The role of the state in the global economic competition for artifical intelligence

(1)

The Role of the State in the Global Economic

Competition for Artificial Intelligence

Graduate School of Social Sciences

Master Thesis Political Science

Name: Stan Goudsmit

Student number: 10755446 Specialization: Political Economy

Supervisor: J.G.W. Blom Words: 18673 Date: 21-06-2019

(2)

1. Introduction

1.1 Research Question

Over the recent years multiple developed countries around the world have publicly published industrial policies. In 2017 Australia released its newest industrial policy ‘Australia 2030: prosperity through innovation’. In December of that same year the United Kingdom (UK) announced the publication of the ‘Industrial Strategy: Building a Britain fit for the future’. Germany, another economically leading country, followed the examples given and in February of this year the ‘National Industrial Strategy 2030’ was printed. These publications are some of the most modern examples of a trend that has started very recently. In 2010 the Japanese minister of Economy, Trade and Industry announced the creation of a strategy for his country to withstand the emerging of increasingly aggressive industrial policies of the top advanced countries (Akkemik 2015). Industrial policy would be back in fashion with policymakers in high-income countries (Lin 2013). This last decade is therefore characterized by, as what some scholars (see: Warwick 2013 & Owen 2012) have called, the global revival of industrial policy.

Moreover, the last two years another growing phenomenon can be noticed. Several leading countries have created Artificial Intelligence (AI) strategies. Policy makers seem to believe that AI is a crucial factor within the industrial strategy. It’s argued that AI can create a new industrial revolution (Horrowitz et al. 2018). Considering the fast increase of computing process capabilities and data, AI is now at a point that shows high potentials for future life (European Commission (EC) 2018: 19). As Spencer (2017) writes, many believe that between 2025 and 2050 AI will completely change the way of life. These days, countries are preparing for this coming age of AI. Realizing the AI potentials would promise considerable benefits for businesses and economies. The previous two years more than 20 countries in total have published their national AI strategies to gain a competitive advantage (Rao 2019). Canada was the first country ever to publish its fully funded strategy plan in March 2017, named the ‘Pan-Canadian AI Strategy’ (Dutton 2018: 4). Macaulay (2019) shows that by now almost all powerful countries have released an AI strategy. As a result some countries, such as the UK, Australia, and Germany, have an industrial strategy accompanied by a more detailed and in-depth strategy for their national AI sectors.

(4)

This series of publications of national industrial strategies and AI strategies has triggered the debate on the role of the state in the latest phase. From the 1980’s onwards the leading paradigm in the debate on the role of the state argued that too much government involvement was seen as highly undesirable. The taboo that industrial policy has carried in approximately the last four decades can be seen in the title of a working paper published by the IMF this year: ‘The Return of the policy that shall not be named: principles of industrial policy’ (Cherif and Hasanov 2019). Industrial policy, and hence the involvement of the state, had been boycotted from 1980 onwards due to the Washington Consensus (Owen 2012: 32). Since the reign of Reagan and Thatcher there has been a belief that the state should be ‘rolled-back’. The highly used industrial policy by states before 1980 was blamed for picking winners, supporting national champions and subsidizing ‘lame-duck industries’ (Tomlinson 2018). To avoid this government overload, the previous phase was filled with a predominant neoliberal discourse with privatization, deregulated free-markets, and a limited role for the state. Defending industrial policy as an economist was close to being a shameful action (Xinquan and Guijin 2011). A closer look at the front pages of the recent printed industrial strategies shows still somewhat of a taboo on industrial policy. The word ‘policy’ has been replaced by ‘strategy’ to avoid the negative factors associated with the phenomenon in the past.

However, the recently published industrial strategies and AI strategies seem to mark the end of a phase that lacked openly published industrial policies in most Western European countries. The new period looks to publicly challenge the underlying consensus that state planning and public investments can’t be a driving force in the economy. Page and Tarp (2017: 1) argue that a growing recognition of markets’ imperfections and the contribution of the state to economic growth through state interventions and policies keep gaining ground. This period heralds a new chapter of the long-standing debate on what the role of the state should be for economic development. Pemberton (2017) mentions in his article that states have been struggling for decades on how to increase a country’s economic growth. There has been little consensus on what role the state should play, in particular in relation to the private sector. Rodrik (2004) shows how the general belief under economists for economic hardship has

(5)

frequently moved between market failures and government failures. Moreover, the competitiveness of states on the global level due to globalization has fueled the debate on the role of the state within the field of International Political Economy (IPE) since the last decade of the previous century.

One of the main scholars to research the role of the state in this phase of globalization is Cerny. With his Competition State theory Cerny (1997, 2010) tries to explain the changed role of the state since the 1980’s. He argues that developed countries transformed from welfare states to competition states (Cerny 1997). The state had no longer the welfare of its citizens as its core policy, but competitiveness on the global level would be its new main focus (Cerny 2010). The competition state has a limited role to only regulate markets and, only if necessary, fix market failures (Uvalic 2014). Two of its main features were reducing government spending and deregulation of economic activities. In the Competition State theory the entrepreneurial power of the private sector is the key for economic success. However, the theory does acknowledge the existence of industrial policy (Cerny 1997: 266). But as Genschel and Seelkopf (2014: 236) show the industrial policy under a competition state has the facilitation of the private sector through general promotion of innovation and growth as a central focus point.

On the other hand, in 2013 Mazzucato presented her book ‘The Entrepreneurial State’. In her theory she fights the assumption that the state has always taken a backseat to ‘unleash’ the power of entrepreneurship in the private sector (Mazzucato 2013: 15). Moreover, in her argument for the entrepreneurial state there is no revival of industrial policy. Despite the fact that the proactive entrepreneurial role of the state has been hidden from the public, it has always been present (Mazzucato 2013: 21). She points to the argument of Block and Keller (2011) that the industrial directives coming from states have been hidden due to the possible backlash coming from the conservative right parties. Following her argument states have proved to go beyond fixing market failures by taking on a more strategic role in what Cerny argues to be the period of competition states. According to Mazzucato (2013:21), states have taken on an entrepreneurial role by doing crucial investments and creating markets. As a result, Cerny and Mazzucato

(6)

have different views on how to understand the modern role of the state in the latest decades.

