U10_dq

Perspectives on risk management and behavioural finance Abstract The application of Behavioural Finance to Risk Management is still in its infancy and few models have evolved as to how to apply the theories and research findings to practical day-to-day risk management problems. In fact, the very topic of this Special Issue — Is there a role for Behavioural Finance in Risk Management? — is still a moot question. In addition to commissioning the papers in this issue, the editors asked a number of respected figures in the wider risk management community to provide their insights on the topic. In particular four thought-leaders were asked for their perspectives on two specific questions: (1) Should banks and regulators include the findings of psychological/behavioural research in their risk management frameworks; and (2) In the light of new UK legislation,1 should ‘reckless’ behaviour be regulated? The responses of these experts are enlightening, but as David Hillson notes the question is not whether behaviour should be considered in risk management ‘but how?’ Hopefully the perceptive answers will trigger debate among risk management professionals as to how the theories can be applied.

Keywords: behavioural finance, risk management, risk perception, reckless behaviour, financial models, judgment

Perspective 1 — David Hillson 3 Lower Heyshott, Petersfield, Hampshire GU31 4PZ, UK E-mail: [email protected]

Known globally as The Risk Doctor, David Hillson, FRSA, FIRM, FCMI, CMgr, HonFAPM, is a thought-leader and expert practitioner who consults and writes widely on risk management. His ground-breaking advances in risk methodology, risk attitude, risk appetite and risk culture have been recognised by a wide range of awards, including ‘Risk Personality of the Year’ in 2010–11. David also holds a number of honorary fellowships reflecting his risk expertise.

Question 1: Should banks broaden their risk management

scope and include the research results of ‘less precise’ sciences

(eg psychology) in addition to those of classical disciplines

(eg mathematics, statistics) into their risk management

frameworks? If so, why? If not, why not?

It is important to recognise that risk is not

managed by processes or systems, nor is it managed

by tools and techniques. Risk is managed by people

acting as individuals and in groups, responding to

uncertainty and making decisions that they deem to

be appropriate. Unfortunately despite the attractions

of a contrary view, human beings are not

dispassionate rational actors who make decisions

based on pure utility. Decades of research and

millennia of experience indicate that perception

drives both our assessments of the situations in which

we find ourselves and also our decisions on how to

react or respond and perception itself is subject to

multiple influences, both hidden and overt, including

a wide range of heuristics and cognitive biases. The

situation is only made more complex when

uncertainty is present.

Consequently any approach to understanding and

managing risk that ignores the human factor is

flawed and incomplete and it is likely to lead to

suboptimal outcomes. So at first sight the answer to

114 Journal of Risk Management in Financial Institutions Vol. 7, 2 114–121 # Henry Stewart Publications 1752-8887 (2014)

the question posed is an unambiguous ‘Yes’. The risk

management frameworks of banks (and other

organisations of all types) should take account of

human psychology as well as statistics.

This simple answer begs a more difficult question:

‘How?’ While it is relatively simple to state in

principle that risk psychology deserves a place in the

risk management framework, it is less clear how this

can be achieved in practice. It is also not entirely

clear that consensus exists within the risk profession

on which elements of risk psychology are important,

how they can be described and measured, whether

their effects can be influenced or should instead just

be recognised, and so on. Until we have broad

consensus on what we mean by ‘risk psychology’,

with an accepted body of knowledge and a

developing practice to implement the principles, it

may be safer to give a different answer to the basic

question. So perhaps the right answer is not a straight

‘Yes’, but a more considered ‘Yes but . . . .’ Two further issues arise from the detail of the

question as posed. The first is the implication that

psychology is somehow ‘less precise’ than other

sciences. This misunderstands the nature of science,

especially experimental science, and particularly as it

is applied to managing risk. The challenge of risk

management is to understand the degree of

imprecision in the various parameters that define

uncertain situations that matter. Under these

circumstances an approach that is overly-reliant on

high degrees of precision may be at a disadvantage.

There is nothing wrong with using the science of

imprecision to understand risky situations where

precision is reduced, variable or absent.

