Professional Documents
Culture Documents
Dimensions of analysis
Character of state: Type, extent intervention (IP tradition or not), capacity, specific
institutions.
Character of the private sector: Type companies, role finance, type of innovation.
Character of state-society relations: Relative strength, cooperative or not, skills regime…
The approach started with the distinction between two types of models of
capitalism/development within the west, namely Anglo-Saxon and Continental countries:
First, the relation to the market and private actors. In Anglo-Saxon countries, public
opinion is much more positive towards the market than in continental countries.
Governing the market: process versus content. The Anglo-Saxon countries, the market is
governed in its process: the way it works, that the market remains free, that there is
enough competition. In the continental countries, the outcome and content need to be
governed and corrected (e.g., income redistribution).
Role of banking and capital markets: In the Anglo-Saxon countries capital markets are
much larger and developed relative to the size of their economies compared to
Continental Countries. The idea being here is that business rely much more on capital
markets for their funding; companies are competing with each other for their funding.
Whereas in the Continental countries; funding has come much more from the banking
sector rather than capital markets.
The organization of businesses:
France
“Bureaucratically-planned capitalism”, being like Germany but more centralized.
Much more centralized than Germany, meaning that the state directed much more where the
economy goes. An example is ‘pantouflage’ . The term refers to the practice by which high-level
French civil servants, usually former students of the École Polytechnique or École nationale
d’administration obtain work in private enterprises.
Dirigiste role state: The state has very detailed plans on where the economy needs to go,
also regarding postwar funds, MoF, and industrial policy.
The Grands Corps is centralized.
Grande ecoles : big schools that educate students to become top bureaucrats.
o i.e., high level of company or industry specific training.
Netherlands
“Liberal corporatist capitalism”
The Polder model and tripartite consultation (state, businesses, labor unions)
There is decentralization (liberal part), but it is not that these groups are competing for
each other for influence. All of them are brought together to cooperate and make certain
decisions.
Collective leadership and anonymous business. Leadership should not be very personal.
Informal planning to gather all the relevant groups together. (achterkamer politiek)
Large financial sector
Relatively organized labor and equality
Egalitarian and free education
Limited industrial planning
PPPs
The United Kingdom
“Personal Capitalism”, meaning that the country very liberal with strong competition and little
focus on cooperation. Like the Netherlands the country is decentralized and liberal, but
compared to the Netherlands the UK is much more focused on competition.
Seminar 1
If our minds are biased, so is our artificial intelligence.
Definitions:
Artificial Intelligence The theory and development of computer systems able to perform tasks that
normally require human intelligence, such as visual perception, speech recognition,
decision-making, and translation between languages
Machine learning A subfield of AI gives the machines the ability to learn progressively from
experience.
Lecture 2: Varieties of Capitalism in East Asia
The developmental state in East Asia
A different literature
Little connection with VoC literature
The developmental state literature looks at what developmental states are, and why they
are geographically concentrated in East Asia?
C. Johnson, MITI and the Japanese Miracle, 1982: This book is about the Ministry
International Trade and Investment in Japan and Johnson argues that the Japanese elite goes
to MITI, and that place has all the talents and elites to make the steps and planning to make
Japan grow. Hence, Japan is not growing because of the market, rather Japan is growing is
because of the state. The state is the engine of growth, rather than the market.
A. Amsden Asia’s next Giant (South Korea): This book looks at South Korea, and Amsden
concludes that – just like in Japan - the state is the driver of the economic miracle.
Wade, Governing the market (Taiwan): This time a framework of Taiwan, in which again the
state plays a leading role in economic success.
Sources of Embedded Autonomy (p. Evans): looks at what makes these developmental states
different from states that fail to reach development. They have something orthodoxically,
namely their embedded autonomy. The bureaucracies are embedded; meaning that they
know their way in society and businesses, and bureaucracies have personal relations with
society and businesses. These embedded characteristics are formed by informal links.
Autonomy means that the bureaucracy has enough influence and talent so that it can stand
up against the lobbying of the business sector.
In Japan, the state is less centralized but rather there are powerful businesses – such as
Toyota and Nintento, that form the ruling alliance to the Zaibatsu. The Zaibatsu refers to
large capitalist enterprises that developed in Japan after the Meiji Restoration.
1920 was a year of change, as their Ministry of International Trade and Investment (MITI)
became the lead organization. Included was a post-war shift towards demilitarization,
democratization and decentralization.
Between the 1950s and the early 2000s the Keiretsu System became prevelant in Japan. The
Keiretsu System entails one business with a huge network in which other businesses are
connected as suppliers or financial institutions that help the business to grow. This is how
Japanese businesses are organized. There are strong relationships within a business keiretsu
(network).
Part of the Keiretsu system is lifetime employment: a strong position for labor that is
integrated in businesses themselves rather than through labor unions. Businesses
themselves have a strong loyalty towards their employees.
In 1985 the Plaza Accord was signed, which dramatically increased the value of the yen
relative to the dollar, with the unintended consequence that it paved the way for Japan’s
“Lost Decade” of sluggish growth and deflation. The Plaza Accord led to a tremendous
speculative bubble in Japan of inflation. In the early 90s the bubble busted in the balance
sheet recession. Since then, the economy has not been growing that hard due to deflating
pressures. Now after Covid there are first seeing inflation again. Although it is still the biggest
exporter of the world, the financial crisis in the 90s damaged the economy.
