POLS3001 – Foreign Policy Analysis

Lecture 1 – Rational Choice Models of Decision Making

Lecture 2 – Behavioural Models of Decision Making
Lecture 3 – Evolutionary Models of Decision Making
Lecture 4 – Forecasting
Lecture 5 - Intelligence Analysis
Lecture 6 – Governmental and Societal Levels of Decision Making
Lecture 7 – Public Opinion
Lecture 8 – Economic Leverage
Lecture 9 – The Enemy
Lecture 10 – Culture and Identity
Lecture 11 – Structure of the International System
Lecture 12

1
 Lecture 1 – Rational Choice Models of Decision Making

The Rational Actor Model

 The rational actor model is a natural starting point for foreign policy analysis
 Yet it has been criticised on a number of grounds
 Supposedly ignores norms, culture, imperfect reasoning powers and information
processing
 However, I argue that this generally represents a misunderstanding of what the
rational actor model is

 We start off with a single decision maker

o E.g. Prime Minister, President
 He/she is faced with a limited budget of time and resources, a set of goals and a set
of beliefs about the world
o E.g. Albanese must make decisions about foreign policy as well as education,
economy etc
o Smaller country, smaller resources (military) – choose how to use our
resources as best as possible
o Beliefs about the effectiveness about certain and types of action, beliefs
about goals and likely actions of other actors (not necessarily beliefs about
right and wrong)
 This decision maker faces a set of alternative options
 He/she estimates the ‘expected utility’ of each option
 Rational actor model takes your goals as being exogenous
o Doesn’t concern itself with explaining where your goals come from
o Simply seeks to try and predict and explain what your behaviour is, or is likely
to be, based on what your goals are, given the constraints that you have in
terms of your time and your resources and your set of beliefs about the
world
 Criticism – when it goes beyond these basic assumptions, it ends up making some
specific predictions about how people are likely to behave both in foreign policy and
in the world more generally
o Prove to be things a lot of people disagree with

What is expected utility?

 Expected utility is simply the product of the utility you would gain from an outcome
weighted by the probability you attach to it, minus the cost
o Utility: how much you like a particular thing
 U(x)p(x) – c(x)
 For instance, the expected utility of buying a lottery ticket where the prize is $10m,
the cost is 50c and the chance of winning is 1 in 10,000,000 is
($10,000,000/10,000,000) = $1 - .5 = 50 cents
 What is the expected utility if the chance of winning is 1 in 100,000,000?
o ($10,000,000/100,000,000) = $0.1 - .5 = -50 cents
 What if the prize is $50m?

2
 o ($50,000,000/10,000,000) = $5 - .5 = $4.50
 What if the price of a ticket is $1.35?
o ($10,000,000/10,000,000) = $1 – 1.35 = -0.35 cents

Weighing options
 We are then said (assumed) to rank each option by its expected utility
 We choose the option with the highest expected utility
 The expected utility of the next best option to the one we chose is called our
‘opportunity cost’
o Rational choice predicts that as the opportunity costs of a particular decision
go up, our chances of doing it will decrease
o And if the opportunity costs go down, our chances of doing it will increase
 The trust ‘cost’ of our action is not just the monetary cost but rather the benefits we
would have obtained by pursuing our next best option
 We must also weigh our options at the margin – that is, what is the expected utility
of doing ‘more of X’ given where we are now
o Whatever money, time or resources we have already invested into X should
not matter

Risk attitudes
 Suppose I told you you can have $50 for sure, or we toss a coin – heads you win
$100, tails you win nothing?
 Same deal, but if the coin turns up heads you get
o $120?
o $130
o $140?
 Same deal, but if the coin turns heads you get
o $80?
o $70?
o $60?

 Risk neutral – situation where you

are happy to get X amount of
dollars for sure, or a lottery/gamble
whose expected value is X dollars
 Risk adverse – the amount of extra
money you would have to get from
the gamble in order to be willing to
give up the sense of certainty of
having that amount of money for
sure
 Risk acceptant – amount of money
you are willing to pay in order to
have the fun and thrill of risk and
gamble

3
Transitivity
 Oranges > Apples & Apples > Bananas => Oranges > Bananas
 These are known as transitive preferences
 These assumptions must hold, or we could be arbitraged out of everything we have

Completeness
 We have preferences over all the possible outcomes and can rank them

Ordinality vs Cardinality
 We can say Oranges > Apples > Bananas (for other people)
 But we simply can’t tell how much more you like Oranges than Apples
 Early (or proto) rational choice theorists like Bentham and Mill thought that you
could measure preferences on a cardinal scale (i.e. I get 10 utils of pleasure from an
orange, 8 from an apple, 5 from a banana)
 Now, however, we don’t talk about utils anymore and instead focus simply on
preferences – he prefers oranges to apples and apples to bananas

Revealed Preferences
 So how do we figure out what others’ preferences are?
 The only way to do so (by rational choice theory) is to infer them from the choices
they made in the past
o Reverse engineering their preferences from their past behaviour and then
projecting this into the future via some kind of formal model which takes into
account the reactions of other people and then tries to figure out what that
person is going to do

Individual rationality does not imply collective rationality

 Allison makes the mistake of confusing the rational actor model with the unitary
actor model
 But decisions that are rational for individual are often perverse for a group
considered as a unitary actor
 Groups composed of rational individuals do not have preferences which are rational
in the sense I have just described

Example – Brexit
1st Preference 2nd Preference 3rd Preference
Nicola No Brexit Soft Brexit Hard Brexit
Theresa Soft Brexit Hard Brexit No Brexit
Nigel Hard Brexit No Brexit Soft Brexit
 Each individual has transitive preferences – they are rational actors
 If these 3 individuals put it to a vote amongst themselves there is no outcome at all
o The outcome is entirely different depending on how you choose to arbitrate
amongst the votes

4
For the group
 No Brexit beats Soft Brexit
 Hard Brexit beats No Brexit
 Soft Brexit beats Hard Brexit
 No Brexit beats Soft Brexit
 An entirely non-transitive and therefore irrational rest of results/preferences
 Collective rationality doesn’t necessarily follow from individual rationality

Game Theory vs Decision Theory

 All the situations here involve one single decision maker possibly in the context of
uncertainty
o This is decision theory (branch of rational choice theory and of the rational
actor model)
 But in foreign policy, this is rarely the case
 We base our decisions on what we think other decision makers are going to do, who
are in turn deciding based on what they think we will do
o This is the domain of game theory

Basics of a game
 Players (at least two, could be more)
 Strategies (decisions they can take – defect, cooperate, etc)
 Payoffs (what does each player get in each outcome)
 Information sets (what do they know about each other’s’ strategies and payoffs)

Key contributions of game theory to foreign policy analysis

Nash Equilibrium
 Suppose two players are playing a game, how do you predict the outcome?
o For player 1, figure out what his ‘best response’ to every one of player 2’s
strategies. Do the same for player 2
o Draw the two best response correspondences on a piece of paper
o Where do they meet? This is each player’s best response to the other’s best
response. It’s a Nash Equilibrium. The predicted outcome of the game

Power of the Nash Equilibrium

5
  It’s unlikely anyone would deviate from a Nash Equilibrium since they can only do
worse for themselves by doing so
 It’s a very general tool for predicting social behaviour

Mixed Strategy Nash Equilibrium

 In many competitive situations, the Nash Equilibrium is not to pursue one sole
strategy with certainty
 If you do that, your opponent can exploit this knowledge to their advantage
 In that case, you should rationalise your choice of strategy so as to keep your
opponent guessing
 Two examples from sports – tennis serve and football penalty kick

Signalling
 Game theory also teaches us that sometimes we must take costly actions so as to
signal our type both to allies and competitors
 Our ‘type’ is the kind of actor we are – tough, loyal, competent, etc.
o Qualities anyone would want others to think they have
 Signalling is the process whereby we undertake an action which only someone who
really was tough, loyal, competent, etc. would actually do (because they are costly)

Collective action problems

 Have something where there is an action or a good that could be produced or
provided for a group of people (for an entire community), which would be of benefit
for everybody
o But it is disproportionately costly for one particular individual to do it, so
nobody wants to actually do it themselves
o Free riding – everybody wants somebody else to do it for them
o It never gets done, benefit is not provided

Alliances are collective action problems

 NATO obvious example
 E.g. crisis in Ukraine
o Country accused of not doing very much to help – Germany – wants someone
else to provide training, cut of gas, implement tax

What game theory says about alliances

 In alliances where one member is much bigger than everyone else, that member will
shoulder a disproportionately large share of the burden
 But this member will continue to do so, though they might complain about it,
because it’s in their interests
 This member might however attempt to use side-payments to induce other
members to contribute more (incentives in other domains)

Bargaining

6
  Bargaining games – situations in which there is a
potentially mutually profitable deal to be done
between two actors, they’re haggling over the
division of this good between themselves
 D1 – Walk Away or Breakdown Point for player 1
 d2 – Same but for player 2
 Shaded area is bargaining range – set of all possible
agreements that you could get
o Set of all agreements which are feasible, and which also leave both parties
better off than the situation in which they walk away and terminate the
negotiations completely
 Any agreement must be on the line
 Nash Bargaining Solution – at the tangent to this curve

What game theory teaches us about bargaining and negotiation

 The advantage in a negotiation goes to the side with

Critiques of rational choice theory

 Popular critique is that it ignores non-material motivations
 Another critique is that it endows people with superpowers of reasoning that they
don’t have
 A final critique is that it simply doesn’t have good predictive power

Non-material motivations
 No rational choice theorist would deny that human beings are motivated by things
other than money or survival
 Preferences can be composed of just about anything (eternal salvation, national
pride, victory of the working class) and, provided they are complete and transitive,
they are rational

However
 This only partly gets rational choice off the hook
 If preferences can come from anywhere, then it is hard to predict much from a
rational choice model per se
 It can make them hard to falsify
 The ‘revealed preferences’ approach can go some way towards solving this problem
but you need a lot of background information on the actors

Synthesis
 Where game theory rhetoric models are joined to in-depth knowledge of the actors
and their preferences, however, they can be very useful
 They sensitise us to the point we can’t just assume that the outcome will be
whatever the actors prefer, because they are conditioning their decisions on what
they think others will do

The Behavioural Critique

7
  The rationality postulates of game theory were essentially what Nash, or von
Neumann or Morgenstern thought a rational person would do
 Beginning in the 1970s, a new wave of economists and psychologists started
examining whether real people actually act the way rational choice theorists say they
should by carrying out laboratory experiments
 They found substantial deviations from what game theorists say ought to happen

The Rational Choice Response

 Rational behaviour tends to emerge over time as people learn by trial and error what
the best course of action is
 Rational behaviour is also likelier to emerge where the stakes are very high –
incentivises careful thought
 Rational behaviour is also more likely where agents interact with one another
repeatedly over long periods of time – so that each can learn each other’s’
preferences and the best response to them

Evidence
 In behavioural experiments, individuals tend to start off behaving ‘irrationally’ – but
end up behaving more rationally as they learn
 Ken Binmore, a game theorist, put his money where his mouth was and made the UK
Government AU$40bn designing a 3G Bandwidth Spectrum Auction

Summary
 Rational choice theory is the first cut for predicting, explaining and advocating
foreign policy decisions
 It’s often caricatured
o Some of the caricatures have weight, others less so
 Rational choice theory does not rely on the assumption of Spock-like rationality and
amoral selfishness
 However, rational choice models are most useful for the foreign policy analyst when
combined with solid background knowledge
 Rational choice models are likely to be right where:
o The stakes are high, so agents have incentives to think hard
 Key things to remember:
o Mixed strategies: it can pay to randomise to keep your opponent guessing

8
 Lecture 2 – Behavioural Models of Decision Making

Behavioural Approaches

Origins of the behavioural approach

 Foreign policy analysis draws on a number of other disciplines
 The rational actor model of decision making is drawn from economics
 Today we look at an alternative approach, which draws primarily from psychology
 This is known as the ‘behavioural approach’ because it seeks primarily to figure out
how humans make decisions inductively by observing how we behave, especially
through lab experiments
o Inductive reasoning – seeks to take certain things that are observed about
human beings, and how human beings actually make decisions based on this
 Starts off with the observation then using this to build a theory

