Can reformed and improved risk management techniques make the financial system better adaptable and more resilient to financial crises?

(1)

Amsterdam, June 2017

Master Thesis Research Project: MSc Political Economy

‘Can Reformed And Improved Risk Management Techniques Make The

Financial System Better Adaptable And More Resilient To Financial

Crises?’

Name: Milan Bednar Student Number: 11242043

Date: 31. 8. 2017 Word Count: 19,541 Email: bednar6@gmail.com

Thesis Supervisors:

Prof. Geoffrey R.D. Underhill (University of Amsterdam) Thomas van Galen (Cardano)

Second Reader: Dr. Eelke Heemskerk

(2)

2

Abstract

The aim of this thesis is to show the presence and provide a solution for what appears to be a false perception of the frequency of financial crises in the financial sector. Specifically, this thesis claims that the answer lies in the field of risk management and that by using the insights from BE and scenario planning we can build a new hybrid Three Swan Model that provides a superior alternative to the conventional approaches to risk management. Ultimately, this alternative can help us improve the internal risk management models of market actors, which, in theory, could enhance the adaptability of the financial system and in turn make it more resilient to financial crises.

Keywords: Risk management, financial crises, scenario planning, scenarios, behavioural economics, financial stability, black swan, financial sector

Acknowledgement

This thesis was written during an internship with the Cardano Insights team at Cardano Risk Management B.V. in Rotterdam. Hereby, I would like to thank and send my regards to all those who helped me with finishing this project. Namely, I extend my sincere gratitude to Prof. Geoffrey R.D. Underhill who provided me with the necessary academic guidance throughout the course of my writing. In parallel, I would like to thank Thomas van Galen for agreeing to be my internship supervisor and for providing me with indispensable support. A special thanks also goes to Stefan Lundbergh, Andreea Popescu and Nicole Xu from the Cardano Insights team who helped me structure and polish my unstructured thoughts. Finally, I would like to thank my parents Jan and Jana, who allowed me to pursue my studies and whom I love dearly. At the same time, I want to thank both of my brothers Lubomir and Viktor who never make my life dull and who continue to support me at all times.

Milan Bednář

(3)

3

Many authors (Freix et all. 2015:152-3; Reinhart and Rogoff 2009:142; Kindleberger and Aliber 2005:3; Sandleris and Wrights 2011:1-2; Weise 2012; White 2010:3; Jorda et al. 2011:372) dedicate their research to investigating the real costs of financial crises and their research strongly suggests that financial crises have much higher costs than the amount of money that is spent on bailouts, which is often assumed to be the main cost attributed to financial crises. Furthermore, in 2007 Nassim Taleb coined the concept of a ‘Black Swan’ to explain events that are extremely rare, unpredictable and have a major impact. Worryingly, since then many have mistakenly treated financial crises as black swan events and as such they are deemed rare and unpredictable. This led to an exclusion of financial crises from strategic discourse, even though it is recognised that financial crises represent a severe threat to the stability of the financial system as a whole (Authers Jun. 2010).

Authors, such as Reinhart and Rogoff (2009), Kindleberger and Aliber (2005) or Solimano (2016), take a different perspective on the topic of financial crises. Rather than treating them as black swans they argue that financial crises occur regularly and can be investigated from a historical perspective. This brings up an important question; has Taleb’s concept of a black swan been, whether intentionally or not, misunderstood by the financial sector all this time? Whilst conventionally financial crises are perceived as rare, unpredictable, and with high impact but in reality they are not by definition black swans then the whole financial system suffers from a case of false perception about financial crises. The possibility that such a misunderstanding is present in the financial sector requires a careful investigation of the issue and an immediate exchange for a more accurate understanding of probabilities as well as an immediate integration of the improved understanding of probabilities into the internal operations of both regulators and financial institutions.

Moreover, the post-2008 financial landscape is defined by a sharp increase in the regulatory burden aimed at avoiding a repeat of the 2008 financial crisis and it’s continuing

(5)

5

aftermath (Litan 2012:276; Gadinis 2013:334-5; Ötker-Robe and Maria Podpiera 2013:24-25,29). Risk management (RM) is a new and popular field used, in theory, to improve the firms’ resilience to financial shocks, nevertheless the field of RM, naturally due to its novelty, must share some of the blame for the problem. The theoretical promise and novelty, but at the same time internal weakness of contemporary practice, make RM arguably the most suitable and also controversial set of policy tools that could aid in making the financial system more robust. Therefore, this thesis will thoroughly explore the field of RM and its contemporary problems. The possibility that there is a general misunderstanding of financial crises in the financial system combined with manifest weaknesses within the field of RM that, in turn, serves as a defence against crises lead to the following research question: can reformed and improved risk management techniques make the financial system better adaptable and more resilient to financial crises?

The structure of this thesis will consist of three main sections. First, I will discuss the literature on the frequency of financial crises and some other relevant points that emerge alongside the topic of financial crises. For instance, if financial crises are truly more frequent and they threaten the stability of the global financial system then we have to define what is meant by financial stability before we set out on the quest to improve it. In addition, the literature suggests that financial crises are multivariate and occur in clusters. Furthermore, their frequency and impact varies across the world which might be due to differences between emerging markets and advanced economies. Crises are more frequent in emerging markets but they are not limited to these economies and understandably if a crisis occurs in the core financial centres it is more likely for the crisis to become global than if a crisis occurs in an emerging market economy. Finally, I will point out why a more frequent occurrence of financial crises is becoming a much more serious matter due to the risk of contagion.

Second, I will explore the literature on RM and I will discuss three problems that hinder the field of RM. Currently, the field of RM suffers from 1) the misconception of the nature of risk, 2) the lack of a common language of risk, and 3) a deeply ingrained ideological controversy about

(6)

6

how to analyse and deal with risk in practical terms. In regard to the misconception of risk, behavioural economics (BE) tells us much about how and why errors in RM emerge. Primarily, this is to do with heuristic biases originally studied by Kahneman and Tversky (1971; 1979), for example, representativeness bias, overconfidence, anchoring, etc., which consequently lead to substantial errors within many of both the qualitative and quantitative RM models. On the subject of the lack of common risk language, I will illustrate how different agents employ different definitions of risk. This lack of a shared language leads to diverging opinions of how risk should be managed. Unless there is agreement on the core risks and how they work in practice, the RM debate takes place on the basis of contrasting terms of reference. Finally, the controversy within the field of RM stems from the clash of various RM frameworks. In turn, based on these varied understandings of the problem at hand, each of these frameworks claims to be a recipe for success, yet they contradict each other on the conceptual and therefore practical level. In particular, some frameworks advocate softer qualitative RM methods (heat maps or scenario planning) and some are more ideologically aligned with the purely mathematical models (ALM). As a result, the field itself is very divided and both sides of the argument are persuaded that their version ought to be without the other. This thesis thus partially focuses on a critical analysis of contemporary RM techniques and poses the question as to how to improve them and thus reduce the risk of financial crises.