Secondly, the question specifically suggests that the

‘research results of [psychology]’ should be included

into risk management frameworks. This may be

unnecessarily limiting, as the results of pure research

are often not presented in a format that is amenable

to practical implementation. Instead it is important to

take such results and turn them into implementable

guidance, which is a non-trivial exercise requiring

specialist expertise and skills. Those who are tasked

with modifying their risk management frameworks

to take account of risk psychology should not need

to be experts in human behaviour. Instead it is the

responsibility of thought leaders and practice experts

to translate research results into pragmatic guidelines

that others can use.

Question 2: Should ‘reckless’ behaviour be regulated?

If so, why? If not, why not? How would one measure

‘recklessness’ in making credit or market risk decisions?

While the proposed measure in the UK has

obvious attractions, especially in the light of the

recent global financial crisis, it has significant flaws

that might well prove fatal. Chief among these is the

difficulty in defining recklessness (or as the UK

government act calls it, ‘reckless misconduct’). The

precise definition of recklessness in law has been

contested and has evolved. It generally involves a

person pursuing a course of action while consciously

disregarding the fact that the action gives rise to a

substantial and unjustifiable risk.

Herein lies the problem: What one person or

group or organisation regards as ‘substantial and

unjustifiable risk’ can equally be perceived by

another as ‘insubstantial and fully justifiable risk’,

resulting from different inherent risk appetites and/or

different chosen risk attitudes. It is unreasonable and

unrealistic to expect a regulator to produce

unambiguous guidelines on what level of risk

constitutes ‘substantial and unjustifiable’ in any given

situation.

The second potentially fatal flaw in the proposed

approach is alluded to in the question: How can

recklessness be measured? The UK government act

specifically states that ‘A regulator’s policy in

determining what the amount of a penalty should

be must include having regard to . . . the extent to which the contravention was deliberate or

reckless . . .’.2 This implies that recklessness is not binary but it is a continuum, such that the extent of

recklessness will determine the amount of penalty,

but if recklessness is a function of inherent risk

appetite and chosen risk attitude, both of which are

internal and intangible attributes of individuals and

groups, it is not amenable to reliable quantification.

In the absence of an agreed definition of

recklessness that takes account of differing

perceptions of risk and with an inherent inability to

measure the degree of any recklessness that might (or

might not) be present, it therefore seems impossible

for ‘reckless behaviour’ to be regulated in any

meaningful way.

Hillson, Sobehart, Ursachi and Riedel

# Henry Stewart Publications 1752-8887 (2014) Vol. 7, 2 114–121 Journal of Risk Management in Financial Institutions 115

Perspective 2 — J.R. Sobehart Credit and Operational Risk Analytics, Citi Franchise Risk Architecture, 2 Court Square, Long Island City, NY 11120, USA E-mail: [email protected]

is a Managing Director at Citi Franchise Risk Architecture where he works in credit risk capital, stress testing and portfolio loss models for wholesale and retail credit exposures. During his career, he has worked in a wide range of topics in risk management, finance, physics, computation and mathematical modelling. He also acted as a technical reviewer for many professional journals and book editors.

Question 1: Should banks broaden their risk management

scope and include the research results of ‘less precise’ sciences

(eg psychology) in addition to those of classical disciplines

(eg mathematics, statistics) into their risk management

frameworks? If so, why? If not, why not?

Risk taking, decision making and market reaction

are the result of a wide range of events from

economic and political changes to shifts in

confidence and preferences of market participants

and decision makers. Any meaningful framework for

risk management has to be able to accommodate

both the vagaries of the economy and the whims of

human psychology.

The behavioural aspects of risk taking and

decision making are not new topics. For decades,

economists have been debating the effect that

investors’ behaviour and imperfect information have

on asset prices and financial markets. Although there

is no shortage of behavioural models, there is still a

strong need to understand the behavioural aspects of

risk taking and decision making, which are

particularly relevant during market booms,

downturns and financial crises.