Two types of ‘irrationality’

 The behavioural approach takes the rational actor model as a foil, a point of
comparison
 There are two aspects to this approach
o ‘Predictable irrationality’: how human beings in general diverge from rational
action
o ‘Individual irrationality’: how particular human beings diverge from rational
action

Founder of the modern behavioural approach

 Kahneman and Tversky were the first researchers to subject the rational actor model
to experimental testing
 They founded a field known as ‘behavioural economics’ on this basis and were
awarded the Nobel Prize for economics
 They claimed to find a number of systematic divergences from ‘rational’ behaviour in
their experimental subjects
 Many of these are important concepts in the analysis of foreign policy design
makinTheory
 Last lecture we discussed expected utility and risk attitudes
 I asked you what you would prefer – $5 for sure or a 50% chance of either $10 or $0
 Now what if I ask you the following question – you give me $5 for sure, or I toss a
coin and if it come heads, you give me $10, tails $0. Which do you choose?
o People’s risk attitudes are reversed when they are in the domain of losses as
opposed to the domain of gains
o So people are generally risk acceptant, but they want a higher expected
utility than a gain for sure

Risk Attitudes are Frame Dependant

9
  More willing to accept a gamble which
might actually result in us losing even
more money than we are to definitely
handing over something for sure
 Results in a risk curve
 Risk neutral (in both domain of gains
and losses) are the straight line
 Risk adverse in domain of losses + risk
acceptant in domain of gains – mirrored
sections
 Most people are risk acceptant in
domain of losses + risk adverse in the
domain of gains

Implication of Prospect Theory

 The implication of prospect theory is that we are risk adverse in the domain of gains
and risk acceptance in the domain of losses in general
 However, where probabilities are very low or very high, we behave differently
depending on whether we are in the domain of gains or the domain of losses
o High probability + Domain of Gains = Risk Aversion (100% chance of $9499 >
95% chance of $10,000)
o Low probability + Domain of Gains = Risk Acceptance (5% chance to win
$10,000 > 100% chance of $501)
o High probability + Domain of Losses = Risk Acceptance (95% chance to lose
$10,000 > 100% chance to lose $9499(
o Low probability + Domain of Losses = Risk Aversion (100% chance of losing
$501 > 5% chance to lose $10,000)

Implications for Foreign Policy Analysis

 Prospect theory has been proposed as an alternative ‘micro foundation’ for foreign
policy analysis and international relations (to rational choice theory)
 One important implication of prospect theory is the idea that leaders will be more
likely to make risky decisions (like going to war) when things are going badly (when
they are in the domain of loss)

Alternative explaining
 “Wag the Dog Effect” – gambling for resurrection
 If the war is going badly, and you gamble and you lose, this isn’t too bad
o Certainty of the end of political career if you don’t make the gamble vs the
possibility that you make the gamble and win so you end up resurrecting your
political career

System 1 and System 2

 Two different ‘systems’ in the mind posited by behavioural economics Kahneman
and Tversky
o Dual systems model

10
  System 1 is low effort, highly efficient for most purposes, disastrous for high level
cognitive tasks
 System 2 is good for high level cognitive tasks but demands lots of energy, and given
humans are conditioned to conserve energy, we don’t want this
 Engage System 1 for as long as possible

The ‘Cognitive Biases’ Literature

 Kahneman and Tversky’s research spawned a whole school of work into ‘cognitive
biases’ – systematic departures from ‘rational’ decision making
 Some of this research was carried out by K&T themselves, some by other
behavioural economists and psychologists such as Dan Ariely or Phillip Tetlock

Availability heuristic
 Cognitive distortion where you systematically overestimate the chances of
something happening if you personally can recall an example of it
 But the ease with which you can recall an example of something has nothing to do
with how likely it actually is – it’s partly driven by media attention and partly by
randomness

Probability neglect
 Ignoring shades of probability (i.e. buying a lottery ticket because of how great it
would be if you won, ignoring how unlikely it is)
 Treating very unlikely events as impossible, and very likely events and certain
 The three mental settings – impossible, 50-50, certain

Small sample bias

 Chances of unusual or unrepresentative events are higher when you look only at a
small sample of the data (e.g. 8 heads in 10 tosses are more likely then 80 heads in
100 tosses and so on)
 In foreign policy analysis, this means overgeneralising about trends from a small
number of events (Munich-Vietnam)

Sunk costs fallacy

 The Brumbies are playing tonight but it’s cold, wet and miserable. Do you go to see
the game if:
o You have already paid to watch the game and the money is non-refundable
o You haven’t bought tickets yet but there are still enough available

Sunk costs fallacy in FPA

 Induced to explain
o Vietnam – American presidents realised war was unwinnable, kept sending
soldiers, already spent so much needed to ensure these sunk costs had not
just been in vain
o USSR in Afghanistan
o Iraq
o USA in Afghanistan

11
Fundamental Attribution Error
 When I/my country behave badly, it’s because circumstances forced us to
 But, when you/my enemy behave badly, it’s because that’s the way they are

Halo effect
 If someone is good at one thing or did one thing well, we tend to assume they do
everything else well too
 Churchill made many strategic blunders, Chamberlain got lots of things right

Trade off avoidance

 Choosing one alternative over another only implies that, on balance, that alternative
is better
 Trade off avoidance means that we tend to downplay or ignore the downsides of our
chosen alternative

Confirmation bias
 Tendency to look for evidence in favour of our favourite hypothesis
 Instead we should be looking for evidence against it
 Various practices have emerged to help combat confirmation bias – red teaming,
devils’ advocates and premortems

Sacred values protection

 All communities have certain things which they hold to be sacred
 Enemies who transgress those sacred values can elicit an excessive response
 Individuals find it hard to bargain over things which they hold to be sacred (e.g. the
Some of the Rock in Jerusalem)

Importance of norms
 In the prisoner’s dilemma game, the rational choice is to defect regardless of the
decision of the other player
 However, when experimental subjects have played the game in a lab, cooperation is
actually quite frequent
 Similarly, in ‘public goods’ games, the rational actor model prediction is little or no
contribution
 Experimental subjects, however, have been found to contribute far more and in
many cases even to pay a cost to punish those who don’t contribute enough

The Dictator and Ultimatum Games

12
The Trust Game

Cross Cultural Variation

 Behavioural experiments have been carried out in a variety of cultural settings,
Western and non-Western societies, developing and developed
 There is significant variation among cultures in terms of how they play
 Westerners tend to be relatively high on trust, cooperation and altruistic
punishments, East Asians tend to view behaviour holistically, Eastern Europeans tend
to ‘anti-social punishment’
 The groups of people who act most like ‘rational actors’ are in fact certain types of
Amazonian hunter gatherer tribes

Really irrational?
 Some rationalists push back against the behaviouralists on the grounds that many
heuristics are in fact quite rational
 It’s not rational to wait for all available information before acting
 Some heuristics such as the sunk costs fallacy may have some rational advantages
 Are behavioural laboratory subjects really representative of how world leaders
would behave in real life?
 Many proponents of rational choice argue that behavioural experiments in decision
making are ‘gotcha’ questions, designed to catch people out

Individual Irrationality

Particular ‘irrationality’
 Long-standing cottage industry of people who psychologically analyse political
leaders
o Jerold Post one of the key figures
 Diagnoses at a distance of Donald Trump is latest in the genre
 However, political scientists and professional psychiatrists are sceptical

Psychiatrists on Trump

13
“His widely reported symptoms of mental instability - including grandiosity, impulsivity,
hypersensitivity to slights or criticism, and an apparent inability to distinguish between
fantasy and reality - lead us to question his fitness for the immense responsibilities of the
office” – Judith Herman, Nanette Gatrell and Dee Mosbacher, Harvard Medical School and
University of California, San Francisco

Leader psychology
 Scholars in this school believe leaders’ individual personality matters
 Some type of leaders could emerge systematically in some types of society (Rosen)
 Political power could of itself cause individuals to start acting in pathological ways
(Robertson)

Counterarguments
 Psychologists are not in a position to make a diagnosis of political leaders
 Psychologising one’s opponents can lead one to underestimate them
 Leader’s public persona may be a carefully crafted image
 Observational equivalence – appearance of instability may not imply real stability

Example – Saddam Hussein

 Prior to 2003, Saddam was one of the most distanced diagnosed world leaders in
history, including by Post
 An unexpected bonus from the 2003 invasion is that a vast trove of documents
about the internal workings of the Iraqi state under Saddam, including tape
recordings of his private conversations

What they found

 Saddam had a very naïve view of world and US politics
 Believed US was out to get him even when it was not (e.g. 1980s)
 Believed UN could constrain US in 2003
 Convened a conference to ‘celebrate’ Iraq’s ‘Great Victory’ in the Gulf War of
1990-1991

But…
 Saddam understood nuclear deterrence
 Saddam knew the US army would outfight the Iraqi army, though US casualty
sensitivity would bring him victory
 Not clear what better options he had in 2003; not clear in advance that 1991
would be as big a defeat as it was
 Not clear that he really believed 1991 was a ‘great victory’

2016 predictions about Trump

 Trump’s age implies greater propensity to initiate conflict
 Inexperience in government implies frequent mistakes
o Military training minus military experience heightens risk or conflict
 Antisocial orientation and lack of epistemic motivation imply aggressive but
incompetent negotiating style

14
Horowitz et al – the LEAD Database
 Horowitz et al sought to examine the question of whether leaders’ particular
psychologies influence their foreign policy behaviour
 They gathered extensive biographical data on world leaders going back to the early
19th Century
 This included a number of individual level variables such as:
o Birth order
o Family background
o Prior occupation
o Military experience

Findings
 Illness also matters – sick leaders behave more erratically
 Military experience has a complex effect – leaders with combat experience in
democracies tend to be less willing to use force (Eisenhower), leaders with military
but not combat experience (Bush) are more willing
 Experience coming to power as a rebel leader (Mao, Castro) makes one more risk
acceptant
 Political system has a mediating effect – combat experience in an autocratic system
makes one more likely to want to use force
o Old democratic leaders are more likely to start wars, as are young autocrats

Genetics
 It’s highly likely that our psychology and decision making are influenced by our genes
 However as yet ‘behavioural genetics’ is in its infancy – we still don’t know with
much confidence which genes influence which behaviour
 Moreover, even if we did know, it’s unlikely we’d get DNA from present or past
world leaders for analysis
 That said, we can speculate that leaders predisposed to particular types of behaviour
are more likely in some states than others
 Countries with frequent instability and violence are likely to ‘select for’ leaders with
high propensity for violence and risk taking

Conclusion
 Behavioural approaches represent a major alternative to the rational actor model as
a ‘micro foundation’ for FPA
 Behavioural experiments appear to show numerous departures from rational
decision making
 These systematic departures have been advanced to explain numerous foreign policy
decisions
 Similarly, leadership psychology has seen a resurgence of interest, with interesting
and useful predictions about leader behaviour
 Next session, we will examine another alternative approach to thinking about
decision making, different both from the rational actor and the behavioural
approach – the evolutionary approach

15
 Lecture 3 – Evolutionary Models of Decision Making

Evolutionary Approaches

So far…
 We’ve examined the rational actor model
 We’ve also looked at the competitor – the behavioural model
 Today we will be examining the evolutionary approach, in some sense a synthesis of
the other two

How evolution works

 Replicators – entities with the means to make copies of themselves
 Random variation (copying errors) in the transmission of the replicators from one
generation to the next
 Variants which are better suited to their environment survive and reproduce at a
greater rate

Types of replicators
 Genes (Dominic Johnson, Hugo Mercier)
 Strategies, decision making heuristics, ideas, norms (Gerd Gigerenzer, Vernon Smith)

Genetic replicators
 According to evolutionary psychologists, many of our ways of thinking are hardwired
into our genes
o E.g. propensity to be violence under some circumstances
 This is because they gave our ancestors some kind of reproductive advantage in the
prehistoric past
 Whether they are still adaptive or not depends on the extent of the match between
the modern and Pleistocene world
o World that most human beings lived in until recently, no cities, not states, no
writing, no agriculture, hunter gatherer society – “stone age brains”