Characterizing for South Korea is their Chaebol model: a large industrial conglomerate run
and controlled by an individual or family. Picking winners or not. If you pick business as
winners as a state (by funding innovation), then you make businesses inefficient and lazy.
Rather, the market competition should pick a winner.
1970s: HCI drive. An example of what most Asian countries did, namely to look at where
they have an advantage: clothes, textiles etc. In these basic industries you start, and then
move to sectors that are higher valued. So you start with clothes, and then move to steel
production. Then from steel production to electronics, and so on. They would use credit to
stimulate a new industry.
South Korea today: successful in more recent technological trends (gaming, pop-music,
cosmetics).
Hong Kong
Started from British colony
Very connected to global economy. In contrast to Singapore: much more liberal. Although
the state is organized, there is much more a Western free market system and corporate rule.
It links with China are very strong: integration into China.
China
During the communism regime of Mao the central state was strengthened a lot in terms of
social economic indicators. For instance, basically everyone in China is literate and
healthcare is quite developed. That is the legacy of communism: they did not create wealth
perse but created conditions that currently help China a lot.
But when Deng Xiaoping took power, growth really started. “To get rich is glorious”, “bide
your time”.
State planning: the communist and imperial legacy. The state plays a big role and
forces the market to go into a certain direction with 5-year plans.
Crossing the river by stepping stones: Yes China plans, but if they plan in the wrong
way, they can adjust. It is very experimental: you take a step and if you can take the
next step; then you take it. They really try to work what works and what does not,
meaning it is very experimental and pragmatic. This makes their economic policy so
successful.
Picking winners late: unlike Japan or Korea, China does not start with its huge
businesses in the beginning. Rather, it lets the market run free and let SMEs compete
with each other. Then it observes what emerges, and once it has the required
technology, China helps to consolidate the market by picking winners.
Strong scientific focus
A history of openness and diversity: due to its Maritain and ocean nature. Therefore,
much more links with the outside world.
Economic structure: much endowed with natural resources than Northeast Asia.
Natural Resource Curse: if you don’t have resources to get your country wealthy you
must educate your population, if you do have resources you become lazy.
Furthermore, resources induce rent-seeking.
Different state traditions and colonial legacy.
Impact on innovation:
o Openness to foreign firms: not local and thus the governments intervene less.
o The middle income trap: in these countries it is relatively easy to go froom
poor to medium, but after that people keep stuck in the middle income,
because the state does not intervene enough to obtain the higher incomes.
For higher incomes one needs innovation, skilled levels, capital and for that a
strong state is needed.
Exceptions: Singapore and Vietnam.
Vietnam
Very similar to China. Both communist and strong state tradition. Also a long way of dealing
with China, making structure similar and policies mirroring China. Vietnam tries to copy the
developmental state of China.
Seminar 2
Liberal Market Economy: Australia , USA.
Coordinated Market Economy: Germany, Sweden, France
LME
Institutions are coordinated
General skills because the labor market is more fluid
Minimalism influence state.
Union groups are weak or non-existent, hence unionization is low.
CME
Focus on long-term and professional skills because the market is not so fluid.
Maximalist influence state
Strong presence of labor unions
5 spheres:
1. .
2. .
3. .
4. .
5. .
USA in practice
Investment and general trust in cooperation is important. Therefore, there is a lot of
publicly available information. Hence, networking is important.
o Your own networks and connections are important.
Unilateral decision making
Relations to employees is less based on trade unions, and hence workers have less
bargaining power.
Less educated population, because education is much more general focused, because
the labor market is fluid. They limit very specific trainings.
Technology Transfers and Research and Development take place due to internal
competition.
Shock therapy: with also disastrous impacts -> chaos, hunger, shortages , corruption
etc.. As in Ukraine, the liberal Shock therapy created space for people to monopolize
that power. Hence again a rise of oligarchs.
Rise (and fall) of oligarchs -> became economically very powerful, whereas in Ukraine
the oligarchs have remained quite powerful as a decentralized device, but in Russia
they got under the wing of Putin who required that they had to stay out of politics
and assist the state at all time, thereby centralizing the country.
Endemic corruption -> in China corruption decreases as you go higher up the chain,
and thus more persistent low level corruption. In Russia it is the opposite, the higher
up you go the freer you are, and hence the more space they seek for corrupt
practices. (unless you are too powerful and considered a risk by the Kremlin).
Resource dependency -> energy
Weakened scientific base: previously during soviet times it was quite advanced
Pockets of efficiency -> although really resource dependent that does not create a
lot, certain businesses really excel, and are really advanced.
Dismantling the license raj (‘king’): after independence the state with arrogant
bureaucrats were inhibiting growth, this needed to be dismantled for India to
become a strong emerging market. Hence, from 1991 the license raj started to be
dismantled and now India is becoming a stronger economic market.
Relatively weak state capacity: for several reasons. First, India has not had a strong
communist movement that created strong institutions, as the British imperialists
never created strong institutions or bureaucracies, the capacity of Delhi to make
plans and implement is relatively weak. This can be seen in public goods that are
weakly developed (education (especially for women) etc). A strong state could have
provided such public goods, the weak government of India fails to do so.
Competitive domestic market: The country is huge and many services are not
provided by the government, and hence you do have a competitive market:
businesses are not focused on government sectors but rather focused on private
sector.
High skill services: IT sector in India is high-skilled because some schools and
universities are very high skilled. These s
Country size
A large domestic market has an impact on the type of economic model you can follow. A
large domestic market can help to spur certain economic activities. But small countries need
to be tremendously efficient to be internationally competitive. Because smaller countries
don’t have large domestic markets, they are more vulnerable to external shocks. That is why
smaller countries usually have more active models shaping their economy – trying to create
niches – in order to direct the economy in a robust and resistant direction.