Third, globalisation, disruptive innovation and uncertainty are factors driving an incredible pace across the industries and economic sectors, yet some market actors are clearly better at withstanding and adapting to these turbulent times than others (Kotler and Caslione 2009:41). Therefore, a dissection of the robust strategies of those market actors that perform comparatively better to the industry standards will provide useful insights on how to duplicate these adaptability-enhancing strategies to other industries and other market actors in the pursuit of greater financial stability. A large section of this thesis will thus be dedicated to scenario planning (scenarios) as a new and improved form of RM technique. Scenarios are a strategic decision-making technique that is utilised in many industries outside of finance and evidence suggests that companies using scenarios tend to perform better compared to the industry standards due to their ability to spot

(7)

7

trends and threats in advance and quickly react to them. I chose Royal Dutch Shell and Cisco as case studies because Shell is the company that popularised the use of scenarios and Cisco created scenarios in 2010 about the future of the Internet in 2025 and now we can see them evolving in parallel to our reality.

Argument

First, this thesis supports the argument that financial crises are both frequent, predictable within bound, and costly. I will propose a new Three Swan Model (3SM) that provides an analytical framework or ‘heuristic’ supporting the main arguments of this thesis. The 3SM, positing black, grey and white swan events, provides a more accurate real world characterisation of financial risks than Taleb’s dichotomous white-black swan model does. I argue that if market actors adopt the 3SM they will be better at recognising risks of various real world phenomena. Furthermore, The 3SM tackles the problem of the lack of common risk language in the field of RM because it proposes different ways of how different world situation, ergo different swans, should be modelled. I will be arguing that if all the parties adopt the 3SM and clarify, which of the three swans they are currently confronting, then we can arrive at common grounds and hopefully a feasible solution to solving the market phenomena at hand. In order to solve the misconception of risk in the field of RM I will draw on the literature from BE where vast number of papers highlight how the behavioural biases that make perception of risk difficult can be tackled and the perception of risk by market actors can be improved. Furthermore, in order to solve the ideological controversy in the field of RM I will argue the hypothetical middle road and claim that in order to create a truly robust RM model we need to utilise both the quantitative and qualitative techniques as each are suitable for different situations. Evidently, the field of RM faces some significant problems, which have to be resolved before we enjoy a more robust financial system and the improved market actor strategies that RM promises. Finally, this thesis will attempt to link a more robust RM system on the micro (firm) level and the implications this would have for the macro or systemic level – sometimes characterised as the ‘Minsky cycle’ (Minsky 1992:8; Toporowski 2008:726). Certainly, there are several question that would arise if the firms within the financial

(8)

8

system were to adopt supposedly improved RM methods. For instance, can the Minsky cycle self-adjust and short-circuit itself through firm behaviour and therefore avoid the Speculative-to-Ponzi stage and the Minsky moment of the Minsky cycle? Or perhaps, financial crises are inevitable, in which case could the more robust RM system significantly decrease the damage caused when the bubbles burst?

In short, this thesis directly confronts the central problem of micro-level risk management by firms. Generally, I will argue that by using the insights from BE and scenario planning we can build a hybrid 3SM that provides a superior alternative to the conventional approaches to RM and ultimately this alternative can help us improve the field of RM, which will consequently result in more robust market actor strategies and possibly greater financial stability1_{. Hence, this thesis}

advocates that in order to achieve a more stable financial system we must take the bottom up approach and focus on the interaction of firms and regulators. A proper analysis of RM, its shortcomings and the potential benefits that RM offers to the financial system allows one to more adequately understand why financial stability is a more crucial factor than ever, especially due to globalisation, higher chance of crises contagion, and increased interconnectedness between countries and industries through the financial system (Werner and Zamarripa 2011:322). Furthermore, a careful scrutiny of this topic sheds light on market actor behaviour and the difficulty of implementing and aggregating new approaches within any system, even if the new approach should be rationally attractive to all market actors due to its positive impact on the user’s longevity.

[1] Literature on the Frequency of Financial Crises

The theoretical debate in this paper will tie together discussions from several strands of literature. To begin with, I will look into the meaning of ‘financial stability’ since our goal is to

1_{‘Possibly’ is a key word since, as Persaud (2000) or Wagner (2014) argue, sound risk management at the firm level}

may not yield systemic stability and may indeed be precisely what generates systemic instability. Systemic stability is something that, I argue, comes from the interaction of both market agents and regulators.

(9)

9

achieve a more stable financial system, therefore we should have a clear vision of what we aim to achieve. However, I will show that defining ‘financial stability’ is not as easy as it might seem at first sight (Goodhart 2004:2-3; Aspachs et al. 2006).

While investigating the literature on the frequency of the occurrence of financial crises, I will discuss several points that the scholarly work highlights. First, I will introduce the literature that supports the argument that financial crises are historically not as rare as many believe. Second, I will look into the types of financial crises that are relevant today – sovereign debt and banking crises – and I will scrutinise whether countries are capable of solving either one of the two types of crises. This will be done through exploring the concept of ‘country graduation’ that Reinhart and Rogoff (2009) introduce. Third, I will discuss how crises affect advanced economies and the emerging markets. In addition, I will also elaborate on the risk of contagion. Overall, this section will lay down the premises for the 3SM by demonstrating that an integration of an improved understanding of financial crises is necessary due to the false perception about the frequency of their occurrence, the devastating effects and the increased risks that financial crises represent.

1.1.1 The Ontology of Financial Stability

Before we delve into the discussions about the probabilities and frequencies of financial crises and the internal debates within the field of RM we first have to clarify what is meant by the term “financial stability”. Many authors, such as Goodhart (2004:2-3) or de Haan and Oosterloo (2006:256), argue that, as a matter of fact, defining financial stability is not an easy task. De Haan and Ooosterloo (2005:257) also point out that in most countries the law does not provide a clear objective for financial stability supervisors and that there are hardly any accountability measures regarding the objective of financial stability. Not having a clear definition of a long-term financial stability or taking it as a one-time static measurement proves not very helpful to the system. Even though central banks can say whether there is more stability (less accumulation of systemic risk) in the system today than there was yesterday, this approach still does no tackle the issue of how

(10)

10

to aim for an overall decrease in systemic risk accumulation in the long-run. Fell and Schinasi (2005:115) very aptly write that financial stability assessments must not only take stock of disturbances as they emerge, but also identify and examine the vulnerabilities that could lead to such disturbances occurring in the future. I add that we must also attempt to identify ‘new’, not just the current disturbances. Differently put, we should not only perceive financial stability as a measurement of how much more or less systemic risk accumulated today in comparison to yesterday but we should strive to prevent the accumulation of systemic risk in the long-run by constantly improving the adaptability of the financial system.