Despite continuous changes in financial products

and market complexity, market booms, downturns

and financial crises often exhibit common patterns

associated with the peaks and troughs of the business

cycles and, therefore, with the evolution of

economic expansions and contractions. Market

booms, downturns and financial crises are also

associated with periods of a strong divergence

between fundamental risk assessment and the market

perception of risk, which is dominated by cognitive

biases. This can lead to periods of excessive

speculation or lax lending standards — where

opportunities for profits are seized rapidly and

investors are willing to take increasingly bigger risks

at declining margins. Once the underestimated risks

materialise into actual losses and the economic

outlook changes, there is a rush to reverse the

expansion and reassess risk that can occur

precipitously, leaving institutions and investors

vulnerable to sudden changes in economic

conditions.

The dominant view in the economic literature of

risk taking, decision making and market behaviour is

that market participants are driven by rationality —

that is, they are presumed to make rational decisions

and quickly correct their behaviour motivated

primarily by self-interest. In the idealised rational

view of decision making and risk taking, price

changes are only the result of market participants

responding to a constant flow of new information on

fundamentals or exogenous shocks and, therefore,

their predictions of the future value of economically

relevant variables are not systematically wrong in that

all errors are random. In practice, investors and

arbitrageurs are limited in their ability to restore

price changes instantaneously owing to market

uncertainty and trading limitations.

Also contrary to standard views on market

rationality is the fact that sometimes asset price

movements are the result of behavioural patterns and

cognitive biases that may lead to under- and -

overreaction situations where investors seem to stray

from rational decision making. These situations can

be exacerbated when liquidity in one market or asset

class is constrained forcing institutions to reduce their

positions in other markets or assets as well.

Depending on the price elasticity of liquidity supply,

institutions may be forced to liquidate increasing

fractions of their assets at declining prices. This could

result in further market declines that can force other

firms to liquidate similar assets as prices fall or due to

Perspectives

116 Journal of Risk Management in Financial Institutions Vol. 7, 2 114–121 # Henry Stewart Publications 1752-8887 (2014)

demand for margin for their remaining assets. The

flood of sales would act as a positive feedback

mechanism that could accentuate the price drop

propagating the downward price trend across markets

or asset classes. Other feedback mechanisms based on

investor behaviour and cognitive biases can also be at

play during regular market upswings and

downswings. As asset bubbles and market crises

reoccur, and asset prices display persistent

behavioural patterns at different points in the business

cycle, it is becoming harder for supporters of the

idealised rational framework of risk taking and

decision making to water down these events.

The discussion above stresses the fact that the

elegant framework of optimal rational behaviour that

underlies neoclassical economics cannot reconcile itself

to actual human behaviour. The problem with this

framework is that it does not reflect how humans act

in the real world. Humans are not purely rational

creatures. Our brains are composed of a messy

network of functional areas, many of which are

dedicated primarily to dealing with emotions while

others are dedicated to processing, interpreting and

censoring information, which influence our judgment.

The expansion of the brain during human evolution

did not make us closer to the idealised rational agents

of economic theory but, to the contrary, allocated a

significant part of our brains to deal with emotions and

how we perceive the world around us. This in turn

helped us dissect the world and parse reality into

strings of causation even in situations with limited,

uncertain or contradictory information.

People have different ways of forming opinions

and making decisions given the contextual

information available to them and the way they make

inferences about the plausibility of different

outcomes. Judgment is always driven in one direction

or another by the particular way in which people

combine new evidence with contextual information,

emotions and preconceptions. In the mid-1970s

Kahneman and Tversky found that when a person is

confronted with an uncertain situation such as a

financial decision, the individual usually does not

evaluate the information or compute probabilities

carefully.3 Instead, the decision depends on a brief list

of emotions, instincts and inferences. This contrast

with rational objectivity, which demands that

judgment about the likelihood of occurrence of

events needs to be based on all the available

evidence, not just some arbitrary subset of it. Any

such a choice would amount either to ignoring

information that is available, or presuming

information that is not. In practice, people do

precisely this all the time. Our minds constantly filter

information that does not conform to our perception

of the world, sometimes leading to biased inference

and erroneous or unsupported conclusions even

when they may seem perfectly logical to us. That is,

we see less of the real world than we think we do,

and we act accordingly.