Mercier
 Our thinking is guided by a number of genetically encoded modules
 Most of our decisions are made instinctively by these modules without our conscious
brain’s awareness
 Reasoning itself is, however, one of these modules
o Reasoning / thinking rationally is a subconscious module designed to fit a
purpose – serving the purpose to build coalitions with other human beings by
telling a story to them that is convincing, making them want to be an ally

Experiment: The Waiting Room Experiment

 Way of illustrating the purpose of our reasoning power
o Namely, to help us to portray a vision or image of ourselves to other people
as being desirable allies (competent, trustworthy, altruistic, etc)
 Have a number of experimental subjects divided at random to 2 groups

16
 o Group 1 – sitting in a waiting room by yourself and you hear somebody
screaming like they’re in trouble
 Do you intervene
 Interested to see if you go to find out what happens
o Group 2 – you hear someone screaming but you’re not alone
 With someone who is a confederate of the experimenter
 If there’s somebody else there, you’re less likely to help
 Reasoning people give for why they didn’t intervene
o Expected the other person to do it instead (but they never mention this
because it makes them look bad)
 Always looking to our reasoning faculties to justify our behaviour to others
o The motivation for our actions may not even be something we are
consciously aware of

The purpose of reason

 We think of the aim of reason as being to find out the truth
 But for Mercier this isn’t it
 For Mercier, reason exists to justify to others the decisions we make instinctively in
our own (or our genes) interests
 Reason is more like a defence lawyer or a PR consultant than a disinterested truth
seeker

Kahneman in a new light

 ‘Faulty’ heuristics and ‘biases’ can therefore be seen in a new light
 They are not ‘faulty’ at all
o They are doing exactly what they are designed to do
 But what they are designed to do is not what we thought

Example: confirmation bias

 Confirmation bias is seen as irrational
 But confirmation bias generally doesn’t apply to any old theory
 Specifically it applies to theories which serve our interests
 We do not suffer from confirmation bias with respect to theories it’s not in our
interest to believe

What can we do about it?

 Mercier believes that we have also evolved the ability to sport flaws in others’
reasoning
 Hence working in teams and groups is crucial to good decision making, as other
group members will shoot down faulty reasoning – act as checks and balances
 One way to institutionalise this process is through ‘red teaming’ which we will
examine in a later session

Johnson
 Similar to Mercier in that he sees ‘heuristics’ as hardwired genetic adaptations

17
  They only appear to be irrational because we do not understand their true
evolutionary function

Overconfidence
 Overconfidence is a common feature in human psychology
 It’s also often been seen as a key cause of war
 Sometimes overconfidence is quite remarkable – e.g. Saddam Hussein’s apparent
belief he won the 1991 Gulf War

Nothing ventured
 For Johnson, however, this is only part of the story
 Overconfidence encourages us to take risks which, if they come off, will pay off big
 Thus in the long run, although most of these ‘bets’ might fail, they will still be
adaptive

Churchill
 In retrospect, Churchill is considered a hero for his refusal to surrender to Hitler
 But, in reality, the idea that Britain could defeat Germany in 1940 was wildly
unrealistic
 Churchill was lucky – had the US and USDR not entered the wat, we would see him
as being a delusional blunderer

Sunk costs ‘fallacy’

 Is it really a ‘fallacy’?
 Consider
o Reputation for sticking with projects
o ‘Option value’
o ‘Breakthrough problems’

Heuristics as replicators
 As an alternative view comes to similar conclusions for similar but different reasons
 This is a view which sees heuristics themselves as replicators
 Rather than being ‘hard coded’ into our DNA as in Mercier or Johnson, heuristics are
socially constructed but then subjected to a process of natural selection

Smith – Ecological Rationality

 We spend a lot of time trying to figure out what the best decision is in a given
situation (constructivist rationality)
 But crucially, we subject these heuristics to the test of experience
 If the heuristics work well for us in a given situation, we keep them, otherwise we
jettison them
 If we stay in a similar environment for long enough, these heuristics can end up
being very highly adaptive
o E.g. agricultural techniques
 But they are only adaptive for the environment in which they arose

18
Example
 Which three out of these six cities are the (largest by population) in Brazil?
o Sao Paolo; Fortaleza; Rio de Janeiro; Belo Hoizonte; Brasilia; Belem
o You pick the three you’ve heard of (likely)
 Which three of these six cities are the biggest in Australia?
o Sydney; Canberra; Brisbane; Darwin; Adelaide; Melbourne
o Not useful as we are familiar with all the options
 Which of these six cities are the biggest in Kyrgyzstan?
o Osh, Usgen, Bishkek, Jalal-Abad, Naryn
o Not familiar with any of them

The “have I heard of it?” heuristic

 Works well with countries that you know somewhat well (e.g. Brazil)
 Unnecessary and maybe misleading with countries you know well (e.g. Australia)
 Doesn’t work with countries you don’t know about (e.g. Kyrgyzstan)

Gigerenzer
 Gigerenzer believes that human decision making capabilities have evolved to be
highly adaptive in most situations
 Contra Kahneman, he believes there’s nothing ‘irrational’ about them, when you
understand rationality to mean ‘getting the best result’

Example – the constant angle heuristic

 Maintaining a constant angle between themselves and the line of the ball coming
 Corelates with the physics of where the ball is going to land
o Proxys for this difficult process
 Brain substitutes for an impossible calculation for a heuristic

Gigerenzer’ critique of Kahneman

 Behavioural approaches are simply a ‘patched up’ version of a rational choice
 In developing a theory of decision making, we should start by observing how human
beings actually behave, not postulating a theory and then measuring deviations from
it

Commonalities between rational choice and behavioural economics

 Decisions are made in order to maximise a ‘weighted sum’ of utilities (e.g. p(x)U(x))
 The difference is simply over how the weights are constructed

Gigerenzer – the checklist

 Instead of making decisions to maximise expected utilities, what we tend to do is we
reach for a particular kind of checklist of things we need to do in a particular
situation
o And those checklists have been selected through this process of natural
selection
 This is in fact a way of making decisions that has evolved over human history
because it works better than the alternatives

19
Application to foreign policy
 Clearly there is something tot his way of thinking
 Many strategies – balance against power, maintain a reputation for strength – can
be seen as heuristics which have evolved over time
 Heuristics must come from somewhere
 They can be adaptive in many situations
 However…

Bush
 Said he based his decisions on his gut

What are the limits to snap judgements?

 Should we just always go with our instincts then?
 Under what conditions will our instinctive judgements lead us astray?
 How much is the modern international environment like the one in which our
decision making capabilities/heuristics evolved?

Differences
 Scale of modern societies
 Technology, including nuclear weapons
 Extent of international trade
 Complexity of government machinery

Commonalities
 Oscillation between conflict and cooperation
 Groups competing over resources
 Need to gain popular support
 Need to predict others’ motives and actions

Conclusion
 Evolutionary approach shows us how ‘rational’ decision making may come about
through anything but rational means
 It cautions us against simply dismissing behavioural heuristics as ‘mistakes’ or
‘biases’
 However, in the realm of foreign policy, it doesn’t give us clear guidance on when
and how we should rely on instinct
 Are our instincts adaptive in this particular domain? It’s not clear
 Also, what are the conditions under which instinctual decision making might lead to
better FP outcomes?

20
 Lecture 4 – Forecasting

Forecasting

Why forecasting is important

 Allocation of resources
 Ability to make timely decisions (e.g. Rwanda)
 Serves as a check on our understanding of the world – Fred and Jane’s lottery
 Governments devote significant resources to forecasting (e.g. Australia’s ONA)

Forecasting has a long pedigree

 Laplace believed that one could, in principle, forecast the rest of human history
 Karl Marx most famously forecast that history would culminate in the victory of the
proletariat
 Isaac Asimov’s books created the idea of the ‘psycho historian’ who could foretell
the future

But the record of forecasting is terrible

 Marx got it wrong
 Intelligence agencies got the fall of the Berlin Wall, 9/11, Arab Spring, Iranian
revolution, all wrong
 Economic forecasters missed the GFC
 Political forecasters missed Trump and Brexit

‘Dart throwing chimps’

 Psychologist Phillip Tetlock studied the record of hundreds of experts in making
predictions
 He found most do no better than chance, many do worse than chance
o Worse than chance – put money on the opposite of what they say
 If you want a prediction about Chinese politics, better to ask a specialist in Russian
politics and vice versa
o People forecasted better outside of their area of expertise

Why is forecasting hard?

 Our psychology
 Strategic interaction
 Preference falsification
 Complexity/non-linear dynamics
 Lack of incentives for getting it right

Our psychology
 Available heuristic
 Affect heuristic – ‘wishful thinking’
 Probability neglect – ‘three mental settings’
 Underrating randomness

21
Lack of incentives for getting it right
 Few forecasters are assessed on the accuracy of their forecasts
 Media prefer sensational statements which are more likely to be wrong
 Even forecasters who work for financial institutions can often defray the costs of
poor forecasters onto others

Strategic interaction
 ‘Holmes-Moriarty’ problem
o Your enemy can gain an advantage from predicting your actions, so you need
to be deliberately unpredictable (sometimes)
o Common knowledge of rationality
 We see this in, for example, Allied deception operations in World War II
 And of course what’s true of you is also true of your enemy

Preference falsification
 In dictatorial regimes, few feel free to express their real political views
 Even in democracy countries, some political views are more politically acceptable
than others
 As a result, the ‘true’ level of popular support for a leader is often unknown, hence
the stability of a regime is unclear

Complexity/non-linear dynamics
 Small events have large unforeseen consequences
 Assassination of Franz Ferdinand, 9/11, Able Archer
 These dynamics make it hard to forecast very far into the future

Advances in forecasting
 Econometric models / Artificial intelligence
 Prediction markets
 Crowdsourcing (e.g. the Good Judgement Project)

How does one assess a forecast?

 Forecasts should be given as probabilities
o Probabilities lie on the range (0, 1)
 1 means – ‘this event will certainly happen’
 0 means – ‘this event will certainly not happen’
 Brier score – (Outcome – Prediction)2
o e.g. I predict Biden wins with p=.7 and he wins, so my Brier score is (1 - .7) 2 =
(0.09)
o Best Brier score is 0, worst is 1
o Measure of the accuracy of the forecast

What makes a real forecast?

 Forecasts must be specific
o Numeric (75% not ‘highly likely’)

22
 o Clear what they refer to (military clash involving 100 or more fatalities not
‘instability’)
 Forecasts must be time limited
o A recession by 31st October 2019, not a recession ‘soon’

Why quantify?

 Verbal terms lump

together forecasts and
likelihood
 Hard to distinguish
between
 People disagree on what
these words actually mean

Gives you less wriggle room

 Suppose you say it’s likely Erdogan will still be President of Turkey in 2022. What
does likely mean? 55%? 75%?
 If he is, you’re likely to say you meant ‘75%’ vice versa if he is not
 ‘Vague verbiage’ forecasts give you the ability to make your record out to be better
than it really is

Why’s this bad?

 It gives outsiders too much confidence in your abilities
 It gives you too much confidence in your own abilities – prevents you from learning
how you could have done better

Ambiguity
 Chiefs of Staff told Kennedy that the Bay of Pigs operation had a ‘fair chance’ of
succeeding
 They thought they were telling him the chances of success were weak
 He thought they meant the chances were strong

23
Types of forecasts – Econometrics/Artificial Intelligence
 Econometrics – where you have a preconceived theoretical model of the types of
variables which are going to influence the particular outcome that you’re interested
in
 Ben Goldsmith – Atrocity Forecasting Project
 Jay Ulfelder – Political Instability Task Force
 Lockheed Martin – Worldwide Integrated Crisis Early Warning Project
 Reasonable performance with short lead times

Prediction Markets
 Intrade
 Iowa Electronics Markets
 iPredict
 Defunct internal prediction market in US Defence Department
 Based on the insight of giving people incentives to get it right
 Shallow market – not a lot of people can participate
o Vulnerable to being captured by a small number of people with lots of money
and potentially a stake in the outcome

Crowdsourcing
 Good Judgement Project – open invitation project to submit questions you’re
interested in and have people answer them
 Questions come from Government agencies, businesses, media etc
 Small number of participants have consistently very good predictive record
o They are known as ‘super forecasters’
 Open source – anybody can sign up for it

What makes a super forecaster?