ECB (2017) defines financial stability as a state whereby the build-up of systemic risk is prevented. In this case, systemic risk can be best described as the risk that the provision of necessary financial products and services by the financial system will be impaired to a point where economic growth and welfare may be materially affected. As such, systemic risk can be derived from 1) an endogenous build-up of financial imbalances, possibly associated with a booming cycle, 2) large aggregate shocks hitting the economy or the financial system, and 3) contagion effects across markets, intermediaries and infrastructures. Moreover, others look at the stability of the financial system in terms of a trade-off between competition and financial stability (Allen and Gale 2004:453). However, looking at ECB’s definition, I argue, in accordance with Goodhart, de Haan and Oosterloo, that even ECB finds it difficult to capture the definition of financial stability. In other words, if we were to accept ECB’s definition of financial stability as a state where the build-up of systemic risk is prevented, we still lack the answer to figuring out what is a more stable financial system. The provided definition is very binary and implicitly states that there is either financial stability or financial instability. If the status quo is the latter then financial crises are more likely, however the definition struggles to explain how much more stable are we compared to yesterday. Generally, if a crash did not happen yesterday and today, does that necessarily mean that today we are more, or less, stable than yesterday? This thesis proposes that we have to look at the stability in the financial system in dynamically adaptive terms rather than by assuming that due to the lack of a crash we can conclude that the system is more stable.

(11)

11

Therefore, if we want to arrive at a less binary and a more dynamic definition of financial stability we have to take a short detour into metaphysics and ask ourselves the question; what would be the ‘good’ or the ‘ideal’ financial system in Aristotelean terms? This is important because then we can draw the parallels between the current financial system and the ‘ideal’ system. Highlighting the differences will subsequently allow us to see which steps we can take to move closer to the ‘ideal’ option. First of all, it would be a system in which stability would be considered from a long-term evolutionary point of view. Second, the system would possess certain characteristics, namely; 1) there would be room for improvement by the method of trial and error. Innovations and new strategies would be tested and allowed to fail but ultimately there would be an adoption of what works, change in status quo, and repeat of trial and error in pursuit of more robustness. A process that Schumpeter (1994:83) would label ‘creative destruction’. 2) The system would correct itself without outside intervention, alternatively said it would be self-sustaining. On the contrary, the current financial system relies, and arguably blindly depends, on external correction in form of bailouts from tax-payers’ money. 3) The utility function should be satisfied by allowing trade to work in an efficient way. Simply put, there would be an optimal allocation of resources between buyers and sellers. 4) Regulation is present within the system because financial stability refers to the interaction of both the market regulators and the market actors.

Now, it is important to answer why this system would be better than the system currently at place. As mentioned in the paragraph above the system we currently have is essentially unable to stand on its own and systemic crashes can be considered inherent to it (Minsky 1992:8; Mellor 2010:138; Hogan and Sharpe 1997:17). Furthermore, the core of the system has a monopoly on bailouts because the large financial institutions simply have roots running too deep into the socioeconomic wellbeing (Wright 2010:18; Jeanne et al. 2001:427). This results in absolutely zero incentive for the financial institutions at the core to decrease risk taking on large scale because they know they are likely to get bailed-out. Moreover, a decrease in risk taking, which would increase the systemic stability, at the same time represents smaller returns on investment and less

(12)

12

money for the shareholders. Thus, I argue, the current system is far from resembling a self-sustaining financial system.

Unfortunately, we are very unlikely to achieve an ideal Aristotelean system, nevertheless we can still learn much from imagining the ideal financial system. It also exposes how limited our current definitions of financial stability are, that one has to adopt an Aristotelian definition as a benchmark. In particular, we can at least attempt to get as close to the Aristotelean system as possible. The contemporary system is far from doing its best in taking steps towards that direction, especially when it comes to the system’s adaptability. Fingers can be pointed in many directions, whether at the incentives schemes we have built (Leiss 2010:99; Cihak and Demirguc-Kunt 2012:29), the unwillingness to accept new heterodox models and theories challenging the neoclassical status quo (Crotty 2009:127; 2009:564; Wahl 2009:86) or the strong grip of elites over the system (Iversen 2006:611; North et al. 2009:143). The truth is, that we are looking at a nexus of factors, which all shackle the financial system with numerous very strong chains in a very prohibitive way. Indeed, as Underhill (2014:1) or Rogoff (2010) point out, the private market agents are very influential in the lobbying process and thereby in steering the wheel, which dictates the future course that the financial sector will take.

Generally, in this paper an improvement in financial stability refers to a larger picture where we attempt to close the gap between our current financial system and the ideal Aristotelean system. The system has to be seen as dynamically changing over a long period of time rather than from a static short-term point of view (Fell and Schinasi 2005:115), which might seem effective at the time but essentially is like collecting nickels in front of a steamroller (Taleb 2010:217). To ensure the longevity of the financial system we have to turn our focus to its willingness and ability to adapt. A shift, which helps the system to increasingly resemble its Aristotelean counterpart would therefore be considered as stability enhancing. Equally so, tools that are capable of closing the gap between our and the ideal system are deemed as strengthening the stability of the financial system.

(13)

13

1.1.2 History Doesn’t Repeat Itself, But It Does Rhyme

The title of this subsection is credited to Mark Twain’s quote and it very aptly describes the reoccurring trend of the emergence of financial crises throughout the history. Financial crises are, by many, seen as black swans and according to Taleb’s (2010) definition 1) are an outlier that lies outside the realm of regular expectations because nothing in the past can convincingly point to its possibility, 2) carry an extreme impact, and 3) can only be explained and rationalised in retrospect. The anecdote comes from the time before Australia was discovered and when people in the Old World were convinced that all swans are white. The belief held true due to it’s at the time irrevocable empirical evidence, which suggested that only white swans exist in this world, hence the concept of a black swan represented the impossible. The discovery of a black swan then highlighted the fragility of our knowledge because a single observation invalidated all the previous knowledge about swans. Nowadays black swans refer to an unpredicted event that was originally perceived impossible but eventually is proven to be possible. For instance, many of the financial institutions claimed the financial crisis of 2008 to be a black swan and would assert its unpredictability, nevertheless, after the devastating effects of the crisis a whole new strand of literature sprang up explaining the causes behind the 2008 meltdown. Similarly, the crisis had a very high impact and arguably parts of the world are still experiencing the aftermath even nine years later (Ötker-Robe and Maria Podpiera 2013:24-25,29). In these terms, the 2008 financial crisis could possibly be classified as a black swan and thus characterised as rare, unpredictable, and with high impact. This thesis argues against such a classification and advocates the view that labelling financial crises as black swans is incorrect, short-sighted and based on false presumptions.