With the potential of a new industrial revolution caused by AI and a new phase with publicly published strategies of countries, it seems relevant for the field of IPE to research the current role of the state in this current period. Policy makers look to emphasize the importance of the national AI sector for economic growth of a country in the future through their industrial strategies. To get a better understanding of the role of states within the industrial strategies in relation to the development of the AI sector, this thesis will answer the following research question:

What is the role of the state with relation to the development of the AI sector?

To answer the central question of this thesis, it will analyze to what extent the role of the state fits the characteristics of entrepreneurship. Mazzucato (2011b) argues that entrepreneurship is one of the least understood topics in economics. Starting with a closer look on an entrepreneur, this thesis follows the argument of Mazzucato (2013: 65) that an entrepreneur can be both an individual and a group of people who is willing to convert a new idea or invention into a successful innovation. This means that also a state, which consists of a group of people, can be an entrepreneur. Nonetheless, many scholars have linked entrepreneurship over the years to different characteristics. In the beginning of the previous century Knight (1921) has presented the risk-taking aspect of entrepreneurship. Moreover, Drucker (1985) focused on the innovative change that is created by entrepreneurs. Whereas these two scholars focused on the economic perspectives, Audretsch (2012) emphasizes that entrepreneurship also has management features that should be considered. One of the latest definitions of entrepreneurship that combines these theoretical examinations is the one of the EC (2003: 6). The commission defines entrepreneurship as ‘the mindset and process to create and develop economic activity by blending risk-taking, creativity and/or innovation with sound management, within a new or an existing organization’. Many scholars doing research on entrepreneurship (see: Congregado 2008, Syrett and North 2008 & Welter and Smallbone 2011) highlight this definition provided by the EC. Therefore, this thesis will use the definition to analyze the role of the state in the AI sector

(7)

From this definition of entrepreneurship at least three characteristics can be distinguished: risk-taking, creativity/innovation, and sound management. Based on these three features, this thesis will set up three sub-questions to answer the overall question. The overall research question will be separated into the following sub-questions:

1. What level of investments is the state making?

2. To what extent is the state involved in the innovation process? 3. Who/what is managing the development of the AI sector?

The first sub-question will focus on the risk-taking of the state. It will look at the levels of investments made by the public sector for the development of AI. Investments show the risk taken by a business or state, because they are only expected to produce future profit. Investments always carry a risk due to the simple fact that they can fail. The second part examines the involvement of the state in the innovation process of the AI sector. Although an industrial strategy is always focused on innovation, the attention given to a certain sector can be different. The AI strategy can relatively receive more effort of the state when it’s part of a vertical overall industrial strategy. The difference between a horizontal and vertical industrial strategy will be explained in more detail later in this thesis. Moreover, innovation is connected to Research & Development (R&D). Considering this relation, the second part will analyze the presence of a national AI research institute. The last sub-question is focusing on sound management. This thesis will translate sound management to finding the person/institution that has the power to rightly manage things. In other words, the last part of the analysis is focusing on which persons or institution are given the responsibility to control and oversee the development of the AI sector. The next paragraph will elaborate on the structure of this thesis.

1.2 Structure

This thesis will research the role of the state in relation to the development of the AI sector based on an analysis of the industrial strategy and AI strategy documentation of the UK. It will argue that the state takes on a more entrepreneurial role in the strategy to

(8)

develop the national AI sector. The role of the UK state in the national AI strategy shows more similarities with an entrepreneurial state in both the ‘risk-taking’ and ‘innovation’ feature. Therefore, the theory of Mazzucato (2013) provides a better general understanding of the current role of the state in the national AI strategy. However, the findings in relation to the ‘sound management’ feature show that the UK state holds on to a belief of the power and knowledge of the private sector. In the case of the UK, the private sector is in the position to lead the development of the national AI sector. As a result, the theory of Cerny (1997, 2010) still provides an important insight on the role of the state. Nonetheless, the state is far more active than argued in Cerny’s Competition State theory. Moreover, as a result, this thesis will support the argument of Rodrik (2004) that modern industrial policy should be reconsidered. It’s no longer about an autonomous state applying subsidies and taxes, but one that creates a more entrepreneurial collaboration between state and business.

The next paragraph will elaborate on the relevance of this thesis. After, paragraph 1.4 will explain why the UK has been chosen for this study. Chapter two will present the theoretical framework. It will start with a wider view on the state debate to position the discussion between Cerny and Mazzucato within the broader state debate of IPE. This part will be followed with a more detailed and in-depth analysis of the Competition State theory and Entrepreneurial State theory. The review of these theories will be translated into expectations that are linked to the three sub-questions and entrepreneurial features to help answering the sub-questions. The last two paragraphs of the theoretical framework will provide conceptualization on the subjects of industrial policy and AI. This will provide a better understanding of the case used for this research. Chapter three will introduce the methodology. It will show that this thesis is doing a qualitative research through a case study and document analysis. The paragraph of the methods will be followed by a paragraph that shows the operationalization used for the empirical research. The empirical research will be presented in the chapter four. Its structure will follow the line-up of the questions. Therefore, it will start by answering the sub-question on the feature risk-taking, followed by relatively innovation and sound management. All the findings used from the document analysis can be found in the appendix. The thesis will end with a conclusion and discussion.

(9)

1.3 Relevance

This thesis is a valuable addition to the debate on the role of the state in this current period that is dominated by the great economic potentials AI can create for countries’ economies. It will argue that the state takes mainly the form of an entrepreneurial state on to grasp the competitive advantage of AI for its country. This thesis will show that within the current premature AI period the theory of Mazzucato (2013) gives the better general understanding of the role of the state. As a result, this thesis has the strength to adjudicate the academic debate between Cerny and Mazzucato on the role of the state nowadays. With the consideration of the global competition of states for AI, this thesis presents the argument that developed states do not leave the realization of economic growth and development to the private sector and the market in this period. States are proactively playing a role in the collaboration with the private sector. As a result, this thesis illustrates how the paradigm of neoliberal states is challenged.

Moreover, this research will give a new view on the effect of AI by approaching it not from a technological angle but a political economic one. It will stress the urgency states feel for the development of their national AI sectors. With this focus of the thesis it will show why it’s important for IPE researchers to start analyzing the role states are taking in this untimely phase with the potential of a next industrial revolution. Therefore, this research will be a contribution to the research on the current status of the role of the state in the field of IPE due to a new technological break-through. So far the role of the state nowadays in relation with AI has been unexplored within the field of IPE.