 Cognitive style
 Do not try to fit the world into a preconceived theory
o ‘Hedgehogs’ try to fit the world into a preconceived theory – not good at
forecasting
 Understand probability and make fine-grained judgements
 Foxes – ‘know a little about a lot of things’
o No one overarching theory that will explain everything
o Wide range of intellectual interests, generalists
 No particularly strong correlations in terms of training, education, job – many super
forecasters are ‘regular Joes’

Super forecasting tips

 Fermi-ize: break questions down into component parts (what must happen in order
for this to come about?)
o Many of the questions seem so specific that there is no base rate to start
with (more applicable to really difficult questions)
o Breaking down the components and judging how likely each are to happen

24
  Be on guard against wishful thinking
o Tend to think that the things we want to happen are more likely to happen
than they actually are
o Keeping our own emotions in check – avoiding panicking, running with the
herd
 Looking for clashing causal forces
 Balance the inside and the outside views
o Say we want to predict whether Joe Biden is going to be re-elected in 2024 as
the president of the US
o Outside views – try and look at this case as an example of broader
phenomenon
 An equivalence class for Biden’s re-election campaign – left liberal
incumbents, American elections overall, mayoral elections
o The more general, the broader the equivalence class, more statistical power
 Lumping in a bunch of cases not so similar
o Inside view – what is it specifically about Joe Biden in 2024 that is different to
this equivalence class
 That would cause me to move my forecast up or down
 E.g. Age, who he’s facing, policies, international environment and
economy,
o Equivalence class – what is this case an example of? What is the base rate for
this case?
o Inside – what is different about this case that would cause you to push your
judgement up or down
 Constantly learn and update
o Make lots of different forecasts
o Make forecasts about specific variables at different points in time
 Look for ‘equivalence classes’ – similar past cases

2020 Question – will China have a military dispute with the US, Taiwan, Japan, Vietnam or
the Philippines?
 How would we go about answering this question?
 Define a military despite – an exchange of fire that results in more than 10
casualties on both sides, causalities being dead or wounded
 Look at the weak link – which of these countries is most likely to have a military
conflict in China
o Taiwan

Outside view / Inside view

 What is the baseline probability of any two countries having a military dispute?
(Equivalence class – pairs of countries in modern history)
o What is the overall rate of military disputes between countries? – answer is
very low, way less than 1% or 1/10000 (initial baseline from a prior
reading)
 What is different about Chinese relations with any of these countries? Does this
cause us to revise our estimates up or down?

25
 o How is it that China and Taiwan, as a dyad of countries, is different from
any randomly selected dyad of countries in modern history?
o How are China and Taiwan different from say, Peru and Finland?
 Much closer
 Extensive history – numerous disputes
 Both grow out of the republican period in Chinese history
 China claims Taiwan as part of their territory
 Taiwan has a Relations Act with the United States – lower possibility
of war between 2 countries, US acting as a deterrent
 Window of opportunity issue
 Both militarised
 All these reasons which would tend to mean that China and Taiwan
would have a higher risk of having a militarised interstate dispute
than two pairs of countries picked at random from the modern
world
 Cause us to revise our estimate up or down
 Shortcut in this case – which of these countries is most likely to have a dispute
with China?

My answer
 0.01 (very low) baseline probability of any two countries having a military dispute
in one year is very low
 Risk for China and any of these countries is higher than average
 Nonetheless, I believed that the distraction cause by Covid would leave any of the
countries named in no shape to launch a military adventure in 2020

What can I learn from this? (2020)

 China recovered from Covid much faster than I had anticipated
 It also perhaps saw the crisis engulfing other countries such as the US as an
opportunity to engage in saber rattling
 I did not consider the ‘opportunity’ mechanism

2020 Question – will Trump win the re-election?

 How would you go about answering this?

At the time the assignment was due

 Covid was beginning to run wild in the US (over 2,000 deaths on the due date of
the exercise, 33,700 new cases) but the worst was yet to come
 Biden had not formally been nominated but it appeared almost certain he would
be the nominee

My reasoning
 US Presidents in modern times are usually re-elected
o Especially if the economy is strong
 The economy had been strong until Covid hit the US hard
 The strength of partisanship in US means any party nominee is almost guaranteed

26
 at least 48% of the vote and almost half of the states, so it would inevitably be
close

Here’s the bottom line

 I gave Biden a 60% chance of winning
 I thought however that his victory would be more decisive due to Covid
 I also thought he would win Ohio and Florida but not Georgia
 I underestimated the strength of partisanship in American politics

2020 Question – will the UK strike a post-Brexit trade deal with the EU by the end of the
year?

My reason – Fermi-ization
 What would have to happen in order for this to happen?
 The EU and the UK would have to solve a number of seemingly intractable
problems, including fisheries and finance
 Even smaller trade deals, such as the EU-Canada deal a few years prior, tend to
take years and can be vetoed by any one EU member state
o Outside view

Covid
 At this point the UK and EU were both being hit hard by Covid
 Substantial amounts of public service time were having to be devoted to this issue
 I therefore judged the probability of a deal by the year’s end to be about 0.05
 I was wrong

Lessons
 Never underestimate the cynicism of Boris Johnson
 He had promised a deal but knew most voters wouldn’t care much about the
details
 He had the credibility with the Tory right to make concessions to Brussels
 So essentially, he got a deal by folding on most if not all contentious matters

Continued issues
 Forecasting so far into the future (even 5 years) is still hard
 Best we can do is straightforward extrapolation
 Was thought that the area of politics best suited to forecasting was the domestic
politics of advanced industrial democracies
 Influence of technology is impossible to forecast
o In order for a particular technology to have a major influence on our lives we
have to know that it’s feasible
o If it’s feasible, we need to know that there’s a market for it
o If we knew both of these things, it would already exist, it would not be in the
future

27
 o To the extent that technology has any influence at all in our politics or
economics means that there are inherent limits to our ability to forecast in
the long term

28
 Lecture 5 – Intelligence Analysis

Markus Wolf, head of the HVA (East German foreign intelligence service)
 “The courage and suffering involved in obtaining information really has nothing to do
with its significance. In my experience, the efficiency of a service depends much
more on the willingness of those who receive its information to pay attention to it
when it contradicts their own opinions”
 Taking the information that they get from a variety of sources and turning that raw
product into good analysis that gets policymakers to actually listen
o Especially when it contradicts their own prior views
 Analysis of raw intelligence

Richards Heuer Jr, CIA

 “Major intelligence failures are usually caused by failures of analysis, not failures of
collection. Relevant information is discounted, misinterpreted, ignored, rejected or
overlooked because it fails to fit a prevailing mental model or mind-set”

Example – Stalin and Operation Barbarossa

 Stalin had advance warning of the 1941 German invasion from multiple sources
o Richard Sage (‘most formidable spy in history’ – Ian Fleming)
o British Government – official sources
o Cambridge Five Spy Ring – unofficial sources
o Soviet sigint and imint
 Stalin refused to believe any of it
o For a long time believed Trotsky, not Hitler, was his main threat
o Intelligence failure – despite everything these agents did, Stalin wasn’t
prepared to believe it because it didn’t accord with his own prior

Sources of raw intelligence

 ‘Imint’ – image intelligence, photos and videos
 ‘Sigint’ – signals intelligence, signals and codes
 ‘Humint’ – information from human sources
 ‘Osint’ – open source intelligence, any type of intelligence acquired via sources for
which you did not need authorisation
o E.g. article in a newspaper, post on Twitter
o Not type of information but whether that information was supposed to be
secret or not

Sigint
 Interception of enemy communication, followed by decryption if secret
 Very old in intelligence – started with interception of diplomatic correspondence,
then radio traffic, now cyber hacking (reading emails or internal communications)
 Very useful if you can break enemy codes
 But you must be careful to protect the source (e.g. House of Commons debate in
1927 about Soviet subversion)

29
 o Not acting on intelligence on the basis that if they had acted on it, the
Germans would have changed it and would have stopped the stream of
intelligence coming in the future

Osint
 For all intelligence services this is by far the most common source of intelligence
about foreign countries – stuff we can all access
 This was true even for highly effective humint agencies such as Wolf’s HVA
 The value added comes from the analysis plus how osint fits into what is known from
covert sources
 The value of osint is going up very rapidly in modern world as so much information is
available
o Huge volumes of data on the internet
o Commercial satellites
o Social media
 Hobbyists making valuable contributions

OSINT in action – Bellingcat

 Bellingcat is an OSINT actor – a crowdfunded NGO run by the British journalist Elliot
Higgins and his network of collaborators
 Bellingcat uses primary analysis of satellite imagery and social media content to try
to answer questions that would traditionally be the domain of state intelligence
agencies

Example – who downed MH17?

 MH17 was shot down over Ukraine in 2014
 Russia was widely suspected, but denied responsibility – blamed Ukraine
 In one of Bellingcat’s finest hours, they used osint to prove Russian guilt

How did they do so?

 We know that MH17 was down by a Buk missile
 A video posted to YouTube showed a convoy carrying a Buk launcher through the
town of Snizhne just under 3 hours before the crash
 An Instagram post showed a Buk missile with the same serial number as that
which down MH17 moving in an army truck through Alexsayevka, Russia, one
month earlier
 The registration plate of the truck allowed Bellingcat to trace it to the Russian
Army’s 53rd Brigade near Moscow
o A Vkontakte post showed a Russian sergeant posing with trucks and
missiles

Humint
 Sexiest and most impactful intelligence source but also the rarest
 Fraught with problems – inserting one of your own national into a foreign country’s
decision making apparatus is actually near impossible for most countries

30
  Most humint comes from defectors, but should these be believed? Fraudsters,
double agents, fantasists, exaggerators
o Ideological defectors
 However, trustworthy humint is the Holy Grail of intelligence

Example – Oleg Penkovsky

 Penkovsky was a senior KGB officer who turned and worked for the British and
Americans due to his disgust at Khrushchev’s autocratic and reckless behaviour
o Walk in – walked into embassy and announced he wanted to be a spy
 He was most crucial in the Cuban missile crisis – he passed on to the CIA the details
of what Soviet missile sites look like – made it possible to interpret U2 imint
 Eventually captured and executed

Example – Oleg Gordievsky

 Like Penkovsky, Gordievsky was a senior KGB officer who turned to the British
 Suring Able Archer, Gordievsky alerted the British and Americans to the Soviet dear
of a nuclear attack
 He briefed both Mikhail Gorbachev and Margaret Thatcher (unbeknown to
Gorbachev) prior to their first meeting
o Helped persuade Thatcher and Reagan that they could ‘work with’ Gorbachev
 Exfiltrated from Moscow by MI6 when his cover was blown

Example – Rafid Ahmed Alwan (Curveball)

 Iraqi Chemical Engineer
 Claimed asylum in Germany, passed on information to the BND (German
intelligence) about Saddam’s WMD program
o BND passed it on to the CIA
 Alwan’s information about mobile chemical weapons labs formed part of the US case
for war
 But Alwan had made it up – he was fervently anti-Saddam and wanted to provoke a
US invasion

Example – the Cambridge Five

 Five British agents that had turned to the USSR
 They provided information which helped the Soviets in WW2, dismantled many
Western operations behind the Soviet bloc in the early Cold War
 However, initially Stalin strongly suspected they were double agents and discounted
their evidence

Problems of dealing with ‘Humint’

 You can be too trusting when dealing with sources (as the US was with Curveball)
 Or you can be too distrustful (as Stalin was at first with the Cambridge Five)
 What you ideally want is a Penkovsky or a Gordievsky, a genuine agent with access
to the highest levels of enemy decision making
 The Cambridge Spy Ring experience drove many Western intelligence officers (e.g.
James Angleton, Peter Wright) to suspect almost everyone was a Soviet spy

31
  How can we sort the wheat from the chaff (or worse?)
 Some suggestions to follow

Intelligence analysis and forecasting

 Part of intelligence analysis is forecasting
 Unconditional forecasting – what is likely to happen? E.g. will the Iranian
Government fall?
 Conditional forecasting – how is North Korea likely to react if we do X?

