2. Globalization, Growth and Inequality:
83
Recommended Readings • Jones, C. Intro to Growth, chap 2. See below. • Eichengreen et al. Growth Slowdowns Redux
84
Starting Point: Neoclassical Growth Model – Solow (1956) • Equations:
Source: Jones, 2000
85
Assumptions • Perfect Competition. • Diminishing marginal returns and lack of scale economies. • Neutral technological change, affecting L and K equally.
86
Growth rate? • Growth depends on K/L ratio, which is determined by diminishing returns. • Both output and capital stock grow at a same exogenous rate – technical progress. • But how to escape diminishing returns?
87
The Role of Technology in Solow • Technology determines Long Run Growth Rates.
Rate of Tech Progress
Rate of Growth
88
And the consequences of that are ... • Technology as a Public Good. – Non-excludable. – Non-rivalrous.
• Therefore, freely available -> same rate for everyone.
89
Dynamics of Transition • Diminishing Returns • Poor grow faster – Convergence is Expected. • But really?
90
Growth Accounting
• Abramovitz (1956): only a small part of US productivity growth could be explained by factor growth – most is unexplained, residual, to be classified as total factor productivity growth.
91
Growth Acounting in Numbers (Denison 1967, 1976)
Source: Fagerberg, 1994.
92
Conclusion: • Factor Productivity does not account for long run growth. • Most factors induced are affected through causal links not considered. Eg: labor-intesive technology vs capital-intensive technology. • Convergence among countries is questioned Transitional dynamics?! • Assumptions are questioned! 93
But what is convergence really? • Sigma-Convergence: an expected reduction in the dispersion of levels of income accross countries • Beta-Convergence: faster growth rates for poor countries.
94
Types of Convergence • Absolute Convergence: Lower GDP -> faster growth. • Conditional Convergence: GDP per capita converges to a contry-specific long run level determined by specific structural characteristics. • Club Convergence: Some countries have similar trajectories. 95
Implications • Absolute Convergence: Poverty disappears by itself. • Conditional Convergence: investment should be focused on infrastructure and institutions. • Club Convergence: income transfers may be needed, as initial income matters.
96
But the thing is ... • Landes (1998): the difference in income or productivity per head between the richest and poorest country in the world has substantially increased over the last 250 years. • 250 years … Right when we started beating the Malthusian trap …
97
“New Growth Theory” • Motivation: problems associated with diminishing returns assumption. • Emphasis on non rivalry of knowledge, with variable exclusion – therefore, not a public good exactly.
98
The Non-Rivalry Hypothesis and its Consequences ... • Non rival goods may imply fixed production costs and zero marginal cost. • Hence, the need for increasing scale returns and imperfect competition. • Therefore, we would get to the point where we could make technology production endogenous, through the R&D sector.
99
Romer • Growth is determined by R&D in the advanced world. • When knowledge is considered an input, the production function presents increasing returns.
100
The Production of Ideas •
is the number of new ideas generated at any moment in time. • Assumption – productivity through time depends on: – The ideas stock in any given moment. – The amount of researchers.
101
Increasing Versus Constant Returns • Individual researchers take the average productivity of research as constant, so therefore, we have constant returns to research • An individual in research generates new ideas. • However, in the economy as whole, the production function of new ideas does not show constant returns of scale. • Although the variation for each individual researcher is small, it varies with the agregate research effort. It is taken as exogenous to the agent. • If we have spillovers (standing on the shoulders), but if there might be effort duplication.
102
Convergence? • Growth depends on: – the amount of resources devoted to innovation activity (R&D), – the degree to which new tech can be privately appropriated (the degree of monopoly).
• The beneficial external effects of capital accumulation outweigh the detrimental consequences of increasing capital per worker – not necessarily diminishing returns. • Hence, the marginal productivity of capital does not decline with increasing GDP per capita. Rich countries stay rich, poor countries stay poor.