Furthermore, this paper will contribute to the existing literature on industrial policy. As Owen (2012) argues, states have arrived in a phase in which they publicly present their strategies again. This research will strengthen the argument of Rodrik (2004), which states that to get rid of the taboo on industrial policy it should be reconsidered. For, as he calls it, the 21’st century industrial policy to be accepted, it shouldn’t any longer be seen as an instrument of an autonomous state to just complement the markets, or as opponents of industrial policy would have argued ‘to distort’. The current industrial policy is, according to Rodrik (2004), different from the ones used by states before 1980. States have presented industrial policy to give the public sector the possibility to truly contribute to economic growth. The assessment of the modern industrial policy in this

(10)

research gives a better understanding of the role of the state in the recently published industrial strategies in relation to AI.

1.4 Case Selection

Many scholars (Cerny & Evans 1999, Horsfall 2013, Evans 2010 & Forde & Slater 2016) point at the UK as a clear example of a state that has transformed into a competition state since the reign of Thatcher. The principals of the competition state, such as de-regulation, have been embraced through time by the consecutive UK governments. Although there are differences visible to what extent competition state policy have been implemented under the Labour and Conservative parties’ reign periods, Evans (2010) argues in his article that the ‘character’ of the state has generally changed from a welfare state to a competition state. The attack of Cameron in 2011 in which he called the civil servants in government ‘the enemies of the enterprise’ (Guardian 2011) shows that the rationale of the competition state under leading UK politicians was still alive in the beginning of this decade. Following his argument the public sector was still too much holding back the real entrepreneurs in the private sector. This thesis therefore uses the UK government to be able to do research on a state that shows one of the highest correlations with the Competition State described in the theory of Cerny (Horsfall 2013). Moreover, with this choice this paper takes a different leading country than Mazzucato. Mazzucato (2013) uses in her book mainly the USA, which would according to Horsfall (2013) also be an orthodox competition state, as prime example to show the entrepreneurial role of states. The choice of the UK in this research will strengthen the generalization of the argument of the Entrepreneurial State theory in relation to AI. Furthermore, the AI sector is not only relevant for its great future economic potentials but also provides this research with a space of the economy in which the debate can be correctly adjudicated. In chapter 2 of her book Mazzucato (2013) attacks the argument that the state is limited to creating the conditions for technological innovation. The Entrepreneurial State theory claims that the innovation in a new high-tech sector is in great need of state actions due to the extensive risks involved. Mazzucato (2011b) states that the private sector is too risk-averse to get the innovation going in such a new high-tech sector. The public sector therefore takes up the role as entrepreneur through being proactive in a highly uncertain early stage technology development (Mazzucato 2013).

(11)

According to the Entrepreneurial State theory, these sectors with development of new technology should be the optimal place to see the state in its entrepreneurial role. Therefore, AI provides the strength of analyzing an ‘entrepreneurial sector’ in relation to a competition state.

To sum up, the role of the UK state within the national AI sector strengthens this research through the combination of both a new high-tech sector emphasized in the Entrepreneurial State theory and a country highlighted in the Competition State theory. This research therefore analyses an appropriate case through a theoretically balanced approach.

2. Theoretical Framework

This chapter discusses the theoretical perspectives used for this study. It will start with reviewing the broader state debate within the field of IPE. It will argue that the theories of Cerny and Mazzucato are connected within this broader state debate through their focus on countries’ strategies to be globally competitive. Therefore, paragraph 2.1.2 will focus on the literature on global economic competition between states. After, the Competition State theory and Entrepreneurial State theory will be examined in more detail. The last part of paragraph 2.2 will compare the two theories along the three features of entrepreneurship presented in the introduction, and create expectations for the empirical analysis. The last two paragraphs will conceptualize the case used in this study by reviewing the literature on industrial policy and AI. This should provide a better understanding of the case in this research.

2.1 Views on the State

This part of the chapter will first show the discussion on the broader state debate within IPE. It will illustrate where to place the debate between Cerny and Mazzucato in this greater debate by arguing that the two theories are connected through their predominant focus on global competitiveness between states. The second paragraph will therefore go into more detail on the literature on the global economic competition. By doing that it will consider the shared underlying principles of the Competition and Entrepreneurial State theory. This will provide a better understanding of the theories

(12)

within the wider context on the role of the state debate during this phase of globalization.

2.1.1 The Globalization-State Debate

Although this debate is more focused on the prospects of the state with recognized theories such as the ‘Retreat of the state’ of Strange (1996) and the ‘Myth of the powerless state’ of Weiss (1997), reviewing this greater state debate does position the discussion between Cerny and Mazzucato within the broader state debate. It provides an insight into the shared underlying principle of the two theories examined in this research.

Globalization is probably one of the most debated concepts across the academic world. Bertucci and Alberti (2001), among many others, call it a multifaceted phenomenon. Since it’s such a complex phenomenon that has different effects on multiple spheres, it’s argued to be nearly impossible to give globalization a unique definition. However, many definitions given within the academic literature refer to the rise of new technologies, a vast increasing and intensifying interconnectedness, and necessary new thoughts on social, cultural, political, and economic factors. Nester (2010) presents this phase of globalization as a world that is filled with complexities and interdependencies in multiple spheres ranging from economic and cultural to environmental and political. The process of globalization means that states, but also firms and individuals, are more interconnected than ever before. Although globalization seems to be too big and complex to define, most of the broad definitions use the increasing interconnectedness that alter way of life due to new technologies to describe globalization.

A broad and long-standing discussion within the field of IPE about the state and the phenomenon of globalization is the globalization-state debate. Held et al. (1999) usefully divide the globalization debate in three schools of thought: hyperglobalists, skeptics, and transformationalists. Hyperglobalists offer the most extreme view on the power of globalization. They argue that globalization will make the state diminish (see: Ohmae 1990). The school states that the global development will lead to a post-state-centric system (Myint 2011: 39). The state will become powerless due to the unstoppable move of events from the national to the global scale (Nester 2010). On the other side, there is the school of the skeptics. Their view offers a sharp contrast with the hyperglobalists.

