dominated by weak group life, societal rules and laws
cannot always insure against opportunistic defection (Streeck, 1991)
2. Higher transactions costs and costly conflict resolution through
litigation, i.e. a confrontational society (Storper, 2004).
3. Inadequate production of public goods (education and training,
environmental management, or innovation)
Do institutions matter for regional development?
Andrés Rodríguez-Pose
Page 10
Interaction between society and community
 Society and community are thus generally viewed as
mutually opposed
 But they interact in all contexts
 Relations in any space take place in the form of:
1. ‘bonding’: within community relations
2. ‘bridging’: across community relations (Putnam, 2000)
 A system of checks and balances can be developed:
1. Developed communities can offset the potentially negative effects of
society
2. A developed society can offset the potentially negative effects of
community
Do institutions matter for regional development?
Andrés Rodríguez-Pose
Page 11
Society and community interaction
Society
High
Low
Community
Low
Sub-optimal
Worst case scenario
Do institutions matter for regional development?
High
Optimal
Sub-optimal
Andrés Rodríguez-Pose
Page 12
Society and community (II)
SOCIETY
COMMUNITY
LOW
HIGH
HIGH
Responsibility without autonomy: individual
agency but insufficient voice
 dominance of arms-length transactions;
 moral hazards contained when information
transparent only
 asymmetrical strength of individual agents;
 certain interests have no voice/no autonomy
(effort and reward markets are imperfect);
 costly conflict resolution and
confrontational society;
 insufficient public goods;
 tendency toward inequality;
Autonomy with responsibility: a good balance of voice
and agency
 costs for all kinds of transactions minimized;
 moral hazards contained for both transparent and
specific information;
 autonomy strong and participation high (strong
community reduces losses due to anonymity,
fragmentation);
 rent-seeking contained through competition, entry
and exit (strong society modernizes
communities);
LOW
Neither autonomy nor responsibility: chaos and
the law of the jungle (voice of the powerful,
ubiquitous agency problems)
Autonomy without responsibility: collective voice, but
with agency problems
 High transactions costs for arms-length
transactions;
 prevalence of primitive, non-modern
communities;
 rent-seeking widespread;
 individual voice subjugated to groups
(insufficient competition and mobility);
 insufficient generalized trust and confidence;
 skewed distribution of public goods
 high costs for all types of transactions
 generalized instability: weak societal rules,
weak local bonding;
 high moral hazards, generalized
opportunism, low sanctions for defection
and cheating;
 public goods destroyed or stolen or
appropriated;
Do institutions matter for regional development?
Andrés Rodríguez-Pose
Page 13
Integrating institutions in development
 Developing and improving institutional capacity is therefore
increasingly regarded as key for development
 There is a need to integrate institutions in development
strategies
 But this is easier said than done
 Several factors limit the integration of institutions in
development strategies
Do institutions matter for regional development?
Andrés Rodríguez-Pose
Page 14
Key problems
 Measuring institutions is difficult and controversial
1. Local institutional constructs tend to be intangible (Fine, 2000)
2. Identical formal institutions may yield very different economic returns in
different contexts
 Efficient institutions are context- and geography-specific
1. What is solid and efficient in one region may not be so in another
2. There is a need to integrate institutions in development strategies
 The effectiveness of institutions changes with time
1. What are ‘good’ institutions in one period may no longer be appropriate in
another (Storper, 2005)
2. Institutions adapt (institutional migration)
 Institutions are extremely resilient to change
1. Persistence of family structures (Duranton et al., 2009)
Do institutions matter for regional development?
Andrés Rodríguez-Pose
Page 15
Key problems (II)
 Identifying the right mix of institutions is problematic
1. More than the density of institutions…
2. It is the quality of institutions
 Endogeneity between institutions and development
1. Direction of causality difficult to predict
 Endogeneity between institutions and other constituents of
growth
1. Institutions may hide the effect of other factors and especially human
capital (Glaeser et al., 2004)
Do institutions matter for regional development?
Andrés Rodríguez-Pose
Page 16
So, how to link institutions and innovation?
 With difficulty and not devoid of problems…
Do institutions matter for regional development?
Andrés Rodríguez-Pose
Page 17
Approaches to the analysis of innovation
 How can innovation and growth be generated?
 3 traditional approaches:
1. The ‘linear model’
- Analysis of the link between R&D, patents and growth
- Fundamentally quantitative (econometric analysis)
- Conducted mainly by ‘mainstream’ economists
2. The ‘systems of innovation’ approach
- Analysis of the ‘territorially-embedded’ institutional networks that favour the
generation of innovation
- The capacity to set these networks depends in turn, on a series of social and
structural conditions (‘the social filter’)
- Fundamentally qualitative
- Conducted mainly by geographers, evolutionary economists, and some economic
sociologists
3. Knowledge spillovers
- Look at the diffusion and assimilation of innovation
- Quantitative and qualitative
- Economists and geographers
Do institutions matter for regional development?
Andrés Rodríguez-Pose
Page 18
Linking the approaches
Link between
investment in R&D, patents,
and economic growth.
(Fagerberg 1988, 1994 and 1997; Grossman and
Helpman 1991;Maurseth and Verspagen 1999)
Geographical diffusion of
regional knowledge spillovers;
(Anselin et al. 1997, Adams and Jaffe 2002; Audretsch and
Feldman 2003, Leamer and Storper 2001, Storper and
Venables 2004, Sonn and Storper 2005)
Existence and efficiency of
regional innovation systems.
(Camagni 1995, Becattini 1987, Morgan 1997 and
2004, Cooke et al. 1997, Iammarino 2005,
Rodriguez-Pose 1999)
Do institutions matter for regional development?
Andrés Rodríguez-Pose
Page 19
Linking the approaches (II)
Link between
investment in R&D, patents,
and economic growth
Geographical diffusion of
regional knowledge spillovers
Existence and efficiency of
regional innovation systems
Do institutions matter for regional development?
Andrés Rodríguez-Pose
Page 20
Innovative factors behind growth
Investment in R&D and patents, when
other factors are controlled for, do not
lead to greater growth
Lag 0
Constant
Log GDPpc
R&D Filter
Social Filter
Clusterisation Index
2
R
F
Number observations
Do institutions matter for regional development?
2.157***
0.350
0.757***
0.038
0.009
0.008
0.049***
0.005
0.013**
0.005
0.925
614.21
1756
Andrés Rodríguez-Pose
But, social
conditions
(fundamentally
education)
matter
As do
institutional
conditions
(specialisation
in clusters,
focus and
diversification)
Page 21
Innovative factors behind growth (II)
But R&D and patents become significant
with time
lag 1
lag 3
lag 6
1.956***
0.314
0.785***
0.034
1.683***
0.243
0.853***
0.028
1.392***
0.248
0.899***
0.028
R&D Filter
0.015**
0.007
0.017***
0.006
0.015**
0.007
Social Filter
0.043***
0.005
0.011**
0.005
0.031***
0.004
0.006
0.005
0.024***
0.005
0.002
0.005
0.932
582.94
1596
0.947
705.77
1276
0.968
1281.37
796
Constant
Log GDPpc
Clusterisation Index
2
R
F
Number observations
Do institutions matter for regional development?
Andrés Rodríguez-Pose
The social filter
(fundamentally
education)
remains the
most
significant
throughout
But institutions
lose
significance in
time
Page 22