As stated above, the conventional and more popular view tends to classify financial crises as black swans but if we direct our attention to the work of some authors who explore a larger historical picture we can find a rich and in-depth research that directly opposes the idea of classifying financial crises as black swans. The scope of the historical narrative is a key component of such a research as the frequency of the occurrence of financial crises can only be uncovered by

(14)

14

an investigation that looks at centuries of economic history, not just five, ten or even fifty years (Schularick and Taylor 2012:1029). Solimano (2016:283) writes in a complementary manner recognising that financial crises need to be understood in a historical perspective and in systemic terms rather than as isolated episodes. Reinhart and Rogoff (2009:15) illustrate that humankind suffers from a particular this-time-is-different syndrome, which makes us believe that whenever a financial crisis happens it is a different situation than during the last crisis. However, in accordance with the authors’ research and the satirical title of their book, the reality suggests that this time, in fact, is not different and that a repetitive pattern haunts humanity in a form of sovereign debt crises and banking crises for over eight hundred years. Freixas et al. (2015:73) similarly argue that financial crises are recurrent systemic phenomena. Kindleberger and Aliber (2005:239) further support this view and claim that the monetary history of the last four hundred years has been replete with financial crises. Equally, Bordo et al. (2010:27) or Ahdieh (2010:286), argue that besides the breadth and depth of financial crises the frequency is also generally on the rise. Reinhart and Rogoff (2009:1) describe the syndrome as rooted in the firmly held belief that financial crises are things that happen to other people in other countries at other times; crises do not happen to us, here and now. Assuming that we are doing things better, smarter, and that we have learned from our past mistakes is a mistake which makes us blind to the unfortunate reality that a highly leveraged economy can unwittingly be sitting with its back at the edge of a financial cliff (examples in Fig. 1 below). As shown in this paragraph, the subject literature very strongly supports the argument in this thesis that financial crises are not historically as rare as is often believed.

Source: Reinhart and Rogoff (2009:15-20)

Case Why was this time different? (thinking at the time)

The buildup to the emerging market defaults of the 1930s

There will never again be another world war; greater political stability and strong global growth will be sustained indefinitely; and debt burdens in developing countries are low

The debt crisis of the 1980s Commodity prices are strong, interest rated are low, oil money is being recycled, there are skilled technocrats in the government, money is being used for high return

(15)

15

infrastructure investments, and bank loans are being made instead of bond loans, as in the interwar period of the 1925 and 1930s. With individual banks taking up large blocks of loans, there will be incentive for information gathering and monitoring to ensure the monies are well spent and the loans repaid

The debts crisis of the 1990s in Asia The region has a conservative fiscal policy, stable exchange rates, high rates of growth and saving, and no remembered history of financial crises

The debt crisis of the 1990s and early 2000s in Latin America

The debts are bond debts, not bank debts. With orders of magnitude more debt holders in the case of bonds than in the case of international banks, countries will be much more hesitant to try to default because negotiation would be so difficult

The US in the run up to the financial crisis of the late 2000s (the 2nd_Great

Contraction)

Everything is fine because of globalisation, the technology boom, our superior financial system, our better understanding of monetary policy, and the phenomenon of securitised debt

Fig. 1 – Table illustrating five instances of the this-time-is-different syndrome (20th_{& 21}st_century)

To demonstrate some numbers, in the long historical panorama of sovereign defaults, Reinhart and Rogoff (2009:34) account for at least 250 episodes of sovereign external defaults and at least 68 episodes of domestic public debt defaults. Tomz and Wright (2007:353) conclude similarly high numbers and claim that 106 countries have defaulted a total of 250 times since the end of Napoleonic Wars. Ahmed et al. (2010:39) simply labels the sovereign debt behaviour as a long and dolorous history of defaults. In terms of banking crises, since the 1890 a banking crisis has occurred on average roughly every 12 years (Reinhart and Rogoff 2009:243). While Bordo et al. (2000:6-7) write that in their post-1972 sample of 56 countries each of these countries have had a nearly one in eight chance of suffering a currency crisis, banking crisis, or twin crisis in a given year. Numerous other scholars (Bordo et al. 2010:27; Bordo et al. 1999:58; Rochet 2008:21-2; Madiès 2006:1831-2; Richards and Gelleny 2006:781; Hutchinson and Noy 2005:725-6) provide further research into the frequency of the occurrence of financial crises and/or its consequences. Evidently, there is an ample number of cases of market collapses throughout history, which stretch

(16)

16

from the most recent 2008 real estate collapse in the US or the 2001 dot-com bubble all the way to the 1637 tulipomania in Netherlands. Similarly, one could argue that there are potential bubbles at place even today, for instance the Chinese property situation (Lewis et al. May 2017). Indeed, it would appear that history doesn’t repeat itself, but it does rhyme.

1.1.3 Types of Financial Crises

As foreshadowed above, there are two particular forms of crises that are relevant today – sovereign debt and banking crises – and both start with an excessive accumulation of private and/or public debt. Furthermore, as Steil and Litan (2006:102), White (2010:4) and others argue, depending on the particular sovereign’s borrowing behaviour and how the sovereign accumulates debt over time we can determine the sovereign’s vulnerability to financial crises. Based on a similar assumption Reinhart and Rogoff (2009) introduce a new concept of ‘country graduation’ that maps countries’ historical development in relation to sovereign debt and banking crises. The concept of country graduation is fairly simple and it provides a very useful observation of how individual countries develop financially over time. Following the distinction between the two types of relevant crises Reinhart and Rogoff (2009:141,151-3) find two interesting observations; 1) many now-advanced economies have graduated from a history of serial defaults on sovereign debt or very high inflation, and 2) graduation from banking crises has proven elusive. In effect, for advanced economies during 1800-2008, the picture that emerges is one of a serial banking crises. It is necessary then to investigate what are the differences between the sovereign debt and banking crises. Namely, we must explore what are the circumstances and consequences of the individual types of crises and why countries can seemingly graduate from the status of a serial debt defaulter but at the same time evidence contrastingly suggests that they struggle to escape the threat of banking crises.