Well known intelligence failures

 Fall of Shah’s regime in Iran
 Fall of the Soviet Union
 Indian nuclear program
 9/11
 Saddam Hussein’s WMD program

Politicisation
 Can be obvious – political pressure on intelligence agencies to reach conclusions
 Can also be more subtle
o Promotions and hiring decisions
o Inconvenient conclusions more likely to be questioned
 But politicisation is by no means the only problem

Reasons behind intelligence failure

 Complexity effects – one small event can have outsized consequences
 Preference falsification
 Other side’s deception – may be specifically trying to deceive you by feeling false
information
 Other side’s self-deception – trying to fool themselves, or another agency, target is
other people within the same government
 Compartmentalisation
 Cognitive heuristics
o Groupthink
o Confirmation bias
o Failure to put oneself in opponent’s shoes
o Overrating one’s state’s own importance
 Small sample bias (“I know a man who”)
 Bias in favour of causes rather than randomness
 Fallacy of identity (Fischer): big events must have big causes
o Failure of the Spanish Armada can’t just have been caused by a storm
o OR covid starting from one small event

What can be done about it?

The 9/11 Commission Report

 Provided a number of suggestions about how to reform US intelligence agencies

32
Social science thinking tools
 Don’t select on the dependent variable
o That is, look at cases where an outcome occurred and where it did not occur
o Look at intelligence failures and intelligence successes
 How are the former different from the latter?
o Many ‘post-mortems’ look only at failure (e.g. 9/11 Report, Iran Post-
mortem)
 Look at relevant comparisons (equivalence classes)
o What happened in other similar cases?
o Relevant comparison cases may not be the same country at a different time
o Problem is though – what is relevant comparison? Here, theory must be your
guide
 Do not overlook negative evidence
o What evidence is there that suggests my chosen interpretation is wrong?
o For instance, what evidence was there that suggested Saddam Hussein did
not have WMDs?
o What were alternative interpretations of the supposedly supportive
evidence?
o Falsification

Heuers Table (from Omand, How Spies Think)

 Shows how to create a Heuers Table

 Putting ourselves in the world of an analyst who is being asked to assess whether
Iran is trying to get nuclear weapons
 First, come up with a set of hypotheses which are supposed to be mutually exclusive
o They can’t both be true
o Hypothesis 1: Iran plans to conduct nuclear weapons related experiments
(military purpose)
o Hypothesis 2: Iran intends to conduct nuclear research for civilian purposes
 Take the evidence that you have found from whatever source and put it in the
Heuers table

33
  Final part relates to the type of sources, credibility, relevance of source
 Found that every piece of evidence is consistent with Hypothesis 2, not every piece
for Hypothesis 1
o We go for the conclusion which has the fewest pieces of evidence that are
inconsistent with it
o Not the most pieces of evidence that are consistent with it, because that
could be driven by confirmation bias
o So Hypothesis 1

Red Teams – a digression

 Three different types of red team
o Plays the adversary in war games
o Plays the adversary to check security
o Team designed to probe weaknesses in the dominant argument
 When we talk about ‘red teams’ we mean the third sense
 Red teams in this sense are designed to mitigate confirmation bias
 This is what you should be doing as the ‘red team’ in the presentation exercises

Social scientific thinking tools – the hypothetico-deductive model

 Hypothetico-deduction:
o Derive ‘observable implications’ of your theory
 What would we expect to see if this theory were true?
o Make your theory falsifiable
 What evidence could provide your theory wrong?
o Identify alternative theories
o Construct discriminating tests
 What evidence would we see if our theory was correct, and the
alternative was not? (And vice versa)

Social scientific thinking tools – Bayes’ Rules

 Probability a hypothesis (A) is true, given some evidence (B)
o E.g. probability Trump is a Russian Asset = A; his appointment of Flynn as
National Security Advisor = B
 Prior – probabilistic belief that you start out with that a given hypothesis is true,
subjective
o Prior would be quite low (.01)
o Denoted by P(A)
o Probability of not A P(~A)
 Alternative hypothesis (which is mutually exclusive)
o He has a sick admiration of Vladimir Putin (.99)
 P(A/B) = P(B/A) * P(A) / (P(B/A) * P(A) + P(B/~A) * P(~A))
 “The probability Trump is a Russian asset given that he appointed Flynn as his
advisor (𝑃(𝐴/𝐵)) is equal to the probability he appoint Flynn as his advisor if he were
a Russian asset ( 𝑃(𝐵/𝐴) ) times our prior belief that he is a Russian asset (𝑃 (𝐴) ),
divided by the probability he would appoint Flynn as his advisor whether he was a
Russian asset or not (𝑃(𝐵/𝐴) * 𝑃(𝐴) + 𝑃(𝐵/~𝐴) * 𝑃(~𝐴) )”

34
  Update prior based on evidence
 Supposed Trump is a Russian asset, what is the probability that he would appoint
Flynn as his national security advisor? 0.85
 What about if B is true? 0.2

Social scientific thinking tools – sensitivity analysis

 Saying – assumptions make an ass out of you and me
 But in reality, all conclusions rest on some assumptions
o The key question is – are these assumptions reasonable? If not, are our
conclusions sensitive to these assumptions?
 Sensitivity analysis involves examining what you are assuming, asking whether other
assumptions are reasonable, and whether our conclusions would change if our
assumptions did

Application: Operation Overlord

Allied invasion of Europe in 1944 – ‘D-Day’

Where will the Allies invade?

 Two theories
o Normandy
o Pas de Calais
 Calais is close to UK, but heavily defended
 Normandy further, but less protected

For each piece of evidence:

 How likely would this be to occur if the Pas de Calais were the main target?
 What are some alternative explanations of the same evidence?
 How likely would this evidence be if Normandy were the main target?

Sensitivity Analysis
 How sensitive is this conclusion to our choice of priors?
 How sensitive is it to certain pieces of information?
 Which pieces of information are the most discriminatory?

35
 Aim of the exercise
 Helps to show the power of ‘priors’ – our initial belief in whether a theory is true
or false
 Shows how important some pieces of evidence are relative to others – by how
much they shift our priors
 Forces us to be explicit about what alternative theories there could be and flushes
out unexamined assumptions

Takeaways from today’s lecture

 Intelligence failures result from many of the same faulty heuristics that we find on
decision making more generally
 Intelligence is about discerning our adversaries’ intentions and capabilities, allowing
us to predict their actions and their responses to ours
 Good social science thinking tools are crucial in this endeavour

Remember
 Formulate your initial hypothesis – what is the case?
 Look for alternatives – what might be true instead?
 Form your priors – how likely is this theory? How likely are the alternatives?
 Update your priors based on incoming evidence – how likely would it be that we
would observe this evidence if our theory were true? How likely if the alternatives
were true?
 Strong evidence is that which causes us to update our priors considerably
 Ask ourselves why we find this evidence so strong

36
 Lecture 6 – Governmental and Societal Levels of Decision Making

Bureaucratic Politics

Last session
 We talked about the unitary rational actor model of FPA
 Few believe in the unitary rational actor model in its entirety, but many believe it’s a
useful simplification
 There are challenges to many of its tenets
o Today we will look at a challenge to the assumption that states are unitary
actors

The Bureaucratic Politics Model

 Associated with Graham Allison and his analysis of the Cuban missile crisis
 Key postulate – bureaucratic actors within states are the main actors we should
focus on
o State policy emerges from haggling amongst different agencies with their
own interests and worldviews
o A state is more like an unruly committee than an individual person

Examples
 Army vs Navy/Air Force over ‘Defence of Australia’ Doctrine
o Practicalities and priority Australia gives to its services
 Foreign ministries often accused of ‘going native’ – sympathising with foreigners
more than their own country (e.g. FCO & Brexit)
 US military more cautious about use of force than civilians, hews to Powell Doctrine

Why different interests?

 Government agencies want to increase budgets and power because this means more
prestige and money for people in them
 Those working in some agencies may have few alternatives should their agencies be
downsized
 Also, individuals may come to believe that their agency’s mission is more vital than
others (i.e. the Air Force really is the future of warfare, intelligence agencies are the
only thing stopping terrorist attacks)
 Noble corruption – threat inflation

Why different worldviews?

 Selection – do you join DFAT if you have no empathy for foreigners?
o How people select the agencies they want to work for
 Socialisation – other diplomats have a way of thinking which you emulate
 Experience – if you’re posted overseas as a diplomat, you may come to appreciate
foreigners’ point of view more than you would otherwise have done

Standard Operating Procedures

37
  Standard operating procedures are essential for complex organisations to work,
especially under conditions of high stress
 SOPs encapsulate years of hard won learning
 But they can lead to disaster…
o While a particular organisation or bureaucracy standard operating procedure
might be second nature to the people who work in that organisation, it might
not be second nature to people who are outside
o Clashes
o Could also be that the SOP which organisations have might impose certain
rigidities, limits on your ability to make decisions

The Railroad Theory of WW1

 Russia wished to mobilise against Austria to protect Serbia, but not Germany
 But the Russian railroads’ SOP meant that Russia had to mobilise against both or
neither – the Tsar chose both
 German General Staff’s SOP (the Schlieffen Plan) required they then invade Russia’s
ally France to knock it out of the war
 What started off as a local dispute in the Balkans, spread via two rigidities that were
built into the decision making procedures of Russia and Germany

Even worse
 1983 Able Archer crisis – Soviet leadership fears President Reagan will launch a
nuclear first strike
 Soviet Colonel Stanislav Petrov receives radar reports of five incoming US ICBMs
 Soviet SOPs are that is he concludes they are real, Soviets much launch a
counterstrike
 Petrov concluded they were not

How can bureaucratic actors shape foreign policy?

 Sheer complexity allows bureaucrats to pursue their own agendas
 Bureaucrats have informational advantages over the political leadership
o Informational advantages, know more about what goes on in their area of
responsibility
 Bureaucrats can shape the agenda by drawing up only plans they support
 Bureaucrats can leak to the press, political opposition

How can political leaders respond?

 Be aware of key departments’ SOPs and language (e.g. Kennedy and the US Navy)
o Reduce information asymmetry
 Play different agencies off against each other (especially common in autocracies)
o E.g. Iraq under Saddam all armies report to him
 Employ people from outside the chain of command (special advisors) to police
bureaucracy
o Regain control over policy agenda
 ‘Blue dye’ test for leaks

38
 o If you suspect a public servant is leaking information, tell this person a
specific rumour and see if it comes up in news

Limits of the bureaucratic politics model

 Like all models, it only explains part of what’s going on
 Even the most powerful bureaucratic agency can’t ignore strong structural
constraints or public opinion on issues the public really cares about (e.g. Brexit)
 Theory is quite US-centric
o Explains less when applied to other types of states
o Public service generalists have less incentive to defend bureaucratic turf

Society-Based Theories

Societal factors shaping FP

 Political structure
o Democracy/Autocracy
o Types of Autocracy and Democracy
 Economic structure
o Type of market economy
o Politically influential sectors

Democracy/Autocracy
 Key distinction in comparative politics
 Democracies are characterised by competitive elections with universal suffrage and
individual rights
o Polity score
 Autocracies are states which do not meet one or any of these conditions
 Some also posit the existence of ‘mixed regimes’ which combine elements of both
democracy and autocracy

Consequences for foreign policy

 Democracies fight less with other democracies (democratic peace theory)
 Democracies tend to be easier to deal with than autocracies in other ways
o Free press, political debate makes it easier to know what democracies want
o Democracies tend to renege less on commitments
 Checks and balances in a democracy

Distinctions amongst democracies

 Parliamentarianism/Presidentialism
o Powers of head of state relative to legislature
o UK, Australia, Canada, Germany, Ireland vs US, France
 Single member vs multimember districts
o Former encourages two strong parties and single party government
o UK, US, Australia vs Germany, Israel