103
Technical progress Vs Capital Accumulation • They tend to go together. • High growth also implies high growth in physical capital, but this is a result not a cause.
104
Catch Up? • There is nothing automatic about catching up. • Need to work – invest, but how?
105
Catching Up
Source: The Economist
106
Middle Income Trap • According to WB calculations, only 13 countries have managed to move from middle income to high income status from 1960 to 2008. • A Trap implies an equilibrium – club convergence?
107
What it is: • Deceleration of productivity growth. • Solow claimed with diminishing returns when you got rich, your growth rate would decelerated, but some stopped before that. • Mainly Latin America and a few in Southeast Asia. • Many in Africa haven’t even reached this point yet. 108
Causes: • • • • • •
Diminishing Returns to Capital Exhaustion of Cheap Labor and imitation gains Human Capital Institutions Infrastructure Lack of access to finance
109
Middle Income Trap
• Middle-income countries are defined according to WB definition (GDP per capita between $400-$4,000 in 1985 and $1,036-$12,615 in 2013. • Source: Kiel Institute Calculations based on IMF data. Taken from 110 the global economic symposium.
Source: The Economist
111
Source: ECLAC
112
Usual Policy Recommendation • Go back to Solow’s model -> technology. • But why can’t they just increase their productivity by increasite their rate of technology progress? • Usually, human capital.
113
It’s Education, Stupid! Source: weforum
114
GNP below $15,000 but
115
3. Trade?
116
Recommended Readings • WTO - Gravity Model WTO - Trade Analysis
117
Traditional Trade Models • Comparative Advantage and Specialization – Ricardo: • Comparative advantage comes from technology
– H-O: • Comparative advantage comes from factor abundance
– Product life-cycle: • Dynamic comparative advantages
• Others: – Imperfect Competition 118
Gravity Model
119
Gravity Model of Trade
Where is the billateral trade flow between countries i and j, G is a constant, is the GDP of the importing country, is the GDP os the exporting country and is the distance between them. 120
For estimation: • In logs and with an error term:
Where you expect the first two coefficients to be positive and the third to be negative. 121
Issues • Theory? • Control variables • Other uses
122
Why are we looking at this? • • • •
Developed versus Developing North versus South Being big (and rich) pulls: Being far away repulses: – Trade flows – Migration flows – FDI flows – Technology 123
Bottom line: you want to big and in Europe ...
124
3a. Trade facts
125
Distance and Trade
Source: McKinsey
126
Distance 2
Source: McKinsey
127
Trade and the Recession
Source: McKinsey
128
Commodities
Source: McKinsey
129
Trade Partners
Source: McKinsey
130
North and South (and China)
Source: McKinsey
131
Why not NAFTA? China!
132
4. Technology Diffusion
133
Recommended Readings • Fagerberg - Growth differences
134
What is knowledge/technology? • • • •
Intagible Difficult to Measure But not a public a good apparently! Technology as public good explains very little of the observed differences in growth across countries. • A different conception of technology, less public – at least partly appropriable, and less neutral – because technological progress and factor accumulation may to some extent be complementary. 135
Knowledge as Public Good: • Public Good: – Nonexcludable: if research results are disseminated through the regular channels of communication, by means of publication in journals or books or websites, it enters the public realm and therefore becomes available to anyone who searches for it. – Nonrival: knowledge or goods may not be consumed by many individuals at the same time, but additional users of said knowledge will not decrease the amount or quality available to others.
• But is it really? • Knowledge (know-how, expertise) versus Content 136
Codifiable Knowledge: • • • •
Found in Publications. Easily transmitted. Nonexcludable and Nonrival Analytical knowledge – Scientific knowledge, formal models – R&D, universities – Codifiable – Ex. Biotechnology 137
Tacit Knowledge: • • • • • •
Specific capabilities of individuals. Face-to-face interactions. Not easily transmitted with distance. Excludable and Rival. More important for innovation. Synthetic knowledge: – Combination, Aplication of existing knowledge – Practical skills, problem solving – Ex. Engineering 138
Geography Matters! • Tacit knowledge favors clustering. • Production is already clustered, but innovation is even more. • But we will take about that later ...