(13)

According to the skeptical claim, globalization is overstated and the state remains essentially the same (Fougner 2006). Instead of the diminishing of the state, it would remain the central institution. The last group in this debate is the one of transformationalists. They place themselves in between the two other groups. In their belief globalization is an evolutionary process, but states and other institutions have the ability to adapt to it (Held et al. 1999). In the transformational point of view globalization will change the functions of the state (Myint 2011: 402). These three schools provide a broad division of views on the state during this phase of globalization. Considering the focus on the increased global competitiveness of countries due to globalization puts the Competition State within the transformationalist school of thought. Cerny’s (1997, 2010) Competition State theory is developed around the argument that the role of states have changed to make society fit for competition on the global scale. Although the Entrepreneurial State theory doesn’t pay significant attention to the phenomenon of globalization as a cause for the entrepreneurial role of the state, the examples in her argument, such as apple, show the central role of the US state in the global economic competition (Mazzucato 2013: 95). As Pradella (2016) also argues, the overall argument in the Entrepreneurial State theory demands the acknowledgement of the central role states have in the struggle for global competitiveness.

This focus on competitiveness of states can’t be a high priority within a hyperglobalist debate. The lack of an ‘inter-nationalist’ view makes the focus of states on competitiveness irrelevant due to a single market in a borderless world (Fougner 2006). Therefore, the theories of Cerny and Mazzucato are in stark contrast with the hyperglobalist argument of a diminishing state. This puts Mazzucato (2013) between the transformationalist and skeptical school of thought. She supports Kogut’s (2003) argument that competitive advantages of countries still depend on national innovation systems. Notwithstanding the important entrepreneurial role that Mazzucato (2013) gives to the state, she does not elaborate on the matter whether this role is a result of globalization. The examples used in the Entrepreneurial State Theory, such as Apple, the Green Industrial Revolution, and wind and solar power, are all from the last three decades. Although this does not argue that the state has changed during globalization, it does imply the prominent entrepreneurial role of the state in the competitive strategies

(14)

of countries in this phase of globalization. Mazzucato’s (2013) theory is therefore situated somewhere between the transformationalist and skeptic school of thought. This part has showed the different schools of thought in the globalization-state debate. It has showed that the theories of Cerny and Mazzucato don’t fit into the hyperglobalist view due to the underlying rationale of global competitiveness of states in their theories. Whereas the Competition State theory suits perfectly into the transformationalist school of thought, the Entrepreneurial State theory shows not a clear distinction in its argument between the transformational and skeptical claim. Nonetheless, the focus on global competitiveness is what connects the two theories within the broader state debate. The next paragraph will further elaborate on the global economic competition that both theories emphasize.

2.1.2 The Global Economic Competition

This part will review the literature on global economic competition between states. Although the warning of Krugman (1994) against the focus of policy makers on a country’s global competitiveness, it will argue that it’s a central theme in current policymaking of states. With this elaboration on global economic competitiveness the shared underlying rationale of both theories will be reviewed.

25 years ago Krugman (1994) argued that the term competitiveness was useless. Concerns about global competitiveness would only be a dangerous obsession for policy makers. He tried to fight the assumption that the economic power of a country could be determined by its success on the global market. In his argument states are not in any important way competing with each other. This false acquisition, according to Krugman (1994), would derive from the wrong comparison of states with corporations. As a result, the obsession of policymakers on competitiveness would lead to bad decision-making (Krugman 1994: 30). However, as Bresser-Pereira (2008) argues, the argument that competition of states shouldn’t be compared with the competition between corporations does not mean in any way that states do not economically compete. This thesis will argue that Krugman’s (1994) argument does not consider the different economic logic of competition between corporations and states. States focus on higher growth rates instead of higher profit rates. This different logic of economic growth

(15)

through high global competitiveness power of country is a prevalent theme of policy making these days.

Stopford and Strange (1991) were one of the first scholars to argue how the competition between states has changed from territorial disputes to rivalry over global market shares. This economic competition between states has intensified during the phase of globalization (Camagani 2008). This manifests itself through increasing international trade, foreign direct investments (FDI), and the global spread of information, knowhow and technology (Feenstra and Gordon 1996). Nowadays most of the implemented policies by a state will not only influence domestic actors (Myint 2011). A policy implemented in one country can have an effect on another country’s relative position in the global economy. Kharlamova and Vertelieva (2013) claim that every state strives for a competitive global position to be an influential and important player in the global economy. Following the mercantilist rationale, a nation’s competitiveness will stimulate continuous economic growth of a country (Schot and Steinmueller 2018). However, Kharlamova and Vertelieva (2013) don’t only focus on the competition of states for economic growth but also for the risks the global economy entails. The economic crisis, according to Kharlamove and Vertieleva (2013: 39), has shown that the countries with the lowest global competitiveness were the most vulnerable to external shocks. Therefore, competition between states is not only crucial for economic growth, but also as protection for negative externalities of the global economy.

In Europe, promoting competitiveness by states has become a more prominent theme in politics since the treaty of Lisbon in 2009 that partly focuses on making the European region the most competitive economic region in the world (Vukov 2016). Furthermore, both Kozmetsky and Yue (1997) and Mazzucato (2013) try to show how the importance of global competitiveness has played a significant role in the policy of the US. Fougner (2006) argues that it’s fair to say that many, if not all, states are geared to global competitiveness these days. Innovation and technology are aimed at maintaining or increasing a country’s competitiveness. This increased focus shows the belief of policy makers in a country’s global competitiveness nowadays.

(16)

Paragraph 2.1.2 has reviewed the literature on the global economic competition between states. It illustrated how the importance of global competitiveness for countries’ economic growth has become a central theme in policymaking. A high level of competitiveness will provide a country economic growth, and protect it against negative external shocks from the global economy. This focus on a country’s global competitiveness connects the two theories reviewed in this thesis. However, the next part will show that Cerny and Mazzucato disagree on the role of the state in a country’s strategy to be globally competitive.

2.2 The Competition State and Entrepreneurial State

Whereas the previous part has showed the broader context around the debate between Cerny and Mazzucato, this part will have a more detailed and in-depth review on the two theories. It will start with the Competition State theory of Cerny (1997, 2010), followed by the Entrepreneurial State theory of Mazzucato (2013). This paragraph will finish with a short concluding overview comparing the two theories, and creating expectations linked to the theories and sub-questions.