Natural endowments
Whether a country is strongly reliant on agriculture, mining, etc. And if so, what kind of
political economy do they create? Meaning; how the endowment distributed and how easy
is it monopolized?
Definitions AI
Common definition: ‘AI is the imitation of complex human abilities through machines’
Nilsson: a technology that can adequately function by dealing with its environment in a
prescient way
Elements in many definitions: ability to perceive, attain goals, initiate information/actions
and learn from feedback loops.
o But might be short-sided. A classic thermos can do the same…. Measure the
temperature, start and finalizing the display of the temperature level.
Useful one by the HLEG: AI refers to system that display intelligent behavior by analyzing
their environment and taking actions – with some degree of autonomy – to achieve specific
goals.
It is difficult to define AI because it recreates something we do not understand yet, namely human
intelligence. AI does out play people in some cases – like in chess during the 1997 Brain’s last stand.
But in other cases, AI is less advanced, social interactions, norms and values.
The Moravec paradox: it is comparatively easy to make computers exhibit adult level performance
on intelligence tests or playing checkers, and difficult or impossible to give them the skills of a one-
year-old when it comes to perception and mobility. What is easy for humans is difficult for machines,
and what is easy for machines is difficult for humans.
AI as Horizon
The horizon or point is AGI: Artificial General Intelligence. That is what humans have; the ability to
apply different types of intelligence and move between them (drive a bike, have a phone call, and
think about something). The opposite is Narrow AI: intelligence in one specific field. In that specific
field it may surpass humans (calculators, chess), in the others AI might not. For now, AI cannot
change between the specific fields of intelligence.
Are we getting closer to AGI? We cannot predict the future, but research shows that it is much
further away than one might expect. Estimations now point towards 2100.
The wrong question: people worry that computers will get too smart and take over the world, but
the real problem is that they’re too stupid and they’ve already taken over the world. We now have
Narrow AI, which already decides government subsides, police control, etc. So instead of speculating
about the future, we should think about what is already happening around us.
Ancient myths about AI show that AI does not come from nowhere but has its origins in myths which
structure the way we perceive AI. There are many ancient stories about humans and non-humans,
where humans surpass their human boundaries and get punished for it.
The First Wave of AI started with the birth of computers and theoretical foundations: thinking about
what AI would look like (Alan Turing and Norbert Wiener). In 1956 the Dartmouth Summer Research
Project on Artificial Intelligence started. Projects and computer programs were developed that dealt
with logical questions: Logic Theorist (1955); the General Problem Solver that could give the solution
to a problem by following the rules of logic. The problem of the combinatorial explosion: if you have
a problem with a few options, then it is relatively easy to handle. But in the real world, there are a lot
of options requiring a lot of computer and program power. What was also difficult is that at that time
the systems did not have heuristics yet; things that simply the options that we need to check. As a
result, in the 70’s the first AI winter started: the enthusiasm, scientific support, and fundings for AI
dried up.
Two currents in AI
We can distinguish among two approaches in AI.
1. Rules-based/symbolic systems/expert systems: Where all the rules are programmed in
advanced. These rules have logical structures. The algorithm has to work with these rules
subsequently.
2. Neural networks / connectionism / machine learning: How neurons work in the human
brain; the rules are not programmed in advanced but the algorithm has to find its own rules.
‘Fire together then wire together’ = the algorithm has to create its own connection and learn
by itself. This type of AI requires a lot of data, data analysis and the supporting computer
power. This makes it hard – so far – to yield net positive results.
In the 80’s the second wave of AI started, as Japan started to focus on AI in its Fifth-Generation
Computer Systems Project. Other countries copied Japan; US, UK, Europe etc. What worked very well
in this wave were the expert systems: it is a rule-based system but the rules are provided by human
experts. Think about algorithms in the field of medicine, where doctors were asked about rules and
patterns that they followed when making a diagnosis. But also, here there were problems; the
hardware was not good enough and programming languages was hard. Moreover, an expert can tell
you what he sees when making a diagnosis, but it is very hard to explicitly state the rule. And if two
rules contradict each other; the entire system (based on logic) fails. Are the rules should be clear and
explicit; something we cannot do.
The third wave started in the 1990’s when the computer beated Kasparov (chest player). We are still
in the third wave, but something fundamentally has shifted. GOFAI: the rule-based system was too
brittle meaning that they could break down very easily if the rules were not coherent with each
other. Hence, more focus on the second approach on AI.
The second approach started to take off and work much better. Three factors were crucial for that:
1. Processing power: increasing processing power to enable the computer to handle the data
analysis (Moore’s law: every half a year the computer power doubles)
2. Availability data: increase in availability of data. With the rise of social media, we ourselves as
social media users enabled this by putting all our information digitally.
3. Algorithmic science: deep learning – different layers within algorithms. But the difficulty was
how to value the labels in different layers.
Approaches: Gofai, coding common sense: meaning that AI will reason based on ‘common
sense’, for instance coding indictive biases (e.g.; if something can walk, it probably lives),
graph networks, neurology.
Combinations of approaches until we have a big AI break through. Deep learning in itself
cannot solve everything, hence we need to combine different approaches.