139
What is innovation? • Different than invention – just the first step. • Commercial acceptance. • Innovation is more clustered than invention – why?
140
Innovation in a Global Perspective • Changes that are new to the local context, even if the contribution to the global knowledge frontier is negligible = Innovation! • No reason to believe that their cumulative social and economic impact is smaller.
141
Measuring Technology • R&D statistics – Poor countries don’t invest! – Returns vary – Public X Private
• Patent statistics – Only success stories
• Productivity – Total factor productivity (the A!) 142
And Yet, the Diffusion Process • Adoption of new technology, substitution of the old one. • Without diffusion, no social impact. • For Schumpeter:
Invention
Innovation
Diffusion
143
The Decision to Adopt or Not a Technology • The decision is usually not to adopt or not to adopt, but to adopt now or later. • Diffusion rates are slow and acceptance rates vary for each invention. • Expectations on technology duration are important.
144
Valuation • Benefits are a lifetime flow. • Sunk Costs: incurred at the moment of adotion. • Rarely a new technology is abandoned in favor of an old one (after adoption, costs don’t matter anymore). • Tendency to delay adotion due to sunk costs and uncertainty regarding benefits. 145
S-Shaped Curve: Rogers (1962)
146
Reasons for it • Heterogeneous consumers – different expectations regarding benefits. • Assimetric information and learning process – publicity.
147
Difusion in the US
148
Determinant Factors • Relative advantage of the invention. • Compatibility with current way to do things and current social norms. • Innovation’s complexity. • Possibility of testing. • Possibility of evaluation once adopted. • Network effects. • Publicity. 149
Geography Matters: • Certain Locations are more dynamic than others: – Silicon Valley today – Italy in 1400s, 1500s – England in Industrial Revolution
• Production is Clustered. • Life Cycle theory: – Early stages: importance of technology and tacit knowledge – Mature industries: Costs are more important, so dispersion 150
But Innovation is Even More: • High tech sectors are not that sensitive to costs. • Reasons to cluster: – Infrastructure – Presence of skilled labor – Externalities due to proximity of knowledge base
151
How does it Matter? • Interactions and knowledge flows between: Firms (Customers, Suppliers, Competitiors)
Public Agencies (Tech Transfer Centers, Development Agencies)
Research Organizations (Universities, Think Tanks) 152
Clustering: • Silicon Valley, Wall Street, Bangalore ... • Social assets, that exist between rather than within firms - not fully appropriable by individual firms, only local firms can enjoy benefits. • Basic commonalities: same language, common codes of communication and shared conventions and norms that have been fostered by a shared institutional environment; personal knowledge of each other based on a past history of successful collaboration or informal interaction. 153
Local Networks: • Patents: Jaffe et al (1993): – Patent applications in analytically based industries cite other patents originating in the same city more frequently than non-locally. – Temporal dimension – they tend to be more likely to be localized in the first year following the establishment of the patent, with the effect fading over time (diffusion).
• Star Scientists: Zucker, Darby (1996), Zucker, Darby and Armstrong (1998) and Zucker, Darby and Brewer, 1998) the rates of startup of new biotech firms are significantly higher in those regions in which key scientist work. 154
Codifiable Knowledge is not Enough! • Not all results are codifiable – failed experiments are not! • Spillovers occur first within established local social networks of scientist, often by word of mouth, before formal results are published. • Career-Buzz: Great Minds Think Alike (And Want to Live Together!)