Evidence accumulated over years has shown

consistently that investors and decision makers are far

less rational in their decision making process than

economic theory assumes. In most of day-to-day

decisions people use primarily loose associations,

analogies, heuristics and ‘plausible reasoning’ — as

opposed to deductive reasoning, which may lead to

erroneous conclusions based on preconceptions and

biased contextual information. Furthermore,

physiological and psychological studies revealed that

certain aspects of emotions and feelings are

indispensable for rationality and decision making.

Human emotions are deeply empirical since they are

rooted in the predictions of brain cells that constantly

adjust and reinforce their connections to improve

their perception of the world around us. Loose

associations, emotions and feelings assist us with the

daunting task of predicting in an uncertain

environment and planning our actions accordingly

without having to exercise logical deduction or

induction in each action.

In some pathological situations, when people are

cut off from their feelings as a result of illness or

injury, even the most banal decisions can become

impossible — loosely speaking, a brain that cannot

feel cannot make up its mind correctly. Behavioural

studies on financial decisions also found that usually

the pain of losses is about twice as potent as the

pleasure of generating a gain. This effect is known as

loss aversion, which is part of a larger psychological

phenomenon known as negativity bias (loosely

speaking, for most people ‘bad’ is stronger than

‘good’ in their emotional responses). The fact that

people cannot make everyday decisions without

feelings and emotions contradicts the conventional

view of humans as rational agents. Of course, the

Hillson, Sobehart, Ursachi and Riedel

# Henry Stewart Publications 1752-8887 (2014) Vol. 7, 2 114–121 Journal of Risk Management in Financial Institutions 117

discussion above does not mean that emotions should

dominate the decision making process. Strong

emotional reactions and attachment to financial gains

or losses can actually be counterproductive.

However, empirical evidence shows that there is a

range of emotional responses and cognitive biases

affecting risk taking and decision making in general,

which need to be understood better by the risk

management community.

Another area of risk management dominated by

judgment and cognitive issues is actually the

quantification of risk. To overcome the uncertainty

in predicting future loss scenarios, risk management

relies heavily on assumptions about the random

nature of price changes and losses, and on

probabilistic models of loss likelihood and loss

severity based on statistical analysis of historical data.

Note however that the purpose of using probability

theory and statistics for estimating future losses is

simply to help us in forming reasonable judgment in

situations where we do not have complete

information. Also note that the sole purpose of the

statistical analysis of past data is to obtain a subjective

judgment that allows us to estimate the likelihood of

future events based on the belief that there is a causal

relationship between past and future events. To

illustrate this point, given any ‘real-world’ random

sequence of events (such as tossing a fair coin), the

likelihood of the next event (head or tail) is not

derived from perfect information about the

characteristics of the random draw (which are

unknown in advance) but inferred statistically from

events already observed. That is, the likelihood of the

next event is ‘assumed’ to be determined by the

limiting frequency of the events already observed. As

we get more and more observations, this assumption

may or may not become more solid. However, a

prediction based only on past observations is always a

‘subjective’ judgment inferred from the available data

and the belief that nothing else about the problem

will change in an unexpected way (for example,

unintentionally replacing the fair coin in our example

with a biased coin after a few tosses). Every time we

assume that some property of a random sequence

will hold also in the future, we are making an

educated — but completely subjective — judgment

of this type. This is exactly the problem faced by

financial institutions, which often have limited

information on borrowers, counterparties or changes

in the business environment and, therefore, need to

complement their risk models with sound judgment

and assumptions limited by cognitive issues.

# Jorge R. Sobehart

Perspective 3 — Irina Ursachi KPMG AG Wirtschaftsprüfungsgesellschaft, The Squaire, Am Flughafen, 60593 Frankfurt am Main, Germany Tel: þ49-172-3807349 (mobile); E-mail: [email protected]

Irina Ursachi holds an MSc in Mathematics with a specialisation in Financial Mathematics from the TU Kaiserslautern. At KPMG, she is a member of the competence cluster ‘Market Risk Management’ and deals mainly with the valuation of financial derivatives, financial engineering, as well as market and instrument model validation.