The Ecosystem
AI itself is about the algorithm, but for it to function in a setting you need five things to work:
5 dimensions of AI strength
1. Hardware: (internet) networks, chips, and supercomputers to make AI work
2. Fundamental research: academic output, R&D who can devise the AI systems and create the
algorithms
3. Data sets: any algorithm requires data; large data sets, which should be commensurate and
representative.
4. Business to develop the AI: tech firms and large and small as well as incumbents firms.
5. Supportive government framework: AI regulation to make rules and legal code, and the
government should stimulate AI with investments, subsidies, places to experiments etc.
Lecture 5: Normative Issues surrounding AI
Public values under threat
Privacy
There could be data that we think is simple, but algorithms can distil from them much more
information. Amazon for instance, filed for patients to reveal emotion from talking to ALEXA. Hence,
information about emotion could become available from speech.
Is there really good data to train your algorithm on? Is data objective? Additionally, can the algorithm
can get feedback on choices it makes? There has to be a feedback mechanism; that is crucial! That is
why Google’s search output works so well; the moment we do not click on the top recommendation
the algorithm gets feedback. So, if you do not get feedback back, then you confirm and strengthen
biases.
Human control
People have right not to be subject to a decision which produces legal effects concerning him or
significantly affects him and which is based solely on automated processing of data intended to
evaluate certain personal aspects relating to him, such as his performance at work, creditworthiness,
reliability, conduct, etc..
3 models:
Human-out-of-the-loop: when humans are not in the picture. Think for instance the spam
folder in your mail.
Humans-in-the-loop: the algorithm can’t make a decision unless a human does something.
The algorithm can suggest something (I think this person is committing fraud), but a human
still makes the final call.
Human-on-the-loop: the algorithm can do the entire process; but a human supervises the
process and intervenes where necessary.
Opacity
We might slowly get into a situation in which society becomes opaque; in which we less understand
why things happen a black-box-society!
Digital authoritarianism.
AI might be used for democratic means, but can also be used by authoritarian states. Authoritarian
states can use AI for surveillance, controlling people; AI makes spying on people cheap, efficient and
large scaled. Think about the Great Chinese Firewall.
Structural changes
What we need to look at more in the future, are slow moving processes that might not be accurate
that could have a detrimental effect on society in the future. Think about the concentration of power
in the hands of a few big companies.
We currently see that the public sector is lagging behind the private sector in their AI application. The
public sector has higher standards than the private sector has, as the government is very responsible.
But if the private sector moves way faster in AI than the public sector; then a gap emerges.
Normative issues surrounding specific applications
Laws
Killer robots, weaponry, and other sorts of technology receives a lot of attention.
Facial recognition
Leads to controversy and people urge companies and governments to stop using facial recognition.
That is because facial recognition is a major asset for surveillance. So it reduces privacy, but it can
also be very helpful: for instance when a child is missing.
Predictive policing
CAS (criminaliteit anticipatie system) in the Netherlands
The object of this book is to elaborate a new framework for understanding the institutional
similarities and differences among the developed economies. The varieties of capitalism approach is
an effort to go beyond three perspectives on institutional variation that have dominated the study of
comparative capitalism in the preceding thirty years:
Modernization approach
Devised in the post-war decades, this approach saw the principal challenge confronting the
developed economies as one of modernizing industries still dominated by pre-war practices in order
to secure high rates of national growth. Countries were often categorized according to the structure
of their state, into those with ‘strong’ and ‘weak’ states.
BUT: tend to overstate what governments can accomplish, especially in contexts of economic
openness where adjustment is firm-led.
Neocorporatism approach
During the 1970s when inflation became the preeminent problem facing the developed economies,
this approach was generally associated with the capacity of a state to negotiate durable bargains
with employers and the trade union movement regarding wages, working conditions and social or
economic policy. Countries were categorized by reference to the organization of their trade union
movement.
BUT: its emphasis on the trade union movement underplays the role that firms and employer
organizations play in the coordination of the economy.
However, we think these approaches tend to miss or model too incompletely the strategic
interactions central to the behavior of economic actors. By locating the firm at the center of the
analysis, we hope to build bridges between business studies and comparative political economy.
Because its capabilities are ultimately relational, a firm encounters many coordination problems. Its
success depends substantially on its ability to coordinate effectively with a wide range of actors.
There are five spheres in which firms must develop relationships to resolve coordination problems
central to their core competencies:
1. Industrial relations -> how to coordinate bargaining over wages and working conditions with
their labor force, the organizations that represent labor, and other employers.
2. Vocational training and education -> securing a workforce with suitable skills, while workers
face the problem of deciding how much to invest in what skill.
3. Corporate governance -> access to finance and in which investors seek assurances of returns
on their investments.
4. Inter-firm relations -> relationships a company forms with other entreprises, and notably its
suppliers or clients, with a few to securing a stable demand
5. Employees -> problems of adverse selection and moral hazard, information sharing becomes
important
There are two types of political economies: liberal market economies (firms coordinate their
activities primarily via hierarchies and competitive market arrangements) and coordinated market
economies (firms depend more heavily on non-market relationships to coordinate their endeavors
with other actors and to construct their core competencies. (ideal types at poles of a spectrum along
which many nations may be arrayed).
In contrast to LME, where the equilibrium outcomes of firm behavior are usually given by demand
and supply conditions in competitive markets, the equilibria on which firms coordinate on CMEs are
more often the result of strategic interaction among firms and other actors. In any national economy,
firms will gravitate toward the mode of coordination for which there is institutional support.