155
What about Globalization? • Reduction in Communication and Transport Costs. • Death of Distance – corporations with no home country affiliation and national identity? • Restructuring of Industrial settings. • But what about High Tech Sectors? • Does this explain the lack of convergence? Will it decrease in the future? 156
Tech Start Ups
Source: The Economist 157
Tech Start Ups per person
Source: The Economist
158
Why Geography Matters? • Tacit Knowledge!! – the importance of previous links (Bercovitz and Feldman, 2010) • The Role of National Economies and Sovereign Nations. • Bercovitz and Feldman (2010): The majority of external members had some prior social relationship via internal members of the team either as former colleagues or students or as long-time coauthors – reflections of the footprints of a former geography. 159
It matters for innovation but what about diffusion? • Although we see a growing contribution of countries such as China and India, the production of technology is still in the hands of a few. • Therefore, technological change pattern is determined, by and large, by international technological diffusion. • This affects growth and income distribution – convergence affected by whether diffusion os local or global. 160
General Data about Diffusion Among Countries • Comin and Hobijn: Historical Cross Country Technology Adoption Dataset, 159 countries, 200 years. • Great dispersion among countries in the diffusion process. This dispersion is 3-5 times greater than dispersion in per capita income. • Relative position of countries regarding adoption of technologies is correlated among technologies. • Convergence for a technology happens at a speed of 4-7% a year. This rate is 3 times greater after 1925. 161
Capabilities • Performance of a jurisdiction heavily depends on what type and amount of knowledge will be produced internally. • Individual productivity is influenced by location: individuals with a given set of characteristic will have different levels of productivity depending on their location (Rigby and Essletzbichler, 2002) 162
Technological Congruence • Degree to which leader and follower country characteristics are congruent in areas such as market size, factor supply. • Ex. US in the 19th century (Abramovitz)
163
Social capability • Vague: – Technical competence – Experience in the organization and management of large scale enterprises – Financial institutions and markets capable of mobilizing capital on a large scale – Honesty and trust – The stability of government and its effectiveness in defining (enforcing) rules and supporting economic growth
• Western Europe in 20th Century: technological congruence + social capability 164
Absorptive capacity: • The ability of a developing country to absorb new investments more generally, related to the ability to absorb knowledge. • Depends on the availability of capital and skilled workers. • Cohen and Levinthal (1990) applied to the firm level: “the ability of a firm to recognize the value of new external information, assimilate it and apply it to commercial ends.” • 3 phases: search, assimilation and commercial application 165
Technological capability • The ability to make effective use of technological knowledge in efforts to assimilate, use, adapt, and change existing technologies. • Possibility to create new technologies and to develop new products and process. • Includes organized R&D and other capabilities needed for commercial exploitation of technology. • Ex. Korea (KIM : 1980, 1997) how Korean electronic firms gradually upgraded from a passive role of implementing imported technology to a more active role of introducing incremental improvements, and eventually ventured into the forefront of innovation based competition. Implementation
Assimilation
Improvement 166
Aspects of technological capability • Production Capabilities: to operate productive facilities efficiently and to adapt production to changing market circumstances • Investment Capabilities: to establish new productive facilities and adjust project designs to suit the circumstances of the investment • Innovation Capabilities: to create new tech, develop new products or services that better meet the specific requirements of the market. 167
How to Measure – Social Capability: • Technical Competence: – Scientific Publications, Patents, R&D, innovation counts – Educational statistics
• Experience in management: – Managerial and technical skills – Previous Experience
• Financial institutions: – Credit
• Honesty and trust; government stability: – Corruption, property rights, independence of courts, bureaucracy
168
How to Measure- Absorptive capacity: • Production quality (ISO 9000) • Infrastructure: – telecommunications, computers, internet; – transport; energy.
• Openess: – FDI – Immigration – Trade 169
What About Results? • Assumption: indicators referring to the same dimension are likely to be strong correlated. • Fabergerg and Srholec (2008): 115 countries and 25 indicators between 1992 and 2004. Led to the selection of 4 principal factors jointly explaining about three quarters of the total variance of the set of indicators. • Most important – technological capability, such as patenting, scientific publications, ICT infrastructure, ISO 9000 certifications and access to finance. • Correlated high with education – cut across tech and social capabilities. • Interpreted it as a synthetic measure of the capabilities or factors influencing the development, diffusion and use of innovations.