Question 1: Should banks broaden their risk management

scope and include the research results of ‘less precise’ sciences

(eg psychology) in addition to those of classical disciplines

(eg mathematics, statistics) into their risk management

frameworks? If so, why? If not, why not?

The functioning and development of financial

institutions and world economies depend on the way

human beings behave and on the decisions they

make. The misleading interpretation or failures

proceeded by economic decisions of any nature is

often the result of cognitive errors, such as under- or

overestimating the probability of an event, framing

beliefs to an anchor, or emotional biases. Often, the

outcome of these decisions yields either very positive

or very negative effects, provoking the frontiers of

risk management. Behavioural models aim to capture

these effects and should therefore be considered as a

steering tool.

Does this mean that risk management should

incorporate behavioural components within their

internal and external policies? Yes. However, it does

not mean that behavioural models should be core to

Perspectives

118 Journal of Risk Management in Financial Institutions Vol. 7, 2 114–121 # Henry Stewart Publications 1752-8887 (2014)

every business decision. Banks should consider

human behaviour as a variable in their equations,

while also setting the constraints of their business

model on a rational, mathematical and

most-importantly, controllable framework. Whether

or not behavioural effects occur, depends among

other aspects on the time horizon of business

decisions and risk analyses. There are behavioural

components that are deemed to be significant on

long time horizons whereas others are significant on

short time horizons. Long term behavioural

components can be found, for example, in liquidity

models for retail deposits. Short term effects are, for

example, the mass psychology effects that

dynamically appear with shocks within the economy.

The big challenge is to assess which assumption

good risk management should be based on, ie if a

behavioural or a rational assumption of counterparts/

clients is sensible for risk management purpose.

Otherwise put, a good risk manager should know

what is good and what is bad for business and this

should be based on understanding new positions,

industries, as well as one’s counterparties’ business

intentions. A good risk manager should also know

where the threats come from and how they can be

avoided, or, if they cannot be avoided, how they can

be controlled, because risk potential equals loss

potential.

A very important differentiation is to be made at

this point: the incorporation of behavioural models

should be treated as a different issue when it comes

to different institutions, ie we cannot generalise. It is

true that asset prices have been reported to also

follow behavioural patterns: we know that prices

fluctuate based on supply and demand, and of

course, supply and demand are influenced by actors

in the economy, but this does not mean that financial

institutions should drop the ‘random’ assumption on

asset prices and embrace empirical models based on

behavioural patterns. The consideration of such

patterns is likely to differ between different types of

financial institutions. Asset managers are likely to

take into account such patterns within the modelling

of asset prices, while banks would rather consider

behavioural patterns within credit — and liquidity

management.

The incorporation of behavioural finance models

within a bank’s organisation should have two

dimensions: a qualitative and a quantitative

dimension. Qualitatively, risk managers need to relate

the behavioural finance approach to their

governance/business model. This is often related to

operational risk and is best managed when people

within the organisation attain both a deep

understanding of risk and a good governance and

communication model within the organisation. Risk

managers should assess at which point their technical

knowledge is no longer the effective solution to

avoid emotional/cognitive biases and they should

set-up an organisational structure that is not prone to

market events that appear irrational. From the

quantitative perspective, the models that are

employed for instrument valuation and risk

quantification can be adapted to include behavioural

effects. For example, portfolio models can be adapted

to also include mass psychology effects, such as

taking higher risks when the already incurred losses

are high. While banks have already recognised such

behaviour and have taken a step towards minimising

risks via stop-loss limits, behavioural effects could be

treated in risk modelling via distributions that are

transformations of probability measures that

overweigh small probabilities and underweigh large

probabilities. All in all, risk managers should aim for

an ongoing assessment of weaknesses of any kind that

can impact their business, while keeping their

organisation in line with the regulatory requirements.