Consequences for foreign policy

39
  In presidential systems, foreign policy is usually reserved to the President although
the legislature has a certain amount of say
o Presidents often negotiate a deal, then must ‘sell’ it to the legislature
 In multimember district systems, coalition government is the norm
o Small coalition parties can have an outsized effect on foreign policy (e.g.
Kosovo War, Israel)
 Single member – formation of 2 strong dominant parties geographically
o Fewer decision makers to deal with

Distinctions amongst autocracies

 Personalist – based on one individual and his family
o No real ideology, basis in national tradition or ruling party (e.g. Syria, North
Korea)
 Single party – based on a single governing party with a somewhat clear ideology (e.g.
China, Vietnam)
 Traditionalist – monarchical or religious
o Grounded in a real or imagined national tradition (e.g. Saudi Arabia, UAE,
Jordan)

Consequences for foreign policy

 Traditionalist / single party regimes are more predictable than personalist regimes
o Certain number of checks and balances, consensus
 Single party regimes resemble democracies in some respect, though they are less
transparent
 Unsurprisingly, the personality of the leader is most important in personalist regimes

Economic structure
 Almost all states these days have some kind of market economy
 But there are vast differences in how these market economies are organised
 Moreover, economic structure overlaps with political structure in ways which are
important for FPA

Market economies
 Companies compete for business; state is at least in theory neutral between them
 Private companies not supposed to influence or be influenced by state foreign policy
 Market economies are usually also open to foreign trade and investment
 Many scholars credit economic interdependence and globalisation with reducing the
prospects for war

‘State capitalism’
 Many Middle Income Countries practice ‘state capitalism’ – market economies but
with large state owned enterprises or sovereign wealth funds
 The interests of these organisations often shape state foreign policy – these
organisations can also often be tools of state policy (e.g. Rosneft)

Politically crucial sectors

40
  Some businesses/sectors can be politically crucial for two reasons
o They generate significant employment/wealth in politically crucial parts of
the country
o They generate a significant party of the country’s wealth overall
 Politicians cannot afford to ignore sectors/ businesses like these

Examples
 US automobile industry (Ohio, Michigan, etc)
 US armaments industry (substantial portions of the country)
 French/Japanese agriculture
 Industries like these are politically untouchable
o Foreign partners cannot expect deals that go against their interests
 Mining industry in Australia
 City of London in the UK
 Oil industry in many countries
 Generate so much revenue and employment for the government that their concerns
cannot be ignored

Consequences for foreign policy

 Influential sectors lobby for better relations with important markets (e.g. City of
London against ‘hard Brexit’)
 They can also lobby for more aggressive foreign policy (US defence industry) or
campaign for tariff or other types of barriers against foreign competition (US auto
industry)
 But economically influential sectors don’t always get what they want, especially
when ranged against other economically important sectors who want to do
something else (e.g. US retailers & financial services vs manufacturing and defence
industry over China policy)

Conclusions
 Both economic and political structures exert important influences on foreign policy
 Democracies and autocracies behave in different ways, as do different types of
democracies and autocracies
 Similarly, politically crucial economic sectors and businesses exert outsized influence
on foreign policy decision making

41
 Lecture 7 – Public Opinion

Public Opinion

Public Opinion and Foreign Policy

1. Does public opinion matter for foreign policy?
2. What issues are most salient? What does the literature say about them?
3. What does the literature have to say about Australia specifically?

Key considerations
 Public has little incentive to inform themselves about politics in general
 Public has even less incentive to inform themselves about foreign policy,
consequences are remote
 Political leaders generally don’t respond much to public pressure over foreign policy
(Dowding, Martin et al 2015)
 Yet, if and when foreign policy crises do arise, public opinion is very important

Issues in public opinion and foreign policy

 Support for military conflict, especially as a function of casualties (‘casualty cringe’)
 Support for free trade/globalisation
o Rise of Trump, Brexit
o Focus on foreign economy policy
 (In the Australian context especially) favourability towards other countries, especially
the US and China

Public impact on FP
 Bush 2004 victory largely due to war on terror
o Projected an image of being strong and tough, at that crisis point (when
American public felt themselves to be under threat), key to his electoral
appeal
 Conversely Obama’s nomination and victory in 2008 largely result of
mismanagement of Iraq by Bush Administration
o By 2008, had become apparent to US public that the war in Iraq was a
disaster, as was war in Afghanistan
o American public viewed as a wartime situation fumbled by Republicans
o GFC also in operation, advantage for Obama
 Australian PMs frequently consult pollsters over war decisions (e.g. McMahon and
Whitlam over Vietnam)

Main schools of thought

 The ignorant public (Almond-Lippman consensus)
 Casualty sensitivity (Mueller)
 Pretty prudent / roughly rational public (Jentleson, Feaver & Gelpi)
 Partisanship above all else (Berinsky)

Almond-Lippman Consensus

42
Members of the American public were asked where Ukraine is, 2014

 Public don’t care about foreign policy, so they are badly informed about it
 They have no real ‘opinion’ about it
 Public have no input into foreign policy decisions and this is a good thing

Casualty consensus
 Public opinion turned against wars like Iraq and Afghanistan after (in a historical
perspective) very few casualties
 Phenomenon first noted with respect to Korean and Vietnam War
 Became axiomatic – public will turn against wars as causalities mount
o Even, the public won’t accept any casualties at all
o Consequences for types of military engagement

The pretty prudent public

 Voters make a rough calculation of the costs and benefits of foreign conflicts
o Expected value calculation
 Gain x probability – cost
o Making an implicit calculation
 Casualties are important as they factor into the costs, but prospective benefits
matter too
o Extent to which they are willing to pay these costs of war hinges on two
things:
 Chances of success – whether you think the war is likely to achieve
the goals
 Goals of the war – whether you support the goals of the war
 Chances of success and the goals of the war can counteract casualties
o Credible theory of how you win the war that can be clearly explained to the
public – more likely to support the war
 Through Feaver’s influence, this animated part of the Bush administration’s ‘surge’
strategy in Iraq
o If casualties are mounting and you are a policy maker that wants the war to
continue, you must persuade the public that the prospects for success are
increasing

43
 o Strategy that hinged on increasing number of US forces in Iraq – more US
troops, more control, more likely to defeat insurgency – change in tactics
 9/11 – a turning point in public opinion, desire for revenge

Partisanship
 Partisan identity is the ‘DNA’ of US politics
 Berinsky claims US public motivated by partisan cues not by facts on the ground
o Partisan cue – signals you are getting from leaders of your party or media
personalities
 Casualties/prospects for success move elite opinion, which then moves public
opinion
o Influential opinion leaders (could be media or politics) are then transmitting
their views about foreign policy to their partisans
 Iraq is an example
o Democratic opinion leaders united in favour of war because they were
frightened if they opposed the war, they would be branded unpatriotic
o Over course of war, democratic leaders and media personalities came to view
the war as being lost – transmitted this view to democratic voters who then
transmitted back to democratic politicians
 Not clear how much this travels outside the USA

The ‘Bottom-Up’ Theory (Kertzer and Zeitzoff)

 How can elite cues explain?
o Why US Democrat voters take a consistently pro-Palestinian stance?
o Why public in most NATO countries opposed the war in Afghanistan long
before any mainstream parties did?
o Why publics in many EU countries oppose the euro/EU but no mainstream
parties do?
Social Networks
 Individuals form their views about world politics based on general psychological
principles
 They get information from like-minded individuals about particular events
 This is increasingly true as the media landscape fragments – the ‘information
environment’ has changed
o Algorithms support confirmation bias
 Opinion leaders being shaped by this more than they are shaping this

Views of globalisation/free trade

 According to economic models, views of free trade should follow personal economic
interest
 High skilled people in rich countries should be in favour
 People dependent on exporting industries should be in favour
 Locally abundant good in Australia is skilled labour, unskilled labour is locally scarce
o Open trade, export product of labour to other parts of world with scarce
skilled labour

44
Reality
 Opinion on globalisation/trade has little to do with personal economic interest
 Instead, main drivers are ideology/identity
 Two types of people overwhelmingly oppose free trade/globalisation
o Conserve nationalists – Trump, Clive Palmer
o Left wing socialists

The Australian context

 Variety of sources for Australian public opinion on foreign policy – AES, ANU Polls,
Lowy Institute Australia and the World Polls
 Lots of work has been done to understand what drives Australian opinion on foreign
policy

Australia and casualties

 Brown – Australia uniquely averse because of Gallipoli legacy
 Friedman – Australia rather insensitive to casualties
 My research – Australia quite similar to US
o More casualty sensitive in the past, but still, prospects for success matter

Australia Survey Experiment

 Across different scenarios, in different groups, support for war

 Consistent with American literature

Australia and Afghanistan

 When causalities were increasing substantially, public opinion was staying more or
less the same
 Support for withdrawal fluctuating the whole time period when casualties were
increasing

45
ANZUS
 Alliance enjoys consistent majority support
 Main predictors of support – education, ethnicity, political ideology and age
o Higher levels of education tend to be opposed
o Conservatives in favour, older people in favour
 External shocks in the form of wars or terrorist attacks increase support for the
alliance
o Tend to increase support for alliance
 However, alliance is less popular when Republicans are in office
o Trump is especially toxic

Australia and China

 China is Australia’s major trading partner by far
 Yet also has a very different political system and culture to Australia
 Australian opinion towards China fluctuates from year to year
 Gut favourability of the US usually much higher
 Opinion towards China largely shaped by education and attitudes towards outsiders

Conclusions
 In general, the public does not care or have much influence on foreign policy
o Almond-Lippman consensus
 But in crises and war, the public can be crucial
 Public in most countries more casualty adverse than before, but casualties are not
everything
 Australian public still more favourable to the US than China, but unclear if this will
last

46
 Lecture 8 – Economic Leverage

The new frontier?

 Increasingly, scholars believe direct war between great powers is unlikely
 The main reason is the danger of nuclear escalation
 But the great powers still disagree over a number of issues, so how can this be
resolved?
 Some believe economic coercion is the new key

Forms of economic coercion

 Sanctions – uni, bi or multilateral
o Unilateral – US onto Cuba
o Multilateral – US, EU, Australia etc onto Russia (partial, but can be universal
like UN sanctions)
 Foreign investment
o Countries may threaten to remove foreign investment
o Or specifically invest in countries to have leverage over them in the future
 Aid
o Provision of foreign aid – designed to help another country to develop
o Some countries use aid as a tool of economic coercion – US foreign aid to
Egypt, Pakistan
 Cyber attacks
o Malware
o Used to damage physical infrastructure – Operation Olympic Games
o Used to gain intelligence
 Monetary and fiscal policy
o Monetary policy – the government’s position with regards to interest and
exchange rates, tool to manage economy
o Fiscal policy – related to how government manages its budget, surplus or
deficit
 Market access
o Being able to export goods to a foreign market

A few examples to think over

 In the following situations, there are two actors
 They are dependent on one another economically to some degree or other
 They want something from each other
 Who is more likely to get what they want?

1. Brexit
 UK voted in 2016 to leave the EU
 UK Government wanted access to European markets without allowing free
movement of EU nationals to the UK
 EU unwilling to grant full access to European markets without freedom of
movement
 Believed that Britain had more leverage because EU wants to export to UK

47
 o But UK is a smaller market than EU therefore EU had leverage

2. Scotland
 Had Scotland voted for independence in 2014, an independent Scotland wanted to
maintain the British pound as its currency
 This would mean the Bank of England would have to act as lender of last resort to
major Scottish commercial banks
 Different currencies in Scotland and England however would hurt both countries’
economies
 Who has advantage? Scotland – they have less to lose

3. Chinese holdings of US debts

 China is the major holder of US sovereign debt
 Chinese purchases of US debt allows the US to consistently run budget deficits
 But the purchases also provide a safe haven for Chinese savings
 Who has advantage? Mutually assured economic destruction
o Ultimately, US has advantage because they would be more easily be able
to find alternative purchasers of its sovereign debt than the Chinese would
be find alternative asset classes that would offer as favourable a risk-return
ratio as the US

Conventional wisdom about economic pressure

 Advantage lies with the lender and the seller, not the debtor and the buyer
o But there is a symbiotic relationship as each gains something from the
relationship
 Hence concern over Chinese purchase of US debt (in US), dependence on Middle
Eastern and/or Russian oil and natural gas

JM Keynes
 “If you owe someone $1,000, they own you. If you owe someone $1,000,000, you
own them.”
 Dependent on you repaying that money
 Relative opportunity costs of closure

The Arab Oil Embargo

Background
 Israel defeated Arab states in the Yom Kippur War largely due to US assistance
 Arab states decided to wield the ‘oil weapon’ to punish US for its support of Israel
o Fungibility
 Main Arab oil producers embargoed oil sales to the US (and Holland)
o Resulted in higher oil prices, caused inflation

How’d that work out?