170
GDP per Capita X Innovation System:
Source: Fagerberg et al, 2010
171
What is Innovation in Developing Countries: • Diffusion of technology developed elsewhere. • Adaptation of technology. • Such incremental or minor innovations consist of context-specific improvements along the prevailing tech trajectories. However, although minor in a tech sense, these improvements can have major economic significance.
172
How is it done? • More than R&D - typically concentrated in the department of maintenance, engineering, or quality control • R&D is not unimportant – assimilating foreign tech, especially at more advance stages of development. • Correlation between innovative activity and the level of development: the higher the level of GDP per capita, the higher the share of firms reporting to take part in innovation activities. • In the EU, quite a widely dispersed phenomenon, not limited to the most developed parts of the EU. 173
Mobility of Inventors:
Source: Wipo
174
Inventor Immigration:
Source: Wipo 175
Foreign Presence: • While the building of national capabilities may be the aspect of catching up that is most directly influenced by domestic policy, the process also depends on foreign sources of knowledge and tech. • Channels for knowledge: migration, licensing, trade and FDI.
176
Multinationals: • Responsible for a large part of R&D in the private sector in the developed part of the world. • Knowledge is a source of competitive advantage, it can be exploited in foreign markets. • Multinational firms are among the most important technology producers in the world. 177
Multinationals Channels of Influence: • Channels: – Foreign Direct Investment (FDI) – Trade – Licensing – Cross-patenting activities – International technological and scientific collaborations – Demonstration Effects 178
The Role of Multinationals: • 83% of all manufacturing R&D in the US was conducted by parent companies of US multinationals – they typically want to transfer technology they have created to their affiliates abroad. • Firms that engage in international trade and FDI tend to be larger and more productive than firms that only operate domestically (interact with foreign firms). 179
Multinationals and Innovation: • More than 1/3 of the top 100 multinationals are active in more R&D intensive industries such as electronics and electrical equipment, pharmaceuticals, chemicals. • Large multinationals play a dominant role in the innovative activities of their home countries.
180
R&D Expenditure of Foreign Affiliates as a Percentage of Total R&D expenditure of all firms in selected host economies (1998): Country
Percentage of R&D
Canada Finalnd (1999) France Japan Netherlands Spain (1999) UK (1999) US Czech Republic (1999) Hungary India (1994) Turkey
34.2 14.9 16.4 1.7 21.8 32.8 31.2 14.9 6.4 78.5 1.6 10.1
Source: Narula and Zanfei, 2005
181
Internationalization of Research: • Overall, multinationals have increasingly internationalized their innovative activities, and, with a few relevant exceptions (Japan) the importance of R&D activities of foreign affiliates has grown in most host economies over the 1990s. • Nevertheless, most R&D and patenting activities are still largely concentrated in the multinationals home countries, and in a few host countries. 182
Why Internationalize Research? • Firms internationalize their R&D to improve the way in which existing assets are utilized. • Firms seek to promote the use of their tech assets in conjunction with or in response to, specific foreign locational conditions - assetexploiting R&D (Dunning and Narula 1995) or home-based exploiting activity (Kuemmerle 1996). 183
Asset-Exploiting: • Product life cycle theory. • Vernon (1966) and others suggested that foreign subsidiaries replicate the parent’s non-strategic activities abroad with decisions such as R&D and innovation rigidly centralized in the home country. • Vernon emphasized that coordinating international innovative activities was too costly due to the difficulties of collecting and controlling relevant information across national borders. • The R&D activities of foreign subsidiaries were limited largely to the adoption and diffusion of centrally created tech. 184
Asset-Augmenting Activities: • Dunning and Narula, 1995. • The foreign location provides access to complementary location-specific advantages that are not available in its primary or home base. • In many cases, the strategic assets sought by the investing firm are associated with the presence of other firms. • The foundations of competitive advantage no longer reside in any one country but in many. New ideas and products may come up in many different countries and later be exploited on a global scale (Hedlund 1986). 185
Not So Easy … • Developing and maintaining strong linkages with external networks of local counterparts is expensive and time consuming, but once developed, are less costly to maintain. • Firms are constrained by resource limitations and some minimum threshold size of R&D activities exist in every distinct location (critical mass of research). • Difficulties of managing cross border R&D activities. • Many small firms operate as part of a domestic supplier network for larger firms and are thus also bound to their home location or the location of their main customers. 186
And now, back to distance ...