A prerequisite to management and regulation of

risks induced by irrational human behaviour is that

appropriate measures against reckless behaviour are in

place. Reckless behaviour is, eg careless business

activity, rogue trading, manipulation of economic

data in one’s advantage and aiming at arbitrage

opportunities or moral hazard, such as cashing out

large bonuses on short time-horizons.

The rules set out by the regulatory authorities

should indeed be the ones of a fair economic

game, but is this really enough? Nowadays, the

financial system has become so complex, that it is

very difficult for authorities to track every little

step and important decision of a bank. If history

taught us anything, it is that there will always be

clever individuals who are going to find a way to

break the rules without being noticed for a while,

or to find arbitrage opportunities and profit from

them. In order to assure that the behaviour of

Hillson, Sobehart, Ursachi and Riedel

# Henry Stewart Publications 1752-8887 (2014) Vol. 7, 2 114–121 Journal of Risk Management in Financial Institutions 119

individuals is the one intended and placed within

the framework of the rules, behavioural insights

are one important hint on how one can design

control procedures, which can also be tested

empirically.

Question 2: Should ‘reckless’ behaviour be regulated?

If so, why? If not, why not? How would one measure

‘recklessness’ in making credit or market risk decisions?

Reckless market or credit decisions are difficult to

model, however, they can be handled by

implementing corresponding control processes, such

as smart limit systems. The regulation of reckless

behaviour is currently pursued by imposing more

responsibility on risk managers and by diminishing

risk that is not sustainable. Moreover, risks induced

by human behaviour should be linked to the impact

that they have on the payoffs that they induce. While

this is not common practice within risk models

suggested by regulatory guidance and while this is

not yet applied within current state of the art models

of the financial industry, recent advances4 in research

show that it is possible to link behavioural

components to risk measures.

To sum up, a successful risk management practice

should assess at which point and for which decisions,

a behavioural component should be included within

the steering of risks. It should conduct a careful

management of risks, to always aim for the

identification of threats that could impact their

organisation even if these threats seem unlikely and

to dynamically assess model weaknesses and

potentially unexpected market events. Most

importantly, a healthy and well-functioning risk

management practice should aim for the

understanding of risk and its sources — human or

not.

Perspective 4 — Frank Riedel Center for Mathematical Economics, Bielefeld University, Postfach 100131, 33501 Bielefeld, Germany E-mail: [email protected]

Frank Riedel is a professor and Director of the Center for Mathematical Economics at Bielefeld University.

Question 1: Should banks broaden their risk management

scope and include the research results of ‘less precise’ sciences

(eg psychology) in addition to those of classical disciplines

(eg mathematics, statistics) into their risk management

frameworks? If so, why? If not, why not?

I take the freedom to modify and enlarge slightly

the scope of the [first] question by taking banks’ risk

management and a proper regulation, two intrinsically

related topics, into account. Let us thus ask: Do

regulatory authorities and banks’ risk management

departments need to take findings from psychology or

related behavioural sciences into account?

Economic theory is based on decision theory, a

rich building of mathematically oriented, axiomatic

theories that describe ways to come to a ‘rational’

decision in complex environments. Such decisions

are typically based on (rather complex) utility

functions, a way of ranking various alternatives and

assessing which option is better or worse in a given

situation.

Facing such a theory, the common complaint of a

practical man or woman is: ‘But I do not have/know

my utility function’ — and indeed, human beings

are not rational agents nor do they have a utility

function encoded in their brain. Common sense, or,

to enlightened persons, a short reading of Freud’s

‘Psychopathology of Everyday Life’ would be more

than enough to support this conclusion. Behavioural

economics and psychology have reinforced the

common sense and psychoanalytic knowledge with

many a study. If we want to understand humans — a

difficult task — we certainly need psychology.

Does this imply that banks, or regulatory

authorities, should take behavioural findings into

account? No, it does not.