 European and Japanese companies bought oil which they sold to the US
 But supply shock led to a global economic downturn

48
  Global economic downturn led to reduced demand (and hence price) for oil
 Also sparked conservation measures in oil importing countries

The real lesson

 The party with the advantage is the one whose opportunity costs of disruption are
lower
 Disruption of an economic relationship hurts both parties, the key question is who it
hurts more
 If there are more willing lenders out there than borrowers, this is the borrow
 If there are more willing sellers than buyers, this is the buyer

David and Goliath

 Generally, the advantage lies with the bigger actor
 Canada loses more from a disruption to trade with the US than the US would
 Both American and Canadian policymakers know this, and know that each other
knows it
 Hence, the US will tend to get more of what it wants than Canada will
 This explains why smaller countries often group together in trade federations (e.g.
EU, Mercosur)

Exception
 If you have something the other state needs, others may also need, and no one can
get anywhere else, then you’re in luck
 But cases like this are quite rare
 E.g. Pine Gap

Likely winners
1. The EU
2. The rest of the UK
3. The USA

Sanctions
 Empirical record of sanctions very mixed
 Very little evidence that sanctions work in the sense of producing policy concessions
or leadership change
 But sanctions are popular as a bloodless alternative to military action

Why sanctions are an imperfect tool

 Need to coordinate amongst multiple states (e.g. South Africa example)
 ‘Sanctions rents’ may enrich and entrench the political elite (e.g. Iraq)
 May generate public support for targeted regime

Model of sanctions
 Idea that sanctions only work when there are threatened, but not imposed
 Logic of this model – comes from Daniel Drezner

49
  Point where sanctions are imposed – bad outcome for both the sender and the
target
o Lose trade with each other
 One of two things will happen
o Either the sender will threaten to impose sanctions
 And if the target values continuing its behaviour less than having the
goodwill of the sender, then it will back down
 But if the target and sender stand firm, sanctions being imposed will
cause a stalemate because only a target values their behaviour – not
been able to use threat successfully to get the target to back down
 E.g. US and Iran/North Korea – not a big history of trade

The Sanctions Paradox (Drezner)

 States with high prospective future conflict are unlikely to be vulnerable to sanctions
 Similarly, sanctions unlikely to work where opportunity costs gap is small
 Sanctions only likely to work where there are low expectations of future conflict AND
high disparity in opportunity costs

Example: Sanctions on Russia in 2014

 Following Russia’s seizure of Crimea in 2014, sanctions were placed on Russian
entities in the following areas
o Energy
o Defence
o Finance

Energy sector
 Sanctions on the energy sector hurt Russia’s ability to exploit new sources and use
the most advanced technology
 However, after an initial shock, Russia adapted

How did Russia adapt?

Substituting domestic for foreign technology

50
  Substituting Asian (especially Chinese) capital investment
 Substituting Asian (especially Chinese) target markets

Defence Industry
 Sanctions significantly disrupted the Russian defence industry
 However, again the Russians adapted, substituting domestic production for
foreign imports and shifting to closer defence industrial relations with India and
China

Finance
 Financial sanctions damaged the Russian economy in the short run, but again
Russia adapted
 Russia introduced informal capital controls to compel Russians to repatriate
money from overseas and intervened to support the banks
 Russia also diversified sources of capital to Asia

Overall
 Putin’s response to sanctions was essentially to reallocate resources from
consumers to energy and defence, the cornerstones of Russian power
 Living standards for the average Russian declined but this is not Putin’s concern

Aid
 Aid is often used as a tool of influence
 Very few countries give development aid solely for humanitarian reasons
 Biggest recipients of US aid – Egypt and Israel
 Cut off of aid can be a major tool of leverage

Monetary and fiscal options

 Soft loans can also be a form of economic coercion
o Soft loans – below market interest rate, subsidised
 Loans can be withdrawn, or collateral might be demanded
 Some scholars and policymakers have raised such concerns over the Chinese ‘One
Belt, One Road’ initiative (‘debt trap diplomacy’)
 Others consider this overblown

Example – the Euro

 Germany is a major exporting economy
o It is geared to running a large trade surplus
 The individual European currencies were merged into the euro at a rate which was
good for German exporters (i.e. DM was too low)
 The Eurozone thus provides Germany with a well-priced market for its export
goods

The Eurozone Crisis

 As a result, goods from Southern Europe especially are uncompetitive with the
Germans

51
  This is a major cause of the Eurozone crisis of the early 2010s
 But it was in Germany’s interest to maintain the Eurozone by bailing out Southern
European economies
o These economies recognised the euro was better than the alternative

Upshot
 The EU and Eurozone are thus a good example of the use of economic instruments
as an alternative to force
 By the use of monetary and fiscal instruments, Germany has a strong market for
its exports without any suggestion of a resort to force

Foreign investment
 Foreign investment is overtaking trade in goods and services as the main form of
globalisation
 Instead of exporting goods to a foreign country, you buy a stake in a foreign business
 Such investment can either be ‘portfolio’ or ‘FDI’
o Portfolio – if you own less than 10%
o FDI – bigger stake, bigger say in day-to-day running of business

FDI as a tool of statecraft

 Many institutional investors have close relations with their home government
 Sovereign wealth funds are one example
o These are quasi-statal organisations
 Often represent commodities exporters, rarely democratic (Norway one exception)
 Debate over whether they pursue fully financial objectives

Example – China-Taiwan

State banks
 State banks are another quasi-statal financial institution
o An ordinary bank which is majority owned by the government
 Again, these can be more easily used as an instrument of state power
 Russia and China, for instance, have powerful state banks
o Governments happy for them to trade at a loss given that they will provide
the government or follow the government’s foreign policy and ideological
objectives
 Western countries usually divested from state banks in the 1980s
o E.g. Commonwealth Bank before it was privatised in 1980s

Cyber attacks
 Cyber actors can attack foreign companies as a way to weaken competition, steal
secrets or even gain a stronger negotiating hand
 One example is the persistent spate of cyber-attacks on Rio Tinto when a Chinese
company attempted to take it over

Market Access

52
  Returning to the EU
 The prospect of accession to the EU was a major impetus to structural reforms in
Eastern Europe (e.g. Poland)
 The prospect of a closer relationship with the EU and possible future accession is a
major tool for spreading Western European influence further east (e.g. Ukraine,
Turkey, Serbia)

Farrell and Newman – ‘Weaponised Interdependence’

 Optimistic 1990s and 2000s globalisation proponents believed that it reduced the
scope of great power competition
 Because the world was becoming more interconnected, the thinking went, no one
would be overly dependent on anyone else
 Moreover, no state would be able to use economic leverage over anyone else

Economic Networks

Centrality
 It’s a major finding in network science that a small number of ‘nodes’ have lots of
connections and a large number have few
 Some ‘nodes’ therefore have an outsized influence on the network as a whole
o This is a ‘high centrality’ network
 Many aspects of the global economy – especially finance and communications – are
like this

Global Financial Network

Argument
 Since the US, UK and EU are unusually well connected in terms of finance and
communications, Farrell and Newman argue they have an economic power beyond
even that of their GDP
 They can use their network power both to observe what other states and actors are
doing (the ‘panopticon’) and coerce them (the ‘chokepoint’)

Example – SWIFT
 SWIFT is a payment system designed to facilitate transactions between international
financial institutions, established by US and European banks
 It is almost indispensable for any bank looking to do business globally
 Because of the US/European basis of SWIFT, Western Governments can use it to
monitor transactions and cut off access

Example – PRISM
 Because most web traffic is hosted by US based cloud servers, the US Government
has used it to monitor the communications of non- US citizens
 By seeing who is talking to whom, even if they don’t know who is saying what, the
US Government can get an idea of who is collaborating and possibly for why

Counteracting

53
  Powers such as China and Russia are aware of the network power of the US, UK and
EU
 Consequently, they are working to counteract it
 A major goal of the Belt and Road initiative is to establish Chinese standards
overseas and give China more network power
 Russia aims to establish the Eurasian Union as an alternative to the EU, with a
blockchain payment system to bypass SWIFT

Conclusions
 The party which loses less in the event of a disruption is generally the one with the
advantage
 This is not always the seller or the lender
 Sanctions are a generally ineffective tool of statecraft
 Sanctions usually work best when threatened but not imposed
 Yet sanctions are by no means the only tool of economic statecraft
 Other tools include aid, monetary and financial policy, FDI and leveraging network
power via ‘panopticon’ and ‘chokepoint’ strategies
 A major part of modern great power competition lies in the attempt to create and
wield tools of economic statecraft

54
Lecture 9 – The enemy

Quote
- “in football, everything is complicated by the presence of the opposing team”
– Jean Paul Sartre

Why we need to understand our enemy

- Anticipate how they will react to our moves
- Anticipate what they will do and how we can best react
- Accurately gauge their intentions
- Win them over?

Why we misunderstand our enemy

- Fundamental attribution error
- Perceptions of excessive centralisation
- Own side bias – the elephant and the rider
- Simulation theory – putting oneself in the adversary’s shoes

Fundamental Attribution error

- Our good deeds reflect who we really are – our bad deeds are a regrettable
response to external circumstances
- Their bad deeds reflect who they really are – their good deeds are either a) a
response to our pressure or b) a response to some other external
circumstance

Perceptions of centralisation
- We have little problem seeing our own policies as emerging from a messy
bargaining process among multiple actors
- We also have little problem seeing that our decisions are often mistakes
- But we have more trouble seeing that the same is also true of our adversary

Ownside bias?
- Are we the baddies?
 Nobody thinks they’re the baddies..
- Hitler believed he was defending Aryan civilisation
- Al Qaeda/ISIS believe they are defending the word of God
- The Soviets believed they were defending the working class and human
progress
- Even Putin believes he is defending Holy Russia against American aggression

The Elephant and the Rider (Haidt)

- We come to hold the views we hold for reasons of upbringing, culture etc
(the Elephant)
- These are deeply rooted beliefs – our reasoning faculties (the Rider) are
designed to protect, not examine, those beliefs
- Attempts to ‘reason with’ even highly intelligent people are likely to fail

55
Simulation theory and theory theory
- Simulation theory – putting yourself in another’s shoes and asking how you
would behave in that situation
- Theory theory – developing (and testing) a theory of what motivates others
and using it to predict their behaviour
- Theory theory recognises others can have quite different motivations to
oneself

Shore’s argument – Hitler, Stalin and FDR

Stalin and FDR judge Hitler

- Stalin was a ‘simulator’ asked what he would do in Hitler’s shoes
- Stalin was cautious and self-centred, so did not believe Hitler would take risk
of invading the USSR
- Consequently, ignored intelligence warnings of Barbarossa and almost lost
the war

Implications (1)
- Do not assume that the enemy is just like you
- The enemy might value certain things more than you would, and take risks
you would not take (or vice versa)
- Carefully build up a picture of the enemy’s motivations, paying especial
attention to their behaviour in costly and unusual situations

Implications (2)
- When you’re trying to win people over overseas, you can’t just rely on facts
- Facts will be dismissed where they collide with what people want to believe
- Building empathy first is key – once you can build that, then the factual
message will get through more clearly
- Shaming people by attacking their integrity or intelligence backfires

Conclusions
- We need to understand our adversary in order to help us formulate policy
- However we are bad at doing so because we often forget how unlike us the
enemy actually are
- We overemphasize centralisation and intent and overlook the fact that the
enemy sincerely believes himself to be in the right

56
Lecture 10 – Identity

Mossadeq
- “it may be easier to articulate the peculiar difficulty of constraining a
Mossadeq by the use of threats when one is fresh from a vain attempt at
using threats to keep a small child from hurting a dog or a small dog from
hurting a child” – Thomas Schelling

The Coup against Mossadeq

- “What the coup did was to take out the moderate, secular, element of
Iranian politics and enabled radical Islamists and radical leftists to emerge as
key opposition factions in place of it (in the 1960s and ‘70s). The coup had
this big impact of essentially eliminating this pro-democracy faction and that
had a very important impact on Iranian politics in the intervening years.”
- -Mark Gaziorowski

Strategic Culture
- ‘culture’ is a popular explanatory variable in strategic studies
 ‘culture eats strategy for breakfast’ – Peter Drucker
 ‘American Way of War’, ‘Western Way of War’, ‘Eastern Way of War’,
‘Chinese Strategic Culture’, ‘Iranian Strategic Culture
- But it’s a problematic concept
 Hard to define
 Often superfluous
 Often carries dangerous ethno-centric connotations (‘military
orientalism’) (just ask Mohammad Mossadeq)

What is ‘culture’?
- “collectively held ideas, beliefs and norms” – Carole Pateman
- “ideas intervene to define what (national interests, war aims or victory)
mean, and even what the end of a war looks like” – Patrick Porter

Tautology
- Russia intervenes in the Eastern Ukraine because Russian strategic culture
hold the Eastern Ukraine to be of particular value
- How do we know that Russian strategic culture prizes the Eastern Ukraine?
- Well, the Russians just intervened there, didn’t they?