A standard literature suggests that institutions provide capacities for (i) the exchange of information
among the actors, (2) the monitoring of behavior and (3) the sanctioning of defection from
cooperative endeavor.
Deliberative institutions — institutions that encourage the relevant actors to engage in collective
discussion and to reach agreements with each other — are important for several reasons:
It can substantially thicken the common knowledge of the group
It provides actors with an opportunity to establish the risks and gains attendant on
cooperation and to resolve the distributive issues associated with them.
They can enhance institutions can enhance the capacity of actors in the political economy for
strategic action when faced with new or unfamiliar challenges.
o Cross-national comparison should be attentive to the presence of facilities for
deliberation as well as institutions that provide for the exchange of information in
other forms, monitoring, and the enforcement of agreements.
Taking a step beyond many accounts, this approach emphasizes the importance of informal rules and
understandings to securing the equilibria in the many strategic interactions of the political economy -
> relevant is the analysis of history and culture, since the operative force of many institutions cannot
be taken for granted but must be reinforced by the active endeavors of the participants.
There are important respects in which strategy follows structure. For this reason, the approach
predicts systematic difference in corporate strategy across nations, and differences that parallel the
overarching institutional structures of the political economy.
Firms and other actors in coordinated market economics should be more willing to invest in specific
and co-specific assets (assets that cannot readily be turned to another purpose and assets whose
returns depend heavily on the active cooperation of others), while those in liberal market economics
should invest more extensively in switchable assets (assets whose value can be realized if diverted to
the purposes).
The presence of institutional complementaries reinforces the differences between liberal and
coordinated market economies -> nations with a particular type of coordination in one sphere of the
economy should tend to develop complementary practices in other spheres as well. The institutions
sustaining coordination in one sphere can be used to support analogous forms of cooperation in
others.
Nations with liberal market economies tend to rely on markets to coordinate endeavors in both the
financial and industrial relations systems, while those with coordinated market economies have
institutions in both spheres that reflect higher levels on non-market coordination.
In LMEs, firms rely more heavily on market relation to resolve the coordination problems that firms
in CMEs address more often via forms of non-market coordination that entail collaboration and
strategic interaction.
The financial systems or markets for corporate finance of LMEs encourage firms to be
attentive to current earning and the price of their shares on equity markets. The terms on
which large firms can secure finance are heavily dependent on their valuation in equity
markets. This encourages firms to focus on the publicly assessable dimensions of their
permanence that affect share price, such as current profitability.
In the industrial relations area, firms in LMEs generally rely heavily on the market
relationship between individual worker and employer to organize relations with their labor
force.
The education and training systems of LMEs are generally complementary to these highly
fluid labor markets. Vocational training is normally provided by institutions offering formal
eduction that focuses on general skills, especially suited to job growth in the service sector
where such skills assume importance.
Inter-company relations in LMEs are based for the most part, on standard market
relationships and enforceable formal contracts.
In is important to note, even within these two types, significant variations can be found, these are
only ideal types. Each economy displays specific capacities for coordination that will condition what
its firms and government do.
Why do particular nations tend to specialize in specific types of production or products? The
comparative economic advantage account have dealt with a serious blow, because of the expansion
of intra-industry trade and increases in the international mobility of capital.
-> comparative institutional advantage: the institutional structure of a particular political economy
provides firms with advantages for engaging in specific types of activities there. Institutional context
can condition rates of growth and technological progress.
The impact of the institutional framework on innovation: radical innovation (substantial shifts in
product lines, the development of entirely new goods, or major changes to the production process),
and incremental innovation (continuous but small-scale improvements to existing products lines and
production processes. Radical innovation is especially important in fast-,loving technology sectors,
which call for innovative design and rapid product development. Coordinated market economies
should be better at supporting incremental innovation.
Therefore, to the extent allowed by transport costs and the efficiency of international markets, there
should be national patterns of specialization in activities and products; and these should reflect
rational responses to the in institutional frameworks identified here rather than random geographic
agglomeration.
The principle problem facing policymakers is one of introducing economic actors to cooperate more
effectively with each other. Economic policies will be effective only if they are incentive compatible,
namely complementary to the coordinating capacious embedded in the existing political economy.
Coordination-oriented policies should be more feasible in nations with both a coordinated market
economy and a political system in which producer groups enjoy substantial structural influence.
Coordination-oriented policies will be more difficult to implement in liberal market economies
because their business and labor associations usually lack the encompassing character required to
administer such policies. Liberal market economies should find it more feasible to implement
market-incentive policies that do not put extensive demands on firms to form relational contracts
with others but rely on markets to coordinate their activities.
The character of the political regime may condition the levels of asset specificity found across
nations. The fluid market settings of liberal-market economies encourage investment in switchable
assets, while the dense institutional networks of coordinated market economies enhance the
attractiveness of investment in specific or co-specific assets.
Much of the adjustment process will be oriented to the institutional recreation of comparative
advantage. Institutional complementaries should play an important role in the processes of
adjustment. On the one hand, they raise the prospect that institutional reform in one sphere of the
economy could snowball into changes in other spheres as well. On the other hand, institutional
complementarities generate disincentives to radical change.
The developmental state literature is typically identified with an institutionalist approach to politics,
focusing on the autonomy or insulation of the government from rent-seeking private interests,
delegation to lead agencies, and coherent bureaucracies. The developmental state literature
resembles the Varieties of Capitalism approach to comparative-historical research, in which the
identification of types provided the basis for a critique of more generalized models.