Let us begin with elaborating on the case of countries graduating from a status of a serial defaulter on sovereign debt. An external debt crisis is defined as the failure of a government to meet a principal or interest payment in the due date (or within the specified grace period). These

(17)

17

episodes also include instances in which rescheduled debt is ultimately extinguished in terms less favourable than the original obligation (Beers and Chambers 2004 in Tomz and Wright 2007:353). The domestic debt crisis is the same but with the addition that domestic debt crises can involve freezing banks deposits or forcible conversion of such deposits from dollars to local currency (Reinhart and Rogoff 2009:11). Historically the most serious defaulters, which we nowadays consider as advanced economies, are for instance Spain with 13 defaults, France with 9 defaults or Germany with 8 defaults (Reinhart and Rogoff 2009:99). Reinhart and Rogoff (2009:86) sadly and humorously write that Spain has established a record that as of yet remains unbroken and still Spain and France are considered as countries that have officially graduated from a serial defaulter to a non-defaulting sovereign. The question therefore is: how can a country graduate from its state of a serial defaulter to a non-defaulting sovereign even though the countries that currently have a label of a non-defaulting sovereign historically defaulted among the highest number of times?

The answer does not necessarily lie in the frequency of the defaults but more importantly in when did those defaults occur. Although, both Spain and France have been considered serial defaulters between 16th_{to 19}th_{century they have not defaulted on their debts in the 20}th_{or 21}st

century, therefore they can be considered as “successful graduates” from the serial defaulter status. Today’s developed countries were going through the same phase of development as the current emerging markets and they experienced recurrent problems with external debt defaults in the same way that emerging market countries do today. Sovereign debt defaults can thereby be linked to the country’s status of an emerging market but once a country converts to an advanced economy it is likely to stop defaulting on its debt. This is because emerging market economies have to first undertake several very difficult steps, such as liberalising capital account or flowing its currency, before they can satisfy the conditions of being an advanced economy. Many of these steps can bring the emerging market into financial trouble and therefore force it to default as the ability to further borrow is limited compared to the ability to borrow that advanced economies have. Kaminsky and Reinhart (1999:39) present supporting evidence of this link between crises and financial liberalisation, which suggests that on the path to a financial maturity countries are likely to experience sovereign debt defaults and crises due to trade liberalisation,

(18)

18

floating of exchange rate, and capital controls removal, the same way that Spain or France have. It seems that patience and pursuit of country graduation are the only ways to solve the problem of pathologically defaulting sovereigns. However, it takes a long period of time for a country to transcend from an emerging market economy into an advanced economy, therefore we are likely to see more sovereign defaults in the future2_.

We have established that given time countries are capable of graduating from the status of a serial defaulter into a country that does not default on its debt. On the contrary, according to Reinhart and Rogoff (2009:153), a graduation from banking crises has proven elusive thus far. Steil and Litan (2006:99) support this point by stating that crises are brutally difficult to contain and resolve with no one having the required economic medicine. The reason why countries are unable to graduate from banking crises, in contrast to sovereign debt defaults, is because of the different dynamics that come with both types of the crises. Above we have discussed the nature of sovereign debt defaults, how they ensue on countries’ path to financial maturity, and how countries are capable of graduating from the serial defaulter status. The difference is, that unlike the sovereign debt defaults, banking crises more often act as an amplification mechanism rather than the trigger of a recession and they rarely occur in isolation (Solimano 2016:283; Freixas et al. 2015:77). De facto, they tend to be accompanied by other crises, such as exchange rate crises, domestic foreign debt crises, asset market crashes, and inflation crises (Eichengreen and Bordo 2002:28; Reinhart and Rogoff 2009:145). Another significant factors is the banking system’s vulnerability to bank runs, which forces the banks to sell off their assets, often at a lower price in fire sales, in order to weather the ensemble of money withdrawals that people demand (Mian et al. 2014:20). The high demand for money withdrawals from banks can turn some assets, which are liquid in normal times into very illiquid ones and banks can face foreclosures because of their inability to sustain the demand for the money in their vaults (Hinds 2006:136). One could argue

2_{Also, the sovereigns have barely any incentives that would push them into repaying their debts since the}

repayments dynamics have changed. Instead of using so-called gun-boat diplomacy where sovereigns were able to force other sovereigns to pay up their debts through the use of military is not applicable anymore. The fact that the ‘international seal of approval’, as Keohane calls it (1998:3), frowns upon the use of gun-boat diplomacy is of course a good thing but at the same time it also erodes the incentives to repay

(19)

19

that a single bank facing a bank run can always borrow from other banks, however if the bank runs occur throughout the whole system the inter-banking credit market freezes and banks do not lend each other as confidence evaporates. Since banking crises are related to other forms of crises, which are arguably inherent to the system due to market dynamics (consumer euphoria during good times, countries striving for trade liberalisation, etc.), in turn a graduation from banking crises does not seem possible3_.

1.1.4 Emerging Markets and Advanced Economies

Thus, crises, whether in form of sovereign debt or banking crises, plague both the emerging markets and advanced economies. Nevertheless, these countries face different risks when having to deal with either of the two types of crises. Crises seem to be more frequent in the emerging market economies, due to the inevitable steps that have to be taken on the journey towards trade liberalisation, however they are not limited to these economies. A sovereign default can lead to a downgrade of the sovereign’s credit profile and a consequent decrease in the number of creditors willing to provide future financing. As Hanson (2005:16) writes, emerging markets can get suddenly cut off from the international credit market because of the perceived net worth falls of the borrowing governments.

On the other hand, a banking crisis and a loss of confidence in global financial centres can produce a sudden stop of lending to the countries at the periphery (Reinhart and Rogoff 2009:74). Essentially, the capital flows channels stop pumping capital to the emerging markets, unrelated to whether the underlying economic situation in the emerging markets is good or bad, and with the inability to obtain credit from elsewhere the economic activity in emerging market economies

3_{One could argue that the problem stems from market actors’ violation of the optimal game-theoretical behavior}

and that the financial system faces a typical ‘problem of the commons’ and free-riding scenarios. This in turn could be solved by a strong regulatory pressure from the central bank. The problem with this argument lies in the presumption that market actors would not get around the regulatory measures. As we saw in the past the shadow banking sector found its way around and fueled asset bubbles anyhow.

(20)

20

contracts and debt burdens press harder against declining governmental resources. Banking crises in the core financial centres inherently consist of such a domino effect that has drastic implications for the emerging markets because they rely on the capital inflows from the advanced economies but if the advanced economies are unwilling to offer the credit then the emerging markets suffer the consequence despite whether their economies have been performing well or not (Acharya et al. 2011:70). Kindleberger and Aliber (2005:107) especially stress the chance of international contagion in their work. Reinhart and Rogoff (2009:74) argue similarly by claiming that banking crises have historically been contagious in that investors withdraw from risk-taking, generalise the experience of one country to others, and reduce their overall exposure as their wealth declines. The consequences are clearly deleterious for emerging markets’ ability both to rollover and to service external sovereign debt. Unlike the sovereign debt defaults, we can see that banking crises represent a much higher chance of domino effect contagions, which can easily spill over further because the capital flows channels freeze and the capital desired and needed by emerging markets stops flowing (Soederberg 2002:617). This contagion effect is further exacerbated in many emerging markets because they tend to overborrow when the times are good, which inevitably leads to much higher vulnerability when the capital flows from the advanced economies freeze. Therefore, the problem at hand points to the financial system as the core of the problem, not simply emerging markets and their debt. In either case, emerging markets are more vulnerable to crises than the advanced economies are4_{(Patel and Sarkar 1998:58), whether they are currently}

performing well or not, simply because of their dependence on credit from the advanced economies.