187
What is Distance? • At first, geography. • But it may be any kind of cost ... • In a way the model is capturing costs of transactions: – Language – Customs – Institutions – Bureaucracy 188
Keller and Yeaple (2008, 2009) • Each country has a large number of firms, unique variety of a differentiated final good. • A firm sells its final good to foreign consumers by assembling a range of intermediate inputs, produced at home or abroad. • The costs of technological transfer come in form of communication costs: in the process of communicating the knowledge behind each task from multinational parent to affiliate, errors can occur which make affiliate less efficient than parent production. 189
Understanding Costs: • More complex tasks involve higher costs of transferring the technological information need for offshore (affiliate) production. • Others are easier to transfer because the codifiability is higher or the degree of standardization is higher. • While the tasks can be completed by the multinational affiliate, alternatively the tasks can also be performed by the multinational parent in the home country, after the assemble intermediate good is exported to the affiliate subjected to shipping costs.
190
The Tradeoff • Two types of costs: – Transfer costs – Shipping costs
• When to ship the product and when to transfer the tech and produce abroad??
191
Considering Costs Still • Inputs with high technology transfer costs -> produced at home and exported • Inputs with low transfer costs -> produced abroad. • Both shipping and transfer costs increase with distance. • But at the margin, both costs have to equalize! Shipping Costs
Transfer Costs 192
A Gravity Pull: • If S < T -> export. • If S > T -> transfer tech, and produce locally. • If S = T – an equilibrium – For close countries, both S and T are small. – For faraway countries, both S and T are large. – Model predicts that technology complexity of inputs imported increases with distance. The further away the country and the more complex the good – transfer costs increase more rapidly than shipping costs for these goods. 193
But that is not all .. • Costs increase with distance. • Firms will be forced to set higher prices in order to break even, and that reduces their sales. • The result is the so called gravity pattern for affiliate sales. • FDI falling in geographic distance that finds strong empirical support in the data. 194
Evidence: • Keller and Yeaple (2009): the gravity for multinational sales is stronger for relatively complex goods, and the share of affiliate imports tend to be high for complex goods because technology for those is difficult to transfer. • Trade becomes on average more technologically complex as trade costs between exporting and importing countries rise.
195
Complexity of Exports and Trade Costs:
196
Geographic Pull Through Time:
197
Results: • Period of 1970s and early 1980s: – At a distance of 1000km away from the technology sender, on average only about 20% of the tech knowledge was still available. – At 2000 km, this fraction had fallen to about 5% and once the distance between the tech sender and recipient had reached 4000 km, there was virtually no tech knowledge left.
• Period of 1986-95: – the fraction of technology knowledge available is about 70% at 1000 km, 50% at 2000 km and still 25% at 4000 km. • Strong evidence that the information and communication technology innovations mentioned above led to more international diffusion of tech over time.
198
How does technology diffuse? • Market Transactions. • Externalities – Spillovers. • Channels: – Trade (interaction) – FDI – Import – Export
199
Spillovers • All this Internationalization and extra contact must have some influence in domestic production chains. • Externalities of Technology Diffusion – Spillovers!