Orthodox interpretations of East Asia’s rapid growth were deeply flawed, descriptively, theoretically
and methodologically. Neoclassical models were rooted in the claim that rapid growth was caused by
the adoption of market-conforming policies, most particularly with respect to export-led growth.
However, it turns out that these policies were not adopted to the extent believed, and rested on
other causal factors such as institutions and underlying social coalitions. The developmental state
literature posed itself as alternative to an emerging neoclassical consensus in development
economics.
The developmental state literature takes history seriously. Historical analysis is required because
successful strategies emerged only through a process of trail and error and learning by doing.
Moreover, the developmental state considers capital accumulation as the principal general force for
growth, and interprets superior East Asian performance as a result of that – rather than from free
market policies or interventionist economic policies. Johnson’s core theoretical insight is that the
process of economic development is characterized by a myriad of market failures that can be solved
only by government that can coordinate private sector activity, interning in markets and providing
selective and conditional support to firms.
Three kinds of coordination problems received particular attention: those that arose in moving from
agriculture to industry and within industry; those associated with financial markets; and those that
surrounded the transfer and adoption of technology.
The dominant political-institutional model of growth in economics and political science focused on
property rights, the rule of law, and checks on state power. But the nature of growth miracles in East
Asia suggests the deep anomalous quality of Asia’s growth: it is hardly a region characterized by
strong checks on state power, well-developed rule of law, independent judiciaries, or codified
property rights.
On the one hand, developmental states were characterized by substantial delegation to executives
and in turn from executives to capable and appropriately incentivized bureaucracies. On the other
hand, the developmental state was politically insulted from social forces that might distract from its
accumulationist and industrial policy aims, including not only the left and working class but also the
private sector itself.
In Amsden’s felicitous phrase, a key feature of the development state was tis capacity to ‘discipline’
the private sector, since private-sector capture could distort the allocation of subsidies from their
developmental purposes.
The idea that institutions could help solve these signalling, credible commitment and monitoring
problems between the state and the private sector was at the heart of the developmental state
approach. The developmental state model emphasized strong or ‘insulated’ states, coherent
bureaucracies, and institutionalized business-government relations that checked rent-seeking and
made government support conditional on private sector performance. But as case studies
proliferated they revealed tremendous variation across countries, including not only between the
Northeast and Southeast but also among Japan, Korea and Taiwan. The central lesson was that
growth was not simply the result of market oriented policy or even appropriate interventions but of
underlying state capacity.
Following the ‘varieties’ focus on corporate governance and labor relations, the four core empirical
features of Hierarchical Market Economies (HMEs) in Latin America would be:
1. Diversified Business Groups
Large companies in Latin America are divided between large domestic business groups and MNCs.
Often, these domestic firms are family-owned. They are clearly hierarchical because most firms are
directly controlled and managed by their owners, either prominent families or foreign firms.
2. Multinational Corporations
Most large firms are subsidiaries of MNCs, which have been long dominant in manufacturing and in
recent decades have also expanded into finance, utilities and other services. These MNCs are mostly
from the United States. MNCs administered, in hierarchical fashion, technology transfers, capital for
investment, relations with suppliers and customers and especially trade. Hence, MNCs are the key
conduits for organizing access to capital, technology and markets for Latin America.
3. Atomistic Labor Relations
There is an absence of institutions both for intermediating employment relations within firms and for
fostering greater investment in skills and training. Unions are small and represent a decreasing share
of workers, as the informal sector is so large. Labour relations are atomistic and often anomic
because most workers have fluid, short-term links to firms and weak or no horizontal links to other
workers through labour unions. As a result, labour and employment relations are individualized,
disinter mediated and consequently hierarchical. Hierarchical relations also characterize employee
relations.
4. Low Skills
Education levels are comparatively low, despite recent advances, and public and private investment
in training is minimal. What explains this low level of investment in skills? The common fear of
poaching discourages investment; if one firm invests in training workers, other firms can then poach
and hire away the trained workers. This is a true coordination problem.
These four characterisitics are interlinked and reinforce each other: institutional
complementarities. For instance, between atomistic labor relations and low skills: When
labour turnover is high and unions at the firm level are weak, em ployers have even weaker
incentives to invest in worker skills both because they expect workers not to stay long, and
because they lack the institutional means for negotiating with workers an explicit
distribution of gains over time from investing in trainig
In some respects, HMEs resemble CMEs (in their non-market forms of corporate government), and in
others they tend towards LMEs (as in labor markets). But, HMEs are not simply hybrids or mixtures
because of the hierarchical relations that characterize HMEs. Hierarchy pervades the core relations of
capitalism more in Latin America than elsewhere. For example, whereas post-secondary or on-the-
job training is more market-based in LMEs and more negotiated in CMEs, it is often unilaterally
decided by firms of businesses associations in Latin America.
Such hierarchical relations also characterise employee relations more generally, where employees
lack formal grievance procedures and representation and informally lack voice, because most of
them are quite temporary. Unions have little influence on hierarchies within the firm, in part because
so few workers are unionised, and in part because where unions do exist they are often distant from
the shop floor. Finally, industrial relations are further structured by top-down regulations issued by
national governments and enforced by labour courts.