2.2.1 The Competition State

The Competition State theory is developed at the end of the previous century. According to the theory, advanced Western states, and to a lesser extent countries worldwide, would have undergone a thorough transformation. Driven by political and economic globalization the traditional form of the state was forced to change (Cerny 1997). Cerny (1997, 2010) argues that the Competition State is the historical successor of the welfare state. The competition state got the opportunity to flourish after the breakdown of Keynesianism caused by the crisis of the 1970’s (Horsfall 2013). It gave scholars such as Cerny (1997, 2010), but also Evans (2010), the chance to alter the paradigm of a big welfare state. The Competition State theory claims that the overall direction of the change of states is clearly in the direction of the competition state due to globalization (Cerny 2010: 18).

The shift from the welfare state to the competition state has different implications. The commodification process is the underlying driving force of states into their new competition forms (Cerny 1997: 266). The state has played an important role in the strategy for global competitiveness through liberalization, deregulation and

(17)

privatization. These types of state interventions in competition states are therefore more focused on marketization than the decommodifying interventions of the welfare state. This move of the state has increased the dependence of citizens and industry on the global market. Competition states open up to the global market, because it’s believed that domestic sources of inputs are to small to be economically efficient (Cerny 2010: 17). As a result the, competition states rely significantly on market mechanisms.

This pro-market policy of the competition state should free the way for the private sector to make a country globally competitive. Businesses are the key players in competition states to globally compete (Lee 2008: 28). Only through a limited state there will be no ‘crowding out’ effect of private investments (Cerny 1997: 259). The US would for example heavily have relied on market forces to give private firms the incentives to conduct R&D (Katz and Ordover 1999: 137). Genschel and Seelkopf (2015: 238) see this decrease of public investments as an adaptive response to the government overload failure of the welfare state. This overloaded state would have soaked up resources which otherwise might have been available for private capital allocation, would have raised the cost of capital, and channeled resources into non-productive directions (Cerny 1997: 262). Therefore, the Competition State theory argues that the state should conduct a policy with minimal market interventions. It should, according to Cerny and Evans (1999: 22), only intervene by incentivizing entrepreneurial behavior of private enterprises through for example the national tax system.

Despite the fact that the Competition State argues that market interventions of the state should be limited, Cerny (1997) still argues that a hidden industrial policy is actually alive and well too. The industrial policy of a competition state would promote innovation and growth in general (Genschel & Seelkopf 2015: 236). It does not try to ‘pick winners’ in the economy through an activist industrial policy. The state does play a role, but this is a facilitating one. It should only involve itself in funding education, science, basic R&D, and stick to promoting competitiveness for the private sector (Cerny & Evans 1999). It believes that the market has the ability of creating winners, and that the state should not be in the way of letting the real innovators and entrepreneurs in the private sector do their job.

(18)

Nonetheless, the competition state thesis has been subject to criticism. Hay (2004: 41) mentions in his article that the Competition State argument should be seen as work in progress because it still lacks explanatory power of the current transformation of the state. The theory is a rather descriptive theory of the transformation. Moreover, the arguments’ simplicity and generality are simultaneously offering a great explanatory potential and problem. Genschel and Seelkopf (2015) summarize the critics by outlining the three principals central to the competition state theory: structural determinism, convergence, and the decline of the welfare state. Hay (2004), for example, disagrees with the existence of strong structural determinism in national policies and institutions. Moreover, the second principal ‘convergence’ is being seen as too simplistic and generalizing. Cerny (1997, 2010) assumes with his theory that there is only one best policy option of liberalizing. This would result in a general trend of states transforming into a competition state. Hall and Soskice (2001) use the ‘Varieties of Capitalism’ argument to argue the opposite, namely divergence. They argue that developed economies of capitalist countries are very varied between coordinated and liberalized market economies. As a result, the Competition State theory has often been blamed to be just a narrative of the Thatcher and Blair UK governments (Horsfall 2013: 55). The last principle is the one in which Cerny argues that the welfare state is a disadvantage for the competitiveness of a state in the global markets. Luiz (2000), for example, writes in his article about the ‘irony’ of the free market that would be in need of a stronger state. Moreover, the compensation thesis states that economic openness and public spending share a positive relationship (Busemeyer 2009: 456). As one factor goes up, the other is following.

Furthermore, Reviewing the literature on the competition state, there has been very little empirical research on the competition state. Vukov (2016) and Horsfall (2013), however, attempted to link the theory to empirical evidence, but the results were not convincing for the support of a general move of countries into a competition state. It, therefore, misses convincing empirical arguments to strengthen the theory.

This paragraph has reviewed the Competition State of Cerny. Its main argument is the emergence of a new state form since the early 1980’s due to globalization. These changes are all aimed on making the state more competitive internationally. Three main

(19)

features of the competition state are liberalization, deregulation and privatization. With this pro-market policy the state has the ability to facilitate the entrepreneurial power of the private sector to make a country globally competitive. The next paragraph will present the Entrepreneurial State theory of Mazzucato (2013).

2.2.2 The Entrepreneurial State

In 2013 Mazzucato published her book titled ‘The Entrepreneurial State’. In her book, she fights the assumption that a country can only be globally competitive, and hence increase its economic growth, by the entrepreneurship and innovative power of the private sector. Mazzucato (2013: 22) tries to change the dominant picture of a sluggish, bureaucratic, and slow public sector. Much of the innovation in the past has been realized by the ‘visionary’ strategic interventions of the state (Mazzucato 2013: 29). A state, according to Mazzucato (2011b), should do more than ensuring a macroeconomy and infrastructure that is favorable for the private sector. Following the argument of the Entrepreneurial State theory a far more proactive state is required for economic growth. The Entrepreneurial State theory shows the importance of public investments in innovation. The state in its entrepreneurial role is the leading agent of succeeding in technological progress. The achievement of technological breakthroughs creates the innovation that allows companies and even the entire national economy to grow (Mazzucato 2011b). Therefore, Mazzucato’s (2013) entrepreneurial role goes beyond the economists that saw a role for the state to fix market failures. With emphasizing the crucial central and active role of the government the Entrepreneurial State combines the Keynesian and Schumpeterian state (Mazzucato 2011a). The theory claims that states have been leading takers and market shapers. The state has been the crucial risk-taker by funding the most uncertain phases of innovation processes, because the private sector has proved to be too risk-averse (Mazzucato 2011b). Within the R&D process states have played a role not only in the beginning and basic R&D, but was also present in the commercial viability stage of the innovation process (Mazzucato 2011b: 41). This results from the fact that many small firms actually ‘die’ between the invention stage and the later commercial transformation and application phase. Herewith Mazzucato (2013: 79) attacks the ‘globally accepted’ assumption that wealth creation has been mainly private sector-led. Whereas Cerny (1997, 2010) used the argument of the ‘crowding out’ effect, Mazzucato (2013: 31) believes in the opposing ‘crowding in’ effect.