200
Externalities • Firms make royalty payments for their use of patents, licenses and copyrights and this is recorded for many countries in services balances. • But tech transfers occur not only through market transactions but also as externalities as tech is not fully codifiable. • Although it is not through market, it is observable through activity – international tech spillover, most important channels are trade and FDI. • Imports are also an important channel of diffusion. Although of less importance, so are exports. • Innovation involve by nature asymmetry of information.
201
Horizontal Vs Vertical Spillovers • Horizontal FDI spillovers: Technological learning spillovers to other firms in the industry through its business operations – effects are which within industry. • The introduction of a multinational makes local competitors copy its process and products. • Because the introduction of a business requires major investments, the size of the externality must be less than the value of the cold chain. • Vertical Spillovers: Across Industries. • If the affiliate buys inputs from local suppliers, there may be so-called vertical backward tech spillover. They arise when a multinational affiliate provides tech to its supplier to a price that is below its market value. 202
But Not Everything is an Tech Externality! • The need to distinguish tech externalities from pecuniary externalities – the latter arise when the arrival of a multinational affiliate induces local firms to produce a greater variety of inputs, which may benefit local downstream producers. The case is distinct because even though there may be productivity gains, there is no transfer of tech. • The entry of an affiliate may affect productivity in ways that is not a tech spillover: it might increase competition, but no tech diffusion is involved. 203
Two channels for FDI spillover: • Affiliates could either disseminate tech to domestic firms of their host country (inward FDI tech spillover) • They might pick up tech from the firms in the host country (outward FDI tech spillover). • Studies show that inward tech spillover are smaller than outward tech spillover.
204
FDI: • Spillovers with FDI are usually important –FDI is important is for countries as a whole and specifically for tech transfer. • But not all case study evidence points to tech spillover associated with FDI – multinationals often do not know the tech employed by their suppliers.
205
Reasons for More Diffusion Inward: • Multinational affiliates are larger and more tech intensive than the average firm in the host country, and this might be the reason why they are good at sourcing tech. • Multinational parent set up the subsidiary with the explicit goal of tech sourcing, while the average host country, in contrast, has not made a comparable location decision. 206
Problems with Measuring Spillovers: • Spillovers measures don’t exist, only proxies. • Capture spillovers only partially, because measures don’t account for cost of acquisition (learning).
207
But How Do We Do It Anyway? • Patents: citation of earlier patent- strongly suggests learning took place but can’t estimate how large these benefits are. • R&D spillover regressions: R&D of firm j is positively correlated with TFP of firm I (spillover from firm j to i)
208
Statistics: • General Statistics of R&D and Patent. • Statistics of R&D and Patents that show influence. • Relationship between productivity and foreign R&D conditional on imports from that foreign country, while some have considered FDI as well. • If the tech capabilities of a firm have improved, can causality be established? Tech capability of the receiving firm is only imperfectly observed (eg its productivity) – which part can be considered a tech spillover? 209
R&D as a Percentage of GNP (World Bank): 3,5
3
2,5 India Brazil
2
China East Asia & Pacific (all income levels)
1,5
European Union Russian Federation United States
1
0,5
0 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011
210
Gross Domestic Expenditure on R&D by sector (% of GDP)
Source: Eurostat
211
Other Forms of Internationalization: • •
•
•
Global inter-firm alliances have become increasingly popular over the past 2 decades (Hagedoorn, 2002) Non-internal R&D activities that rely on interfirm cooperative agreements have grown rapidly in number. In some circumstance international cooperation may substitute for internal innovative activities such as that of R&D alliances aiming to enter foreign markets protect by non-tariff barriers as it is the case of environmental regulations. This tends to develop in areas in which partner companies share complementary capabilities and these alliances create greater degree of interaction between the partners respective paths of learning and innovation. Participation in these alliances tend to be correlated with firm size in tech intensive sectors by associating complementary resources, and competencies, it makes it possible to explore and exploit new tech opportunities.
212
As an After thought
213