The varieties of capitalism approach entail the idea that basic institutions of capitalism differ from
one country to another and that these differences are not accidental, but linked to strong
institutional complementarities. It is exactly this core hypothesis – that the character of institutions
within a successful economy are mutually reinforcing – that is at the heart of VoC theory. The VoC
approach departs from the juxtaposition of liberal market economies (LMEs) and coordinated market
economies (CMEs). Their main difference is the different type of innovation procceses that are
central to their production systems. CMEs are assumed to have a premium on incremental
innovation and LMEs on radical innovation. More general, usually five interdependent elements of
the VoC can be highlighted:
1. The financial system (the means to give rise to investments)
2. Corporate governance (the internal structure of the firm)
3. The pattern of industrial relations
4. The education and training systems
5. The preferred mode for the transfer of innovations.
In LMEs, the most important forms of coordination are competitive market arrangements and
formal contracts. In coordinated market economies, nonmarket forms of coordination, such as
interfirm networks play a crucial role.
From a comparative political economy perspective the Czech Republic, Hungary, Poland and Slovakia
are increasingly considered as four cases of the same basic variety – sharing similar socio-economic
institutions while being distinct from LMEs and CMEs. More specifically, the central characteristic of
the region is its external dependency. The author therefore suggest that ECE signifies the emergence
of a third basic variety: a dependent market economy (DME) type of capitalism.
DMEs have:
A comparative advantage in the assembly and production of relatively complex and durable
consumer goods.
These comparative advantage are based on institutional complementarities between skilled,
but cheap, labor
The transfer of technological innovations within transnational enterprises
The provision of capital through Foreign Direct Investment (FDI).
The DME approach incorporates transnational influence – in particular the role of transnational
companies – in an approach (VoC) that traditionally tends to consider socioeconomic systems as
closed containers.
The analysis also highlights some challenges for the long-term future of these DMEs: their
comparative advantages are constantly being threatened by countries located further to the east and
will continue to remain limited to segments of their economies, thereby leading to increasing
political tensions.
The crucial importance of an extraordinary crisis for the emergence of new socioeconomic institutions
is in this case the collapse of communism.
In many countries, there has been a shift from the rather broad and optimistic tone, to the ethics and
regulation of AI. Due to it’s rapid spread in the public and private sphere, the mission is now to
determine the best way to embed AI into society.
What is AI?
The strictest definition understands AI as the complete replication of human intelligence by
machines, although this is an (yet) unfulfilled goal. Up until now, AI has been able to exert specific
features of human intelligence, such as interacting, self-learning and predicting. It is in general hard
to define AI, as the concept is embedded in that of human intelligence, which is a goal that we do not
fully understand. The formal definition the WRR uses states:
“AI refers to systems that display intelligent behaviour by analysing their environment and taking
actions – with some degree of autonomy – to achieve specific goals.
When we look at national competitiveness, the question of how to assess a country’s ai capacities
arises. Although there are several ai indices, this query is complex and needs to take several variables
into account. Based on the work of several authors, we discern five variables:
1. The quality of fundamental research
2. The availability of data
3. The required hardware
4. A dynamic private sector to commercialize the technology
5. An enabling government
To conclude, across all of the five tasks, AI requires work from society as a whole to become
embedded within our daily lives. The five tasks can be supported by developing an institutional
infrastructure that facilitates the exchange of knowledge and the coordination of state activities.
Hence, the final recommendation states to establish a policymaking infrastructure for AI< starting
with an AI coordination centre that is embedded into the political process.
AI Now 2019 Report
Some concerns
Bias Built In
In AI-algorithms, there is a proved bias in advertising networks, facial recognition, search engines,
welfare systems and even in algorithms using criminal sentencing.
‘ Garbage in Garbage Out’
Short summary:
A noticeable trend has been the increasing reliance of public decision-making bodies on algorithms,
i.e., computer programmed instructions for taking a given set of input to produce output.
Problematic however is that algorithm-driven decision-making does pose a significant threat to the
legitimacy of such processes. Namely: would the resulting outputs be morally problematic? And:
would such processes be morally and politically legitimate? Could algorithm-based decision-making
somehow undermine this legitimacy?
Danaher argues that it could. According to the author,t here is a serious problem concerning AI’s
opacity, the potential incomprehensibility to human reasoning. Hence, increasing reliance on
algorithms gives rise to the threat of algocracy – a situation in which algorithm-based systems
structure and constrain the opportunities for human participation in, and comprehension of, public
decision-making. To solve this issue, the author the author proposes two solutions – although neither
is expected to be entirely successful: “resistance and “accommodation”
What is algocracy?
The author uses the term algocracy to describe a particular kind of governance system, one which is
organised and structured on the basis of computer-programmed algorithms. More specifically, it is a
system in which algorithms are used to collect, collate and organise the data upon which decisions
are typically made, and to assist in how that data in processed and communicated through the
relevant governance system.
Humans can be more or less involved in the data-mining process and in the kinds of decisions made
on the foot of the data-mining process. We can distguish between three roles, and we illustrate it
with robots:
1. Human-in-the-loop: Robots can only select targets and deliver force with a human
command.
2. Human-on-the-loop: Robots can select targets and deliver force on their own, but there is
human oversight and the possibility of human override
3. Human-out-of-the-loop: Robots act autonomously, selecting and delivering force without
oversight or override.
What is the threat of algocracy?
Algocratic systems can be said to raise two moral and political concerns:
1. Hiddenness Concern: This is the concern about the manner in which our data is collected and
used by these systems. People are concerned that this is done in a covert and hidden
manner, without the consent of those whose data it is.