(20)

She claims that public and private investments have a positive relationship. Instead of making the private sector less ambitious through public investments, the private sector would follow and boosts it total amount of investments. The public sector should, nonetheless, show more character in demanding more effort of the private sector. A concrete example of a global business success in which the state has played a major role is, according to Mazzucato (2013), the American company Apple Inc. Apple was only able to become such a global player because of the integration of new technologies that were developed through state investments (Mazzucato 2013: 93). Moreover, another American example mentioned to support the theory is Silicon Valley. Mazzucato (2011a: 135-136) argues that the region became globally competitive after decades of strong state-led vision, state investments in most uncertain research, and nurturing of regional innovation systems. With these examples Mazzucato (2013: 3) tries to counter the argument that the state is lousy at picking winners and should stick to the basics such as funding education. The theory, therefore, supports the argument of Rodrik (2004) to rethink public and private actions. The previous technological break-throughs would never have happened without an entrepreneurial state (Mazzucato 2011b). Moreover, Dosi & Lovallo (1998) claim that states shape expectations of future growth, which would be crucial for firm entry. The state has been the agent shaping the future opportunities, often by investing in areas before the private sector dared to invest (Mazzucato 2015). Her argument however does not try to undermine the importance of the private sector. The Entrepreneurial State theory wants scholars to reconsider the role of the government in the government-business relationship. Countries have experienced innovation-led growth through a pro-active entrepreneurial role of the state.

Nonetheless, the theory of the Entrepreneurial state also had to face critiques. Mingardi (2015) shows her doubts about the generalizability of the argument due to the use of only American examples (Pradella 2017). This makes the external validity of the theory questionable. It can be compared with the critique on the Competition State theory of being a narrative of the transformation of the UK state. Moreover, her argument would have been more convincing if she would have included more companies and countries to see how they differ. It would have given her theory a broader support. With a

(21)

comparative design focused on several companies and countries, Mazzucato could have included the differences between companies to strengthen her argument. Since this research is using the UK as case, it could strengthen the external validity of the Entrepreneurial State theory in relation to AI.

Furthermore, Westlake (2014) criticizes the narrow focus of the research. The theory neglects all the investment other than R&D related. With such a narrow view the argument would be easily made that the public sector is the leading risk-taker. Her argument that the state is the leading agent in investments would also be rather only stated than being explained (Mingardi 2015). This lowers the explanatory power of the argument made in the Entrepreneurial State theory.

To sum up, the Entrepreneurial State theory has been presented in this paragraph. The theory argues that the state has a far more prominent and pro-active role within a country’s strategy. The state would have more leading and risk-taking entrepreneurial actions. The picture of a sluggish and slow public sector is countered with one of a visionary and innovative character. The next paragraph will briefly conclude with a comparison of both theories in relation to the sub-question. It will create expectations based on the literature reviews.

2.2.3 Conclusion & Expectations

Following the three features of entrepreneurship provided in paragraph 1.1, this part will form three sets of expectations by linking them to the theories of Cerny and Mazzucato. First it will compare the two theories in relation to risk-taking, and after respectively to innovation and management.

Cerny and Mazzucato have different views on public investments. The Competition State theory prefers a lower level of public investments due to the crowding out effect. These public investments would be in the way of the investments of the real entrepreneurs in the private sector. The state should, therefore, limit its actions to facilitating the private sector through for example tax systems, and only funding basic R&D and education. On the other hand, Mazzucato (2013) believes in the crowding in effect. She believes that public investments would boost private investments. Moreover, the state as a leading risk-taker would have started innovation in sectors in which the private sectors would

(22)

have shown to be too risk-averse. According to Mazzucato (2013), the state as leading risk-taker with significant amount of public investments has played a crucial role in high-tech sectors. Based on these opposing arguments, the first set of expectations will be the following:

1. The Competition State: the level of public investment would be relatively lower.

1. The Entrepreneurial State: the level of public investments would be relatively higher.

The second set of expectations is on the innovation feature of entrepreneurship. Cerny (1997, 2010) argues that the real innovators are in the private sector. The state should promote innovation and growth in general instead of picking winners through an activist industrial policy. On the contrary, Mazzucato (2013) argues that the state has the power to pick winners. She points at Silicon Valley as an example. Moreover, in this success the state has played a crucial role in R&D. Public R&D would play a big role from the moment of the invention to the successful commercialization phase. Public labs would be crucial because many firms, and hence their private R&D labs, die during the innovation process. From these two views on innovation, the second set of expectations is:

2. The Competition State: the role of the state in the innovation process would be relatively lower.

2. The Entrepreneurial State: the role of the state in the innovation process would be relatively higher.

The last set of expectations is related to the feature of sound management. Cerny (1997, 2010) argues that the state should be limited to regulating to avoid government overload. Through deregulation and privatization the Competition State theory makes the private sector the leading player in developing an industry. On the other hand, Mazzucato (2013) sees a more proactive and leading role for the state beyond just

(23)

facilitating the private sector. She argues that the state is a leading agent in achieving innovation. From these different opinions, the next set of expectations is used:

3. The Competition State: the private sector would have the expertise to manage the development of the sector.

3. The Entrepreneurial State: the public sector would have the expertise to manage the development of the sector.

These three sets of expectations form an overview of the theory reviews. These expectations will help answer the three sub-questions, and hence the overarching research question. The next paragraph will conceptualize industrial strategy.