To recap, we have established that financial crises are not analogous to Taleb’s definition of a black swan, except for the agreement on the high impact. This has paramount implications for the financial system because financial crises are especially dangerous in our modern era due

4_{I is very important to point out here that I am referring strictly to vulnerability due to the dependence on credit}

from advanced economies. That does not mean that crises always have more severe effect in emerging markets than in advanced economies. Examples of this are India or China that performed comparatively better during the Global Financial Crisis than many advanced economies have.

(21)

21

to the complexity and interconnectedness of the world with the financial system and the diverging effects that crises have in different countries. Therefore financial crises pose a much higher chance of banking crises or asset market crashes spilling over to the rest of the world and resulting in more global financial crises. As Reinhart and Rogoff (2009:223) point out, in analysing extreme shocks such as those affecting the US economy and the world economy, the standard macroeconomic models calibrated to statistically “normal” periods of economic performance may be of little use. The false perception that plagues the financial system must be exchanged for a more accurate understanding of probabilities and this must be integrated into the risk management systems of both regulators and financial institutions. It is here that Isiah Berlin’s (1953:1-2) anecdote about hedgehogs and foxes fits well. In one of his essays, Isiah Berlin, compares the character and mental approach to situations by the two distinct animals. A hedgehog knows one big thing – a strategy –, which he follows stubbornly, and the fox knows many things – strategies – and adapts to a given situation accordingly. Projecting this anecdote to the financial system, which currently exhibits the hedgehog’s stubbornness, I argue that we need a third order paradigm shift (Hall 1993:279) from the deeply entrenched neoclassical status quo towards a more fox-like way of thinking. Opening the financial system to unorthodox models and theories is key to achieving a more robust and adaptable financial system.

1.2 The Three Swan Model

The 3SM serves as an analytical framework that supports the main arguments in this thesis and it provides a conceptually more accurate real world characterisation of financial risks than Taleb’s dichotomous white-black swan model does. In particular, the 3SM creates a triad of swans – Black, White and Grey –, which all characterise different kinds of real world events in terms of their rarity, predictability and impact. The purpose is to illustrate that financial crises, discussed in the previous section, are by nature different to the real black swans. Within the 3SM (see Fig. 2) the black swans are by definition truly rare, unpredictable and have a high impact. As an example I give natural disasters because even with the advancements in modern technology natural disasters, such as earthquakes or tsunamis, are considered rare, unpredictable and with an

(22)

22

extremely high impact. The only way to for instance account for them would be with extreme risk averse behaviour such as never visiting areas where these disasters are more likely to happen, or in the world of finance one would have to extremely diversify their portfolio so when a black swan appears only parts of the portfolio will be affected in seclusion. White swans are the polar opposite of black swans and by definition they are not rare, they are predictable and their impact is fairly low. An example are standard market fluctuations, which happen often and can be modelled with good statistical methods (ALM, normal bell curve distribution) that optimise for such fluctuations. Finally, the grey swans cover the idiomatic grey area hence they are not rare, they are predictable to an extent and they most certainly have a high impact. The grey swan category consists of financial crises. As I have argued before, financial crises are not rare if we apply a more extensive scope of the historical narrative. They are predictable because of the early warning signals and red flags that show up with the presence of financial crises. However, unlike with the white swans, the predictability of grey swans is a much larger issue due to the long-term horizon, which makes it more difficult to accurately pinpoint when the crash will occur. Differently put, grey swans are only predictable if one is looking and specifically modelling for them. Lastly, as history has shown us numerous times, grey swans have a very high socioeconomic impact on the world.

Source: Personally created for the purpose of this thesis

(23)

23

To relate the 3SM back to the research question and to the literature on the frequency of financial crises we have to ask ourselves the question: how does the 3SM fit in? Currently, the market actors who consider financial crises as black swans simply ‘define’ the problem of financial crises away by claiming that by definition financial crises are black swans and thus unpredictable. I argue that by adopting the 3SM market actors will be better at recognising risks of various real world phenomena. In other words, within the 3SM framework the problem of financial crises cannot be defined away anymore and by utilising certain techniques that will be discussed later in the thesis the market actors can begin looking and modelling for grey swans.

[2] Risk Management: The Proverbial Diamond In The Rough

Thus far we have established that instability plagues the financial system, while financial institutions retain a false perception about the frequency and probability of financial crises. RM is a very young but a very promising field that has the theoretical promise and the practical possibility of improving market actor strategies and the stability of the financial system in general. Nevertheless, the field of RM also suffers from some internal issues, hence ‘a diamond in the rough’. This section of the thesis will investigate three problems that need to be resolved in order for RM to make internal RM models of market actors more robust and to enhance the adaptability of the financial system to consequently reduce the risk of financial crises. First, I will elaborate on why the way in which risk is typically conceptualised represents an important misunderstanding of the nature of risk. Second, I will discuss why the field of RM lacks a common language. Third, I will tackle the deeply ingrained ideological controversy about how to analyse and deal with risk in particular terms. Generally, the discussion in this section of the thesis should illustrate what are the contemporary problems within the field of RM and how they can be solved.

(24)

24

2.1 The Misconception of the Nature of Risk

There are two main reasons why a misunderstanding emerges in the way in which risk is typically conceptualised. One, the topic of risk is complex and not easy to understand due to the underlying probabilities. Two, we, as humans, are subject to heuristic biases that act as dominant factors in our decisions-making process. Since RM focuses on mitigating future risks and since future is full of uncertainty many of these biases emerge and effectively blur the overall understanding of risk within a given situation. Subsequently, this leads to a biased decision-making process. To begin with, why is the topic of risk complex and not easy to understand? A brief example of a ‘common mode failure’ should suffice as an answer. Consider a flight, Hubbard writes (2009:5), where the tail-mounted turbines fail catastrophically, causing the fast spinning turbine blades to fly out and cut the lines to all three redundant hydraulic systems that make the aircraft uncontrollable and as a result the plane crashes. After the incident the aviation officials would refer to this failure as one-in-a-billion event. However, as a more thorough mathematical inspection reveals, this was a typical common mode failure that most certainly is not a one-in-a-billion event. The common mode failure refers to an event where a single failure causes further failures of multiple other components and results in a systemic crash (Block and Savits 1981:456). Indeed, if the components failed independently of each other at the same time, that would be considered as an extremely unlikely scenario. However, because all three hydraulic systems, in this case, had lines near the tail engine where the initial failure occurred, the single failure subsequently damaged all three hydraulic systems at once. Thus, the common mode failure wiped out the benefits of redundancy. These probabilities are not difficult to explain ex post but to predict common mode failures ex ante in complex systems such as airplane machinery is extremely difficult due to the interdependency of the underlying probabilities of a variety of possible events. Furthermore, imagine how extremely difficult it would be to identify all risk interdependencies within an even more complex system such as the financial system. The failure of AIG, Lehman Brothers and Bear Stearns is also an example of a common mode failure. If all three of these financial institutions went bust at the same time but independently of each other, that would be extremely unlikely. Nonetheless, as we have seen their failures were not independent. We can now see how understanding risk within complex systems can be incredibly difficult.