2. Opacity Concern: This is a concern about the intellectual and rational basis for these
algocratic systems. There are worries that these systems work in ways that are inaccessible
or opaque to human reason and understanding.
a. We might initially favour algocratic governance systems for instrumental reasons,
and be impressed by their greater speed, accuracy and insight. But in favouring
them we may end up with systems that are increasingly opaque. We might see all
these amazing thins algorithms can do for us, but we do not know how exactly they
manage to do this. The result is social spaces opaque to human reason and a set of
decision-making procedures that are depleted of their legitimacy as the scope for
genuie human participation becomes more limited.
Summarized:
1. Legitimate decision-making procedures must allow for human participation in and
comprehension of those procedures
2. Increasing reliance on algorithmic systems limits the scope for active human participation in
and comprehension of decision-making procedures
3. Therefore, reliance on algocratic systems is a threat to legitimate decision-making
procedures.
A growing number of states are deploying advanced AI surveillance tools to monitor, track and
surveil citizens to accomplish a range of policy objectives. To track AI technology, the AI Global
Surveillance (AIGS) Index provides a detailed empirical picture of global AI surveillance trends and
describes how governments worldwide use this technology and by whom they got the technology
supplied. The index tells us that at least 75 of 176 countries are actively using AI technologies for
surveillance purposes, such as smart/safe city platforms, facial recognition systems and smart
policing. Moreover, China is the major driver of AI surveillance worldwide, as a supplier of AI
surveillance technology in different countries, especially the Chinese Company Huawei. However, the
U.S is also active in this regard. Problematic however is that democracies are not taking the
necessary steps to monitor and control the spread of advanced technologies that might cause a
range of violations. The quality of the government is hence the main factor to determine if
governments will deploy the technology for repressive purposes or not. More specifically,
governments in autocratic and semi-autocratic countries are more prone to deploy AI in abusive
manners than liberal governments. Examples are China and Russia, who use AI for mass surveillance
purposes and dismal human rights.
State surveillance is not inherently unlawful. For example, tracking tools play a crucial role in
preventing terrorism. But, technology has also changed the nature of how governments carry out
surveillance and what they choose to monitor, and that profoundly affects an individual’s right to
privacy. There are several obligations outlined by the international human rights law, that are critical
to assess the lawfulness of a particular surveillance action. This entails that the legal framework
should be accessible, clear, comprehensive and non-discriminatory, that the surveillance is applied to
situations for which it is strictly and demonstrably necessary, and that the interests justifying the
surveillance are legitimate.
When smart cities employ safety technologies, we call them safe cities and they employ technology
to predict, prevent, and reduce crime, and address new and emerging threats.
Smart Policing
The idea of smart policing s to feed immense quantities of data into an algorithm – such as
geographics, historic arrest levels, types of committed crimes, biometric data, social media feeds – in
order to prevent crime, respond to criminal acts, and make predictions about future criminal activity.
But gives the police unprecedented access to data
Growing concerns about algorithmic bias and prejudice
o Areas that become the victim of ‘over policing’
Exporting Digital Authoritarianism: The Russian and Chinese models
Whereas China uses the technology for domestic censorship and surveillance, Russia uses the
technology to effectively eliminate potential (political) opposition and undermine democracies
abroad. Moscow is struggling to catch up with China’s high-tech model of domestic control. Rather,
Moscow’s domestic model is relatively low-tech when it comes to domestic surveillance, and hence
makes an appeal for those countries or actors that lack China’s economic prowess, human capital
capacities, and centralized state control. Whilst distinct in its precise execution, both countries have
developed models that enable rising authoritarian regimes to surpass potential opposition at home
while undermining democracies abroad.
China pioneered digital age censorship with its ‘Great Firewall’; a state-controlled Internet and
unprecedented high-tech repression deployed in Xinjiang. Hence, far from sparking a political
opening, within China the internet has been a valuable space for state censorship and surveillance. AI
also plays a role in China’s digital authoritarianism, to identify and track individuals across the
network. Chinese security services in Xinjiang have been able to monitor the movement and behavior
of its residents in paralleled detail, with Uighurs in particular being singled out. Yet what is so
troubling about Xinjiang is not just the tech-driven mass detentions and human rights violations. It’s
the prospect that Beijing will sell the technologies it has pioneered there to illiberal regimes abroad.
China is exporting its domestically produced technology at a low cost but yet high-quality information
and surveillance technology; making it likely that the export will only grow in the upcoming years. For
instance, Dubai has already begun to deploy Chinese technology as part of its “Police without
Policemen” program, an ambitious project to reduce crime, replacing policemen with video
surveillance and facial recognition technology.
Whereas China’s efforts focus on promoting a positive view of China (or repressing negative views
through various influence and intimidation techniques), the Russian approach aims to destabilize
politics and polarize societies to weaken them from within. This zero- sum view of international
relations has become part and parcel of Russian foreign policy.
The Russian government has made recent legal and technical moves which further tighten control,
including legislation passed this year to establish a “sovereign Russian internet.” This enabled the
Russian Government to isolate the Russian internet in whole or in parts from the global net. In the
long run, the Russian model (low-tech, and low-cost) may prove to be more adaptable globally as
emerging authoritarian regimes that cannot afford China’s high-tech model seek greater control over
domestic populations and influence.
The authors recommend that the United States and other democracies should:
Tighten export controls on technologies that advance digital authoritarianism
Sanction regimes engaging in digital authoritarianism and firms that supply them
Develop a competitive democratic model of digital governance with a code of conduct, to
enhance security in the USA and EU while still protecting civil liberties and human rights.
The US and Europe should work to develop common practices, rules and systems of digital
governance.
Increase public awareness around information manipulation, including funding educational
programs to build digital critical thinking skills among youth.