‘Essentialism’
- Holds a state’s decisions in international politics to be a function of
unchanging national culture
- Russia, or Germany, or Iran, do this because there’s some ineffable Russian,
German or Iranian strategic culture that propels them to do it
- Therefore unable to account for change

57
Example
- “hundreds of years ago there seemed nothing surprising in German
barbarism, since the world was full of savages in these early days… Other
people grew up and settled down. The Germans never did. The Brazen Horde
remained savages at heart” – Sir Robert Vansittart, the Black Record

Now?
- BBC Poll: Germany most popular country in the world (2013)
- Germany at peace, democratic since 1945
- Massive donor of foreign aid
- Accepts more Syrian refugees than other EU states

Does invoking culture add anything

- The American way of war (Weigley)
 Dependence on technology, logistics
 Use of annihilation, attrition
 Minimization of casualties
 Preference for righting conventional enemies
- The British way in warfare (Liddell-Hart)
 Preference for using continental allies
 Preference for maritime, indirect approaches
- United states is a large, wealthy, technologically advanced country
- United Kingdom is a small, wealthy island off the coast of Europe
- Aren’t the American or British ways of warfare nothing more than a rational
adaptation to economic and geographic circumstances?
- To say it’s ‘culture’ doesn’t add anything. But it can cause confusion

Dangerous ethno-centric implications

- For propaganda purposes, states always have incentives to paint opponents
as irrational, driven by ancient, incomprehensible hatreds
- That doesn’t mean strategists should buy into this
- It leads to underestimation

The Maori Par of NZ

- Par is a traditional Maori fort
 When British were fighting Maori war against indigenous of NZ
believed that the Maori’s had a superstitious attachment to the Par,
meaning they would never abandoned the par
 Gave them strategic advantage to pin them to par and destroy
them there
 They tricked the British to thinking this so British would attack the par
and be hit by surprise attack by Maori’s

Female Suicide Bombers

- Great deal of reluctant of traditional Islamists to allow women to take part in
suicide bombing due to beliefs towards women

58
 - Once Palestinians decided to use female suicide bombers it gave them
advantage as it meant that women would be less likely to attract attention
from Israeli security forces
- Because Israeli’s believed that there is an inherent limitation in Palestinian
culture that prevented use of women, which lead Israelis to underestimate
opponents and many people to die because this was not true

The ‘Western Way of War’

- A cultural concept devised by conservative Victor Davis Hanson 
conservative American Military Historian
- Argues that there is a line you can trace going all the way back from ancient
Greece to the modern day which we calls the Western Way of War
- War involves conventional combat standing face to face engaging enemies in
an open battle wearing uniforms or some kind of marker what we call
conventional warfare essentially
 Looks down on deception and indirectness and prioritises face to face
fighting
Oh really…
- “he who overcomes the enemy by fraud is as much to be praised as he who
does so by force” – Niccolo Machiavelli
- “He should regulate his march so as to fall upon them (the enemy) while
taking their refreshments or sleeping, or at a time when they suspect no
dangers and are dispersed, unarmed and their horses unsaddled” – Flavius
Vegetius Renatus

What’s left of culture?

- It has to provide some explanatory leverage over and above a standard
rational choice account
- What might this look like?

Intercultural differences – trust and hierarchy

- There is evidence that cultures differ systematically in how much they value
certain things
- These differences can have important consequences for decision making
- Two of these differences concern trust and hierarchy

Hierarchy
- The extent to which is considered legitimate to question those who are
higher up in a given hierarchy than oneself
- Hofstede – power distance index
- East Asians = high power distance index; Australians and Americans = very
low

Trust
- Extent to which you think strangers will try to take advantage of you
- Individuals who do not trust others may even miss out on mutually beneficial
deals

59
 - Economists developed an experiment to measure ‘trust’

The Trust Game

- The experimenter gives you some money (e.g., $20). You can give some % of
this to a second person (a stranger)
- The experimenter will triple the amount you give. The second person then
gives you some % back. The experimenter triples this too
- Say you give 50% ($10). The second person gets $30. Say they give you back
50%. You get back $45
- How much would you give?

Cross-national differences
- Systematically, Greeks and Saudis, for instance, have been found to make
lower initial offers than Britons, Swiss or Australians
- That is, people coming from the former cultures are systematically less
trusting than the latter

Frames of Reference
- Policymakers use analogies from their own country’s strategic history and
thinkers first
- This may shape how they interpret other states’ moves and the best
responses to them
- E.g. Americans think by analogy to the Revolutionary War, Civil War, WW2,
Vietnam. Other ways of thinking about thinks less easy
- Same is true for other cultures. For instance, many Iraqis analogised the US
invasion of 2003 to the medieval Mongol invasion of the Baghdad-based
Caliphate

Goals
- ‘culture’ may provide the goals which states’ strategies pursie
- E.g. ‘sacred’ or quasi-sacred territory such as Jerusalem, Kosovo, and the
Ukraine, Northern Ireland
- But remember
 Many ‘ancient’ disputes are modern inventions
 States have incentives to exaggerate how much they care about these
issues
 Sacred values such as these can be traded off against other things
(e.g. post GFC, most Northern Irish Catholics opposed to unification
with Eire)

Even here…
- Strategic cannons so vast that one can use them to justify pretty much any
course of action
- ‘Western way of war’ supposedly about direct, face to face combat
- But you can find support fort evasive, guerrilla type strategies in Machiavelli,
Vegetius and in Western history

60
Moreover
- Thinkers and decision makers borrow from ‘other’ strategic cultures
- Sun-Tzu and Mao popular in US staff colleges
- Al Qaeda borrow liberally from thought of non-Muslim urban, guerrilla and
insurgency theorists

China
- Push for an IR theory with ‘Chinese characteristics’
- IR theories usually developed with reference to Western History and culture
- Are we missing out thereby?

The Sinic World System

Approx (206BCE)-Approx (1840CE)

- Chinese system not based on formal equality or sovereignty
- Other states in the system, were vassals of the Chinese emperor (e.g. Korea,
Vietnam)
- Could run their own internal affairs, but deeply influenced by Chinese culture

Peaceful
- Two ‘interstate’ wars in East Asia between 1368 and 1840 CE (Chinese
invasion of Vietnam in 1407 and Japanese invasion of Korea in 1592)
- In the same period, in the West, 46 wars between England and France alone

Kang’s argument
- Other states internalised Chinese value system (e.g. Korea, Vietnam, Japan to
some extent)
- Chinese did not believe conquest of these states to be necessary or
legitimate
- Historical memory of the peacefulness of Chinese hegemony permeates
Asian reactions to China’s rise today

Counters?
- Geography? No need for territorial aggrandizement/ less opportunity
- “if we envision IR as a scientific inquiry, then IR theory should be universally
applied. If we do not need a Chinese school of physics or chemistry, why
should we need a Chinese school of IR theory?” – Yan Xuetong

Conclusion
- Strategic culture is a popular concept
- But it’s a flawed one
- At worst, it is tautological, adds little to existing explanations and encourages
underestimation of the enemy
- At best, it can illuminate how one’s adversary conceives of issues, but even
here the relevance is limited

61
Lecture 11 – Structural Considerations

Back to the beginning…

- This course started off looking at foreign policy behaviour that traditional
realism can’t explain
- But realist-style structural variables cannot be ignored in FPA

Australia in the Ukraine

- Defence advisors noted that Australia simply cannot act unilaterally in places
like the Ukraine
- Relative power is a key constraint on foreign policy decision making

Why can’t Australia intervene in the Ukraine? – Discuss

- Ukraine is not going to be bothered if Australia sends troops as they know
Russia did it
- If we were to get into a shooting match with Russian backed separatists?
Australia can’t easily send reinforcements
- No matter how much an individual policymaker might want to do something
for whatever reasons, if the state doesn’t have very much power and not
many interests in a particular area then there is really nothing they can do

What is the basis of a country’s power?

- Correlates of War defines it in terms of
 Total population
 Urban population
 Iron and steel production
 Military personnel
 Military expenditure
- These are defined in relative terms (fraction of overall global resources)

Population
- Key component – more people implies bigger economy and potential military
power
- Population is bounded by geography – Australia can only get so big
- But population isn’t enough, or India would be a global power

Economic Development
- Produces resources which increase bargaining power and potential military
power
- US much smaller than China in terms of population but far bigger in terms of
GDP
- Hence US still more powerful than China
- European countries more influential than African nations much bigger

What causes economic development?

- Lots of debate in economies and political science
- No definitive answer

62
 - There are a number of schools of thought

Geography – Sachs, Diamond

- Today’s wealthy countries were also relatively wealthy in AD1000
- Temperate climate, proximity to trade routes with other highly populated
areas created an initial advantage
- But some areas wax and wane – Europe, Middle East, North and South
America

Political Institutions
- Democracy claimed by some to be crucial
- Democratic governments believed to be less corrupt
- Democratic governments also have more incentive to spend public money on
infrastructure and education
- But the economic evidence is weak

Economic institutions
- Rule of law and a strong court system incentivize enterprise
- Potential entrepreneurs need to feel secure that they will profit from their
business
- Politicians have incentives to loot however
- States which can restrain politicians from doing this will prosper

Military Power
- Wealthy populous states don’t always translate economic power into military
power (US 19th C, UK 19th C, EU, Japan today)
- But economic power is potent in itself and can be translated into military
power relatively quickly (US in WW2)

Current global distribution of power

Today
- US dominant politically, economically and militarily but faces internal
divisions
- China is rising but its unclear whether this will continue
- Russia is condemned to isolation and decline

63
 - EU is economically powerful but politically divided and militarily non-existent

USA
- US combination of economic strength and population size makes it so potent
- US likely to continue to be very powerful
- Internal divisions between ‘red’ and ‘blue; America however have already
harmed America’s standing and will continue to do so

China
- Much larger population than the US
- Projected to overtake the US in terms of GDP but has a long way to go in
terms of GDP per capita
- Continued growth depends on whether it can move up the value chain and
on whether there is a political transition
- Many economists sceptical

Russia
- Combination of tactics from disinformation to cyberwar to little green men
allowed Russia to amplify its power
- But in the long run the outlook is grim for Russia
- Bases of economic power shrinking, population stagnant. Isolated,
incompetent strategic elite. Increasingly dependent on China

European Union
- Larger population and GDP even than the US
- However EU is a union of nations not a unified power like the rest
- Breakup of the EU unlikely in spite of Brexit, but further integration also
unlikely
- EU will thus remain politically and militarily underpowered

Australia’s Place
- US still essential to Australian security
- Hard to see what could replace USA
- But US guarantees less credible now
- Australia is not big enough to defend itself against a great power or largely
alter the Asia Pacific balance of power
- Alternatives to US alliance need to be spelled out

64