(25)

25

Next, BE tells us much about how and why errors in RM emerge due to the heuristic biases that affect our decision-making process. Kahneman and Tversky (1971; 1972; 1973; 1974; 1979) are the founding fathers of the BE field and in collaboration they studied many heuristic biases. To illustrate I will explain three5_{: 1) representativeness bias, 2) anchoring, and 3) overconfidence. The}

representativeness bias refers to a heuristic where a person who follows this heuristic evaluates the probability of an uncertain event, or a sample, by the degree to which it is: (i) similar in essential properties to its parent population; and (ii) reflects the salient features of the process by which it is generated (1972:431; 1973:207; 1974:1124). In other words, we tend to judge odds based on what we assume to be representative scenarios. The famous example created by Kahneman and Tversky (1972:434) is presenting an individual with a series of six coin flips and asking them to determine which result is more likely. If H = heads and T = tails, is HHHTTT or HTHTTH a more likely outcome? As it turns out, both scenarios are equally as likely, however most people assume that because the first series appears less random than the second series, then the second series is a more likely outcome of six random coin flips.

Anchoring is a heuristic where individuals make estimates by starting from an initial value that is adjusted to yield the final answer (Tversky and Kahneman 1974:1128). Simply put, when the same question is asked but in different forms, the different starting points will yield different estimates, which are biased toward the initial values. Tversky and Kahneman (1974:1128) designed an experiment where two groups of high school students had to estimate, within 5 seconds, a numerical expression that was shown to them. The first expression was 8*7*6*5*4*3*2*1 and the other one was identical but in reversed order 1*2*3*4*5*6*7*8. Because the result of the first few steps of the multiplication is higher in the descending sequence that in the ascending sequence, the former expression is likely to be judged larger than the latter.

5_{There are many heuristic biases discovered by the BE scholars throughout the years but it is not the purpose of}

this thesis to extensively investigate numerous cases. I discuss few simply to illustrate how these biases affect our decision-making process, which can lead to error in perception of risk.

(26)

26

Both of the authors predictions were confirmed. The median estimate for the ascending sequence was 512, while the median estimate for the descending sequence was 2250. The correct answer is 40320.

Finally, overconfidence is another bias that relates to individuals’ prediction and how confident they are with their estimates. As the term suggests, we tend to be naturally overconfident in our predictions and we underestimate real risks systematically. Essentially, individuals, whether university students or high-level managers, have a tendency to estimate the probabilities of events too optimistically. An abundance of research (Slovic et al. 1980: 181; Griffin and Tversky 1992; Koehler et al. 1996; Brenner et al. 1996; Broihanne et al. 2014:72) has been done on overconfidence and the general consensus in the literature highlights how the overconfidence in our predictions leads to many systemic errors, of which many can end up catastrophic.

In general, we can see from the three demonstrated examples that the heuristic biases can cause logical errors in our assessments of the probabilities of risks in given situations. On top of that, RM already deals with complex systems, which are extremely hard to navigate. Therefore, the complexity of the system and the heuristic biases that we are subjected to can lead to many misunderstandings in conceptualising risk6_{(Gigerenzer 2008; 2013; 2017). I argue that the}

solution to some of these problems can be found in the BE literature, which researches these biases and in turn tries to adjust individuals’ decision-making process so that the decision maker is more aware and better prepared for accounting for the heuristic biases7_.

6_{That said, some authors suggest that often simple solutions to complex problems are a better solution –}

less-is-more principle (Gigerenzer and Brighton 2008; Gigerenzer 2017)

7_{There are several methods with which the BE scholars tackle heuristic biases, however discussing them also is not}

(27)

27

2.2 Adopting A Common Language of Risk

Currently, the definition of risk is not universal because different agents employ different definitions of risk. Consequently, the RM debate takes place on the basis of contrasting terms of reference (Hubbard 2009:79). As such, this lack of common language leads to diverging opinions of how risk should be managed. Therefore, we have to answer the question: why do different agents employ different definitions of risk and how to solve this problem?

The reason why different agents use different definitions of risk is because of the way one may perceive the risks that real world events present. In other words, agents employ diverging frameworks of perceiving real world phenomena. If an agent adopts Taleb’s vision of the world then financial crises are by definition rare, unpredictable and with high impact. According to the framework that the agent employs they will also choose the instruments that they believe to be the most suitable for managing the risks that their internal framework confronts. In this case of an agent employing Taleb’s framework, the agent will consider financial crises as black swans, therefore in her eyes the only way to mitigate risks of financial crises would be to diversify her portfolio. On the contrary, if an agent adopts a different framework, such as the 3SM, then by the definition of the 3SM framework financial crises are predictable to a certain extent and as a result the agent will use tool like scenarios. The problem emerges when the two agents engage in a debate about how to confront financial crises. Both of the agents will be discussing financial crises, however due to the different internal risk framework they will have diverging opinions on whether financial crises are predictable and on which tools to utilise in order to effectively manage their portfolio.

To solve this problem, I argue that if all the parties adopt the 3SM and clarify, which of the three swans they are currently confronting, then we can arrive at common grounds and hopefully a feasible solution to solving the market phenomena at hand. As shown before, the 3SM proposes different ways of how different world situation, ergo different swans, should be modelled.

(28)

28

Additionally, the 3SM introduces a more accurate real world characterisation of financial risks than Taleb’s dichotomous white-black swan model does. Unlike in Taleb’s framework, in the 3SM agents are offered tools with which they can scan for potential grey swans, thus the agents are capable of confronting these risks actively in contrast to passive diversification and dependence on luck.