2.3 Industrial Policy

This research will analyze the role of the state based on industrial strategy and AI strategy documentation of the UK. Therefore, this sector will review the literature on industrial policy. It will start with conceptualizing industrial policy. Secondly, it will show how the views on industrial policy have changed over time after WW II. This will be followed with a more in-depth review of the debate on industrial policy in the latest phase. The overall function of this paragraph is to provide a better understanding of the sources, namely the national industrial strategy and national AI strategy documentation, used for the empirical analysis in this thesis. Moreover, with the argument of Rodrik (2004) it will show how the predominant entrepreneurial role of the state fits with the current industrial policy. As a result, it will support Rodrik’s (2004) argument that industrial policy should be reconsidered.

2.3.1 Concept and Definition

Before conceptualizing industrial policy it’s important to note that this thesis doesn’t make a distinction between industrial policy and industrial strategy. Rhodes and Brown (2018: 4) write in their paper that industrial strategy is traditionally understood as a set of economic policies that seek to support or develop specific industries with a special focus on manufacturing. These days, however, it’s used in a broader way in which not only industry and manufacturing is included. The only difference is that industrial strategy can also have social and environmental purposes (Rhodes and Brown 2018: 4).

(24)

Since this thesis is only focusing on the economic aim of industrial strategy, it will make no difference between the two terms.

The Donor Committee of Enterprise Development (DCED) (2017) defines industrial policy as ‘the strategic effort by the state to encourage the development and growth of a sector of the economy.’ Owen (2012: 3) refers in his article to industrial policy as measures taken by governments to bring about industrial outcomes different from those that would result if markets were completely free. Cimoli et al. (2009: 2) add that it can be wide-ranging from support of infant industry to policies affecting trade, science and technology. Moreover, these government interventions can, according to Di Tomasso and Schweitzer (2013: 3), be directed to manufacturing as well as services, but also firms, regions or sectors. Industrial policy should therefore not been seen as just financing certain industrial production and manufacturing.

Furthermore, industrial policy can be divided into horizontal/general and vertical/selective. On the one hand, horizontal industrial policy is help offered by the government to all businesses without any form of discrimination. It’s aimed at increasing the quality of the business environment in general (Berry 2016: 835). This means, in other words, that every firm in the sector, industry, or even in the entire economy is supposed to profit equally from it. Uvalic (2014: 2) sums in her article the possible tools for horizontal industrial policy up ranging from creating a supportive business environment and providing incentives to R&D to services for the creation of enterprise networks. Horizontal policy is seen necessary to complement the market that could not be relied upon to generate investments for certain things, such as basic research (Owen 2012: 4). Relating this part back to the two theories, this thesis will argue that the horizontal/general industrial policy fits into the Competition State theory. As mentioned in the previous paragraph, Cerny (1997, 2010) was of the opinion that the state should promote innovation and growth in general instead of an activist industrial policy that would pick winners.

On the other hand, vertical industrial policy is a more targeted state intervention. It’s the more traditional one that has been often blamed for ‘picking winners’, or as some argue ‘picking losers’ (Wade 2012: 226). The vertical version does target specific industries,

(25)

sectors, and firms to improve (Berry 2016: 835). As a result, it favors particular industries over others. With the argument of Mazzucato (2013) that the state has the power to pick winners, such as Silicon Valley, the vertical/selective industrial policy leans more toward the Entrepreneurial State theory. This division will later be used in the operationalization of innovation explained in the next chapter. The next paragraph will first turn to the different phases industrial policy has gone through in Europe.

2.3.2 Evolution of European Industrial Policy

The literature on the history of European industrial policy is often divided in three periods (see: Owen 2012 & Warwick 2013). This paper will also use these three periods to illustrate how industrial policy has evolved over time. The three periods are: (1) early 1950’s to late 1970’s, (2) early 1980’s to late 1990’s (3) Early 2000’s to present.

Phase 1: Early 1950’s to late 1970’s

The first phase after WW II was dedicated to creating national champions, which were seen as crucial factors for the development of the national economy (Owen 2012: 5). The states, mainly in Western Europe, had direct roles in the structuring and operating of the economic market (Thatcher 2014: 10). With this power it favored national champions or specific sectors, which had to give the overall national economy a boost. Some industries, such as coal and steel, and electricity, were seen as the fundament to recover from the war period. These industries received much government support. Naude (2010: 10) writes that states adopted a wide range of very selective industrial policies such as heavy protection of infant industry and import tariffs and quotas. Another reason that European governments provided significant assistance was to close the technology gap with the US (Owen 2012: 5). This gap had been established in the years before this first phase. The war had given the US the lead in many high-tech sectors. Moreover, Western European countries also had to face more competition in the older textile and shipbuilding industries from low-wage countries. The European states used industrial policies to withstand these competing rivals. Therefore, governments justified the strong use of industrial policy to protect its national economies against old and new competitors. Creating national champions was the focus of the industrial policy. If they were able to compete with foreign rivals, the entire national economy would grow. As a result, industrial policy during this phase in Europe was seen as something that was necessary for a country’s economic development.

(26)

Phase 2: Early 1980’s to late 1990’s

Whereas the first phase was one with strong and selective industrial policy, the second phase was one in which this view on industrial policy completely turned around. An underlying neoliberal economic paradigm emerged, and would start dominating the economic policy making of states (Cowling & Tomlinson 2011: 832). The US and the IMF started globally supporting the Washington Consensus for countries to develop economically. The Washington Consensus was founded on the idea of deregulation, privatization, and market liberalization. Interestingly, these features match with the ones mentioned in the Competition State theory of Cerny (1997, 2010). It was in this phase that states would have gone through their transformations from welfare states to competition states. Government failures were seen as worse than market failures during this time (Warwick 2013: 18). Owen (2012: 23) points to ‘eurosclerosis’, which includes a rising unemployment and high inflation in Europe, as an argument of economists to promote the neoliberal discourse. The economic stagnation in Europe could only be solved through a boycott of industrial policy. Industrial policy was from now on a taboo, which was according to Di Tomasso & Schweitzer (2013: 2), based on the assumption that a state should have a role as little as possible. Probably the most well-known citation of this period came from Nobel laureate Gary Becker (1985): ‘The best industrial policy is none at all’ (cited from Rodrik 2008). The highly used industrial policy of states in the first phase had to make room for the functioning and mechanisms of the market itself as a tool for economic growth of a country.