The rules set by regulatory authorities and used by

risk management departments, should ideally be the

rules of a fair economic game. In general, these rules

allow, even promote free economic activity in the

interest of the involved trading partners, but they

Perspectives

120 Journal of Risk Management in Financial Institutions Vol. 7, 2 114–121 # Henry Stewart Publications 1752-8887 (2014)

need to prevent misuse of freedom and misuse of

market power. In the context of financial markets,

good rules support reasonable risk taking, credit

provision and insurance, promoting a society’s

economic well-being and wealth, as well as banks’

profits; good rules also hinder economically

undesirable activities like excessive risk-taking, or

manipulations of risk management that make the

bail-out of large banks necessary, or put a bank’s

survival at risk.

A crucial aspect of good risk management is the

following. The rules should not set incentives to

‘trick the system’. Many tax rules are a case in point,

but, as I have explained in detail elsewhere,5 the basic

risk measure, value at risk, also sets the wrong

incentives and gives the wrong numbers, and it can

be, and was, exploited. value at risk induces smart

agents to speculate on small probability events and is

easily manipulated.

The LIBOR scandal is another example. The

LIBOR, to recall, is determined by asking a

number of banks about their ‘cost of money’ for

certain maturities. The problem is not so much the

traders’ psychology here. The LIBOR system was a

silly game in as far as it is optimal for rational

traders to lie in such an environment and to

manipulate the LIBOR in order to make gains from

their banks’ exposure in derivatives — and this

happened.

The design of good rules is a job for rational

analysis. We want to assess how the game we come

up with is played by smart, rational and, yes, ‘greedy’

actors, in order to assess if we have set the right

incentives. In order to do so, it is necessary to

understand how rational agents behave in the game

defined by risk management rules. Good risk

management and regulation is thus based on rational

analysis, or in other words, mathematical economics.

The recent literature on ‘risk measures’ is a good

starting point here. It has shown how one can replace

the catastrophe-triggering value at risk by better risk

measures that assess the riskiness of financial positions

in a robust way. Such risk measures promote

diversification, and thus prevent speculation on small

probability events, to give one example. Their

robustness makes them immune to human

behavioural biases as well, in a certain sense. It would

thus be worthwhile to finally replace value at risk by

a robust risk measure (as it is currently being done in

some places).

What about the role of psychology then? First,

and maybe obviously to the readership of this

journal, judgment (‘Urteilskraft’) is crucial for risk

managers. This typical human skill is both necessary

and important, and it cannot be totally replaced by

rules, neither by ‘mathematical’ nor by ‘behavioural’

algorithms. Second, for the implementation of new

rules, testing in the laboratory and in the field

remains important. We need mathematical economics

to come up with the right rules, ‘Urteilskraft’ for

managers, and empirical tests to check the

consequences and the robustness of our theories.

References 1 HM Treasury (2013) ‘The Government’s response

to the Parliamentary Commission on Banking

Standards’, HM Treasury, London, July 2013,

available at https://www.gov.uk/government/

news/government-responds-to-parliamentary-

commission-on-banking-standards

2 HM Treasury (2011) ‘A new approach to financial

regulation: The blue print for reform’, HM

Treasury, London, June 2011, available at https://

www.gov.uk/government/uploads/system/

uploads/attachment_data/file/81403/

consult_finreg_newapproach_blueprint.pdf, last

accessed on 20th March, 2014.

3 Tversky, A. and Kahneman, D. (1974) ‘Judgement

under uncertainty: Heuristics and biases’, Science,

Vol. 185, No. 4157, pp. 1124–1131.

4 See eg Greg Davies (2012) ‘Behavioral investment

management: An efficient alternative to modern

portfolio theory’, McGraw-Hill, New York.

5 Riedel, F. (2013) ‘Die Schuld der Ökonomen: Was

Mathematik und Ökonomie zur Krise beitrugen’,

Econ Verlag, Berlin.

Hillson, Sobehart, Ursachi and Riedel

# Henry Stewart Publications 1752-8887 (2014) Vol. 7, 2 114–121 Journal of Risk Management in Financial Institutions 121

Copyright of Journal of Risk Management in Financial Institutions is the property of Henry Stewart Publications LLP and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.