2.3 Ideological Controversy

As is common to any academic and professional field, there tend to be several schools of thought that advocate various approaches on how to effectively or successfully solve problems within that field. The same applies to the field of RM, particularly there is a lot of disagreement on the methods that should be utilised in order to mitigate risks to the user entities with the highest level of efficacy. The field of RM is distinctly polarising because the two opposing ideological camps very vigorously defend their point of view. On one side, there are the scholars who pioneer the use of qualitative risk analysis where it is preferable to be approximately correct rather than precisely correct since we are dealing in business of predicting the future (Calder 2016:83). On the contrary, the second camp advocates purely mathematical models and bases their predictive superiority on the precision of the data without the negative effect of human error, which arguably creates significant problems for the qualitative risk analysis, especially heat maps and rating grids. I will argue that both sides of the RM quarrel build their argument on a valid rationale but in general I will propose that in order to achieve truly robust internal RM systems we need to adopt a hybrid model that uses both approaches, each under certain circumstances. In other words, I argue the hypothetical middle road, which utilises strengths of both of the camps. Additionally, I urge that the qualitative school has to embrace new strategies in order to enhance its predictive capabilities. Namely, I will introduce scenario planning (scenarios), what scenarios are and how they contribute to our goal of making internal RM models of market actors more robust. However, the detailed investigation of scenarios will take place in the next section, which is aimed at explaining the details.

(29)

29

The quantitative camp has many members that very zealously argue in favour of mathematical methods of risk analysis but the discussion pushes even further where the quantitative side of the argument compels for a complete dismissal of the qualitative methods. Hubbard (2009:74) is amongst those who wish completely to abandon the qualitative approach and he argues that we need to discontinue the using the popular but ineffectual scoring methods regardless of how practical they appear to be. Hubbard (2009:77) claims that the qualitative methods are entirely ineffectual and even with their substantial popularity we should completely dismiss them. His argument builds on insisting that numerous arbitrary rules and values created by the scoring methods not only fail to consider subjective risk but they introduce errors of their own, hence they make the final decisions worse. The presence of subjective risk within the models then sways the results and cannot guarantee a reliable output from models that work on arbitrary scales of High-Medium-Low or similar. Cox (2008 in Hubbard 2009:123) strongly argues in the same direction and even labels the qualitative methods as ‘often worse than useless’. Furthermore, Hubbard (2009:122-3) argues, scoring methods are virtually always developed in isolation from any scientific methods in risk analysis and decision making, therefore there is no empirical evidence that these methods improve decision making whatsoever. In his book Hubbard (2009:122-3) lists three main reasons why scoring methods are problematic. First, he argues that the methods have false perception of risk and uncertainty mainly due to the fact that they are developed in isolation and the ambiguity of the scales they use. Second, there is no consistency among the users of qualitative methods about the descriptions of likelihood since they use similar but not universally standardised approach. Third, there is an abundance of internal error because of the unreliability of human input and also because of the structure of the models. Additionally, the problem that the scoring methods have is the problem of design (Hubbard 2009:123). The design problem emerges because the model design fails to address the issue of how to remove subjective judgement of uncertainty. The individuals who use these methods suffer from biases that we have discussed in the previous section and these biases have profound impact on the end result that will be the product of the method. Therefore, according to Hubbard (2009:123), the failure to remove the input error and the internal error results in a very low quality analysis, which

(30)

30

the users are often better without. To illustrate the point, Hubbard (2009:124), uses an example of looking at the world through a frosted lens. If we look at a bridge, which has many engineering mistakes and is unlikely to stand on its own a detailed inspection of the bridge will reveal the flaws in its design. However, if we look at the bridge through a frosted lens the picture gets blurred, the little flaws disappear and suddenly the bridge design looks perfectly fine. In other words, the frosted lens of scoring methods does not alleviate the fundamental problem of limited information. Even worse, it simply makes you less aware of it and with the nice packaging of heat maps or scale matrices the results look plausible.

The quantitative camp’s argument against the use of qualitative methods undeniably has some strong points to back their line of reasoning up. Nevertheless, if the qualitative methods are so detrimental to the users of qualitative risk analysis methods, what should we then result to using according to the quantitative school of thought? The quantitative methods do not base their models on arbitrary scales and grids but on statistical models, which try to reduce uncertainty or provide as much certainty in their simulations as possible. Considered the most sophisticated and the most abundantly used method is the Monte Carlo simulation (Hillson 2014:147). Monte Carlos are mathematical methods that use historical data as an input variable of their models. The historical data then projects thousands of possible simulations into the future and the weighted average is seen as the most likely outcome to occur and the company can therefore optimise accordingly to such a scenario. Authors like Hubbard (2009:208) claim that the single best method to adopt and master for the analysis of uncertainties is the Monte Carlo simulation. Hubbard also argues that Monte Carlo is the single best hope for mastering the analysis of risk in any organisation.

Still, there is no blind belief that Monte Carlos are flawless, as a matter of fact Hubbard (2009:167) writes that Monte Carlos should be subjected to scientifically sound testing methods in order to ensure that the results are reliable. Due to the nature of the Monte Carlos the

(31)

31

mathematical models face two distinct issues. One, they have to be back tested in order to ensure that the models are reliable. Even if the predictions seem to work without disasters happening, it does not mean that the model is effective. Subjective perception of success of the individuals who perform the Monte Carlos simulations does not provide a reason to believe that the method does not have a negative impact simply due to the lack of failure at the moment. Two, the models highly depend on the data that is inserted into the model (Culp et al. 2008:177-8). This can further cause two more issues. First, same as with the qualitative methods, an input of data can create subjective bias depending on the data we insert, from which period is the data taken and how reliable the data actually is in terms of the database (Smith 1979:780). Second, the data can contain a so-called noise. If the noise is present throughout the whole data set, the aggregate of all the noise can actually create a significant amount of bias to the final output of the model. As such, the quality of the data is extremely important for the Monte Carlo to succeed. Thus, since the methods are entirely dependent on the data input, by default the predictive capability of the methods will be restrained by the extent of the data input. To use Hubbard’s own analogy, I would argue that seemingly the quantitative methods use a frosted lens of their own.

Looking at both the qualitative and quantitative methods, we can see that each suffers from certain limitations, which is acceptable since the aim is to predict the future. I argue that either approach should not exclude the other’s utility because both of the methods, the mathematical and the mental ones, are extremely useful for different purposes. The mathematical models are extremely efficient at predicting the white swan phenomena because historical data is a very rich and a reliable input in this case. On the other hand, Monte Carlos will fail at predicting the black8_{or grey swans because of the limited scope that the data input allows (Kloman}

1992:309). On the contrary, mental modelling, while lacking the precision that historical data provides, is not restricted to the boundaries of what happened in the past. Monte Carlos have that restriction because they essentially project optimised historical data into the future (Bood and Postma 1998:2). Indeed, there are many ways that the mathematical models can attempt to

Can reformed and improved risk management techniques make the financial system better adaptable and more resilient to financial crises?

Master Thesis Research Project: MSc Political Economy