Phase 3: Early 2000’ to present

In this phase the discussion is mainly about the ‘revival’ of industrial policy. In the literature several reasons for this revival have been mentioned (see: Warwick 2013). The two main reasons will be briefly reviewed in this part. One of the reasons given for this revival is the financial crisis of 2008. Owen (2012: 41) argues that the financial crisis led to government interventions on an extraordinary scale. Due to the financial crisis of 2008 the EU would reconsider its approach towards industrial policy. In other words, the EU started supporting a more targeted industrial policy to keep the EU competitive (Uvalic 2010: 3). This was a remarkable move, because targeted interventions was blamed for its negative consequences of the first phase, and turned into a taboo in the second phase. Nonetheless, the financial crisis had damaged the faith

(27)

in the market severely. Its long-lasting negative implication forced governments to intervene (Wade 2012: 224). A second reason is the economic successful development of countries that did not follow the rules of the Washington Consensus, with a special focus on China (Stiglitz et al. 2013: 5). European states felt the need for industrial policy to compete with these rising powers. Tomlinson (2014) argued that these countries have shown that states can have a positive influence on the economic development of a country. So both the financial crisis and the rise of new economic powers have dominated the discussion on the global revival of industrial policy in European countries.

This part has showed how in general the use and views on industrial policy in Europe has changed over time. Since industrial policy has proved to be a debatable subject within the field of IPE, it shows why this new phase of industrial policy is interesting to research. The next paragraph will therefore have a more detailed review on the debate on industrial policy in this third, and current, phase.

2.3.3 Review of the New Debate on Industrial Policy

As could be seen in the previous paragraph the debate on active industrial policy has been present in Europe since WW II. Uvalic (2012) shows in her paper that the revival of industrial policy went along with a new rising academic interest on the subject. Rodrik (2004, 2008), Lin and Monga (2010) and Chang (2010), among others, discuss the revival and how modern industrial policy should be understood. This paragraph will briefly introduce the debate on industrial policy for the 21st_century.

Rodrik (2004: 3) starts his argument by stating that the debate should no longer be about the traditional industrial policy with an autonomous government applying subsidies and taxes. The modern industrial policy plays an active role in creating a ‘strategic collaboration’ between the government and the private sector. Moreover, the belief behind the new modern industrial policy is one of constructive roles for states in shaping the national economy, instead of just expanding industry (Stiglitz et al. 2013: 13). This partnership between state and private sector would be the best chance to find the main obstacles for economic growth and to come up with functional ideas to overcome them. The new debate on industrial policy proves, as Tomlinson et al. (2015) write, the increased recognition of a greater role for the state with an active industrial

(28)

policy to overcome the obstacles and externalities experienced in the previous years due to failing market mechanisms. The focus of the debate therefore has moved from the traditional market failures that would have been in need of subsidies and taxes to one that is about a broader failure of the market or even the complete system. Whereas with the first ‘smaller’ failures there was a discussion on whether industrial policy was the right tool to fix it, in this current debate the focus is more on the question of which future direction industrial policy should take.

Nahtigal (2014: 180) writes that the new modern industrial policy will be crucial for the advanced and successful European to stay competitive in the 21st_{century. Rodrik (2004:}

3) adds that the new industrial policy should be seen as a discovery process in which government and enterprises learn from each other. However, the future direction of industrial policy is highly uncertain. As Warwick (2012: 7) shows, countries all have different thoughts on what strategies to use. Some countries see industrial policy as the effective tool to rebalance their economic growth trajectory. In other words, some sectors have become too dominant and therefore a country depends too much on them. Other countries call upon industrial policy to strengthen a specific sector, which the government thinks will be profitable or increase the country’s competitiveness.

Despite the recognition of the power of the current publicly published industrial policies the scholars in the debate admit that industrial policy will not always succeed. Industrial policy has often been blamed for eventually picking losers instead of picking winners (Lin & Monga 2010). However, seeing industrial policy as a discovery process would mean that some parts of the industrial policy could fail. Therefore, Hausmann and Rodrik (2002) argue that industrial policy always brings a certain failure rate. To deal with this failure rate, a government should be capable of letting losers go, and reinvest in new possible winners (Rodrik 2008). The new industrial policy is therefore introduced in the new academic debate as a discovery process of a strategic collaboration between the state and private sectors. Since both have not all the required information, there is always a certain failure rate involved.

This part shows that the focus on the debate on industrial policy has changed. Instead of questioning the existence of industrial policy, scholars are debating on how the new

(29)

industrial policy should be understood for the 21st_{century. The debate shows that the}

use of 21st_{century industrial policy by a state shouldn’t any longer be considered as an}

autonomous state applying taxes and tariffs, but as one in which the state tries to create an entrepreneurial relationship with the private sector. This paragraph provides a better understanding of how a predominant entrepreneurial role of the state fits within the context of the current industrial policy.

2.4 Artificial Intelligence

Paragraph 2.4 will focus on AI. It will first conceptualize AI. After it will review the potential influence and economic effects of AI globally. By analyzing the expected impact of AI, this part simultaneously elaborates on the societal relevance of this research.

2.4.1 Concept and Definition

Hall and Pesenti (2017: 8) describe AI as a set of advanced general-purpose digital technologies that enables machines to do highly complex tasks effectively. The Engineering and Physical Sciences Research Council (EPSRC) of the UK adds that AI technologies have the goal to reproduce, or even outperform, certain abilities that require intelligence as if humans were to perform them. In other words, AI has the goal to replicate, and even exceed human intelligence. The EC (2018: 19), furthermore, conceptualizes it again in a different way by arguing that traditionally AI is seen as machines or agents that are capable of observing the environment, learning, and taking intelligent actions based on what is analyzed in the first two steps. Geere and Highfield (2018) mention that AI is often confused with machine learning. They state that the big difference between the two is that machine learning, as the name already says, is a process in which people try to learn a machine something. With AI the machine can think without having anything programmed to learn it.

2.4.2 The Global Impact of AI

In the IPE literature is often discussed what factors have made some countries rich and others poor. One example is the book of Helpman (2004) called ‘The mystery of economic growth’. Helpman (2004) tries to convince his readers that innovation and institutions made countries economically develop. This combination is crucial, because innovation can only lead to growth when it’s managed in the right way. However, Lin and Monga (2010) show that only recently innovation has been included in the modern economic growth theories. In the more traditional theories, such as the Solow growth theory and

The role of the state in the global economic competition for artifical intelligence